Once upon a time I was an undergraduate at Cambridge. Back in the day when men were real men, women were real women and Cambridge’s contribution to the national sense of humour were people like John Cleese, Tom Sharpe and Douglas Adams. Cambridge students were well aware that we were at the top of the intellectual tree and generally proud of the fact. In fact for many of us (e.g. your humble correspondent) coming to Cambridge was where we discovered that there were people far far smarter than we were for pretty much the first time in our lives. It was humbling. And once we got over that, exhilarating. You could have discussions about quantum mechanics or code optimizations or whatever with people who understood the concepts at least as well as you did. You could be lectured by professors who had Fields Medals or Nobel Prizes; often they had literally written the standard text book in the field. Sometimes they would drop digs about colleagues or rivals and you had to know that “Tony” was Tony Hoare who was in charge of Computer Science at the slightly inferior other place. And so on. Yes it was elitist. I’d say it was proudly so. Yes there were concerns about there being too many Old Etonians and other “public school1” students compared to those from state schools but the solution was seen as going out to state schools and ruthlessly headhunting the top students rather than putting a cap on the number of privately educated students.

Some forty years later it seems that Cambridge no longer aspires to such intellectual heights - at least not for some areas of study. It is true that back then we generally held the Education Department and the associated students of Homerton College to be at a slightly lower level of intellect - and since they were mostly of the female persuasion many male chauvinist jokes were made at their expense. Also those of us reading for the more challenging degrees also disparaged some of the less challenging ones (sociology, land economy…) but even the most elitist MCPs agreed that even the ladies of Homerton had some intellectual assets to go along with what John Cleese might have called their attractive large tracts of land. However I doubt that even the ladies studying education back then would have put up with the quality of tuition offered today by professors such as J Arday, PhD.

However it isn’t just that the quality has gone down, what has also happened is that the racism, sexism and disableism has gone up at the same time as a deliberate part of the process. What do I mean by the racism etc. has risen?

Start with this excellent substack by Katy Barnett

What Katy Did

Arday and the academy

The academy’s latest scandal is the downfall of Professor Jason Arday, formerly of Cambridge University, until he resigned recently. His own account indicated that he had overcome considerable disadvantage: he was a Black person who grew up on a council estate, he was autistic and had been non-verbal until he was eleven, and he was unable to read or wri…

Read more

4 days ago · 65 likes · 21 comments · Katy Barnett

You should read the whole thing because it is truly excellent, but the most relevant bit is this:

Arday’s appointment might make Arday’s enablers feel a righteous glow in the short term—“Look at how wonderful we are, we appointed a Black, disabled scholar!”—but when the web of falsehood disintegrates, it leads people to doubt the worth of all Black or disabled scholars more generally. It also does a disservice to all scholars and students at an institution. From what I have seen on social media, several former Cambridge graduates who had come from less advantaged backgrounds, and had worked exceptionally hard to be admitted, felt particularly betrayed.

I can understand this: they must feel that their hard work has been mocked. Instead of working so hard, why did they not embroider a story of sadness and use that instead? I, too, have felt this bitterness. Am I fool for not using my disadvantage?

If you want to know my “lived experience” before you think I’m worth listening to, you can have it. I have suffered from cerebral palsy since birth, because I was born at 29 weeks—and yes, I have proof of this!—but I do not think that this is relevant to my academic work. […]

I do not think my disability is a reason to promote me, nor do I think it is a reason why I am better than other people. It is certainly not a reason why I am more worth listening to on questions of private law.[…]

Just presuming that someone is worthwhile simply because he is Black and disabled and from a disadvantaged background is patronising. But so, too, is dismissing someone simply because he is Black and disabled and from a disadvantaged background. These two positions are two sides of the one coin: using a simple, superficial heuristic to decide worth, without looking at the reality.

The sole relevant consideration is does this person produce good work, and can they do what is needed? I worry that the coin may flip in society, and we’ll go to a position where women and those from minority or disadvantaged backgrounds are presumed to be worthless. If the academy continues on this path, they will produce the very thing they fear: a backlash against minorities.

This is bit at the end is exactly the problem. If you promote incompetents to positions they are not suited for because of their gender, race etc. then you inspire everyone to look at other people with the same gender, race etc. and wonder if they too are incompetent. This is not a philosophical point, it is a real practical issue. I know people, some of whom read this substack, who make a conscious choice of choosing the older, white, male doctor if they can because they can be sure that this doctor was not given a helping hand due to his gender, race etc. and that therefore he will be more likely to give correct medical treatment.

Medicine is far from the only field where this is the case. You may recall the DC crash in early 2025 where a female helicopter pilot collided with a jet. There was a lot of commentary wondering whether (or outright saying that) she was unqualified and only got the job because she was a woman of the right (i.e. left) political leaning. I don’t know whether that is true or not - from what I recall at least some of the so-called evidence against her turned out to be rubbish - but the fact that such commentary sounds plausible and the responses are mostly of the form “Sexist!!!” with no evidence provided makes my point. Likewise the Florida university bridge collapse where the primary engineering firm was all female and (IIRC) Hispanic.

Share

The fact that the phrase “DEI hire” exists and that everyone knows that a DEI hire is known to be less capable is the result of this racist behavior. And I do in fact mean racist. To quote Katy again:

In fact, their attitude strikes me as profoundly racist and not in the least “progressive”. The subtext is, “Black and disabled people can’t manage it, let’s turn a blind eye to any flaws and let this guy in anyway, so we can look good.”

This is the soft bigotry of low expectations.

It says “we know you brown, gay, etc. sorts are less capable so we’re going to give you a gold star participation trophy and everyone will pretend you are just as good as the real thing.” Effectively what it says is exactly what the racist, sexist good old boys used to say; it’s just that unlike the good old boys who used that as an excuse to hire their incompetent inbred drunkard cousin, they use it as an excuse to hire an unqualified minority candidate and give themselves pats on the back for giving it to “the man”. Forgetting that these days “the man” is more likely to an AWF(U)L2 just like them - see e.g. Hilary Cremin, the woman who cringeingly claimed that Arday was “the best in the world in terms of the research that you do.”

Eric Rasmussen lists all of the people involved in this hiring choice and has photos of them from their official bios (which he also publishes)

Ras-Stack

Who Hired the Ridiculous Jason Arday at Cambridge? I Have the Names.

Jason Arday was appointed Professor of Sociology of Education (in its Faculty of Education) at Cambridge University in 2023. Since then he has been found to be an incredible liar and a miserable scholar. Some items…

Read more

8 days ago · 542 likes · 191 comments · Eric Rasmusen

I think all of them should be signed up as members of the BNP (if it’s still around) because they are just as guilty of racial discrimination. This is more or less what well known US Academic John McWhorter says in this interview3, where he also identifies the people who enable the Ardays of academia as racist:

It’s a perfect example of how the Latinate term “Affirmative Action” is a euphemism for lowering standards. It wasn’t that Cambridge resisted some general tendency to only see excellence in whites. They chose a black person with no concern for excellence. In truth, they consider black people cognitively inferior.

Almost certainly the best advice in this area was given by US Supreme Court Chief Justice Roberts back in 2007 when he wrote: “The way to stop discrimination on the basis of race is to stop discriminating on the basis of race.”

Really. That’s it. Especially if you are Cambridge and, allegedly, one of the top half dozen universities in the world. By all means encourage those from less privileged backgrounds to apply, but don’t give them any other advantages beyond initial access. Just recruit the best, and the best of all races, genders etc. will apply because they know that the Cambridge brand means they aren’t a DEI hire, but are actually capabable. If it turns out that that means that Nathan Cofnas is right and there are only a small number of people capable of the job from certain ethnicities, then just accept that as a fact and move on.

A Branding Coda

Though the Cambridge brand is looking pretty tarnished right now in certain fields. In fact, if you wanted to parody the wokeness of today’s Cambridge along the lines of Tom Sharpe in Wilt, you would fail because the university has gone and done it to itself. Take the current PhD students and their theses in the “centre for gender studies”.

E.g.

Waruguru Gaitho

Legal Mobilization by Black Lesbian, Bisexual and Queer (LBQ) Womxn in Kenya and South Africa

or

Hannah Gallagher-Syed

A Genealogy of Premenstrual Dysphoric Disorder

or

Sarah Knaus

Exploring gendered embodiment in transgender and gender-diverse adolescents at the intersection of exercise and eating behaviours, social inclusion, and mental health: Towards a desire-centred approach in paediatric gender-affirming healthcare and research

Reading this you begin to understand why no one in Cambridge raised an eyebrow at the work of Jason Arday, whose PhD was entitled “An Exploration of Peer-Mentoring among Student Teachers’ to Inform Reflective Practice within the Context of Action Research” and who produced papers with abstracts like:

The permanence of systemic racism in the UK and USA means that Black people are disadvantaged in myriad ways, including within the Academy. While the disproportionate impact of COVID-19, alongside the Black Lives Matter movement, has increased awareness of the challenges faced by Black communities, these issues remain, both in and beyond higher education. Furthermore, there is still a paucity of research individualising the experiences of Black people, who are often homogenised with other ethnic minority groups. This paper explores the impact of COVID-19 on UK and US Black students and academic staff, utilising a critical race theory (CRT) framework. Analysis revealed that Black students and staff experienced COVID-19 against the backdrop of racism as a “pandemic within a pandemic” [...] including racial (re)traumatisation, loneliness and isolation. Other themes included precarious employment and exploitation. Recommendations are offered for penetrative interventions that can support Black students and staff in the wake of strained race relations neglecting their adverse experiences and a global pandemic.

Having thousands of non-academic eyeballs read this kind of thing does lasting reputational harm to the “top university” brand of Cambridge and, as Lisa Mckenzie writes in Spiked, to the “is a serious subject” brand for sociology.

Sadly, as Mark Atwood writes about Science Journalism, the incentives for internal reform are entirely lacking. Thus it seems unlikely that Cambridge will do away with the Centre for Gender Studies and Sociology in Education and equally unlikely that sociology as a whole will remove itself of the grievance studies fields.

However when internal reform is impossible there is always external reform in the form (in the UK) of the party of that name. When Reform forms the next British government and needs to make cuts, it seems fairly clear that there’s a significant chunk of higher education that can be removed with no loss to the national good

I’m sure the affected departments will find some hostage puppies to display to justify their existence. I rather doubt that Reform’s voters will care.

Subscribe now

Back in 2022 at the old blog I wrote this about an Atlantic article by Emily Oster:

I saw one glaring lack:

An apology for getting things wrong

She’s not alone, mind you. I don’t know I’ve seen anyone apologise, certainly no one in a position of authority. That’s a problem.

I’m willing to forgive people but has to be a quid pro quo that is an acknowledgement that there is something to forgive. If you can’t admit you made mistakes and that those mistakes killed people, bankrupted businesses and so on then you don’t get my forgiveness. And as I tweeted1 people who got things wrong up to about June 2020 get a pass because there was a ton of confusion and (especially from the PRC) misleading information so figuring stuff out was hard. But anyone who continued to advocate school closures, isolation of the elderly, mask safety theater and so on after that date needs to apologise.

Anyone who insisted on vaccination of everyone, including the young, and wanted passports, mandates or other coercive measures needs to apologise.

Anyone who mindlessly mocked recommendations for HCQ, Zinc, Vitamin D, Ivermectin and so on needs to apologise […]. So does anyone who deliberately misinterpreted President Trump’s remarks about disinfectant to imply that he said people should inject bleach. In fact the mainstream media needs to do a general apology for their un-questioning slavish devotion to “the science” even when said “science” changed its mind and their related trashing of anyone who didn’t follow “the science” as it was declared to be at the time. Come to think of it the people stating “the science” i.e. Fauci, Ferguson and co, would do the world a huge favor by admitting that they made mistakes and apologising for them.

Anyway. That was nearly four years ago. And I still haven’t seen an apology from anyone that matters. Actually I’m not sure anyone in any government or governmental organization in the world has ever apologised. None of the “scientists” have said a word that could be taken as an apology.

In the midst of all the “under the advice of my attorneys I will invoke my right under the 5th Amendment to the Constitution, to refrain from answering your questions” blather, Herr Doktor Science notably failed to apologise. Not in his opening statement, and not when explicitly asked to by Senator Bernie Moreno. He just repeated his 5th amendment spiel.

I thought I’d put the Covidiocy behind me. We’ve got wars in Ukraine and Iran. We’ve got AI running everywhere. We’ve got Trump 2.0 and (not so good) Andy Burnham in the UK. And so on. The world has moved on. But I read some of the Fauci diaries. I’ve read the commentary on it by and and many others. And my frustration has come raging back.

Subscribe now

I don’t require a jail term, repayment of all the moolah or even an admission of guilt.

Just man up (almost everyone concerned is of the male persuasion), put on your big boy trousers and say sorry.

It can be just that one word.

“Sorry”

No need to say why you are apologising. No need to say who to. Just say the word.

This particularly applies to Herr Doktor Science because he’s got a presidential pardon so saying sorry won’t land him in jail. Unless of course he shares other people’s skepticism about the validity of the autopen pardon he received.

I feel like channeling Oliver Cromwell and beseeching them, in the bowels of Christ, to think it possible they may have made mistakes. In fact talking of Oliver, my mind thinks of Tudor/Stuart era justice and wonders whether an updated version might work well. I’m not suggesting it would be even slightly legal, let alone moral, but a subscription streaming show where Herr Doktor Science and his buddies were subject to having their heads flushed in a toilet or subjected to other school type tortures until they said sorry would make millions directly. The bookies and prediction markets taking bets on how long before each one screamed sorry (or indeed whether they would have breath to scream), would make bank. Probably you could sell ads and sponsorships for millions as well.

“This apology session brought to you by …”

Though maybe not because the CEOs and marketing execs who would have to approve such spends might be nervous about whether they would be next to be made to apologise for, say, hiring illegals or H1-Bs

But seriously. Herr Doktor, follow the example of Japan.

The execs at Tepco caused a radiation leak that killed a few, made a large chunk of Fukushima radioactive and only didn’t have something worse because (AIUI) the then Prime Minister overruled their dithering. They stood in a room, faced the cameras and apologized2.

The Covidiocy scientists and bureaucrats have never done anything similar. They should and the fact that they haven’t tells the world a lot about their character.

Sir Elton John says “Sorry seems to be the hardest word” but that’s no excuse to not say it.

Update: Back in 2022, Herr Doktor Science was rather less reticent, or at least claimed to be:

Share

[Note: As usual this post is long and is best read in the app or a browser ]

Update: Added a quote from Apostalos Stamenos about the impact of False Positives that is useful and worth discussing.

Recently Substack announced a partnership with Pangram1 that allows any substack subscriber to run most text on the platform through Pangram’s AI detection algorithm. This has been controversial and I don’t really understand why. Moreover when I read (listen to)2 the complainers they seem to be stating as fact things that they don’t present evidence for, and which do not correspond to my experience with AI detection.

Everyday Ethical AI

Why I won’t be using Pangram AI scoring

On July 21, Substack Team announced the integration of Pangram “AI detection” scoring into the Substack platform on our longer Notes, comments, and articles. I’ve been seeing lots of commentary on Pangram integration since then. This article is to summarize the issues and to state clearly where I stand…

Read more

21 days ago · 149 likes · 80 comments · Karen Smiley

Karen Smiley wrote a post (linked above) saying why she objected and also updating it with information on how to disable the capability to have the pangram analysis run on a post. She also requests (which seems sensible) that substack allow authors to opt their entire substack content from pangram rather than force it to be done on a post by post basis.

However I read her post with some mystification because I don’t really understand the objections she has.

• As I’ve said repeatedly, I’m not interested in judging or shaming people for using AI.

• I don’t trust the tools to be accurate and fair — especially to people of different races, ethnicities, neurodiversity, academic training, ESL backgrounds, or women. So I don’t want to be a party to discrimination by triggering a writer to be unfairly scored by Pangram.

• Publishers who may have opted out of AI training on their writing shouldn’t have their work exposed to Pangram’s AI against their wishes. (Over the past year, about 1/3 of our SheWritesAI community members have opted out, as have I.) I won’t do that to them.

• Since JHong uncovered that Substack will pay Pangram for each use, I don’t want to drive more profit to them.

The first point is fair. I personally read some writers that I know use AI to write (some of) their text and I don’t particularly judge them for it - see below for an example. But that is because they are using the AI as a bicycle to enhance their writing capability rather than as a train to slop station3. However I also judge people who use AI in their writing and, with apologies to Karen, in most cases when I find someone is using AI am less likely to continue reading them, let alone paying to read them.

Point two about “trust[ing] the tools to be accurate and fair” is where I flat out disagree. It is certainly true that AI detection tools, like plagiarism detection tools, have had a checkered past. However it seems to me that there are two approaches to “this tool isn’t very accurate”. Approach one is Karen’s and that is to not use the tool and tell others not to. Approach two is to improve the tool. That’s my approach. I want to have an accurate AI detector tool beyond my own wetware (which, as it happens, seems to be pretty good at it). The best way to get that is to find a pretty good tool (e.g. pangram), use it and give feedback when it gets it wrong. To me, and I suspect to many, accurate AI writing detectors are valuable and we need them to be as good as possible. I will add that, in general, the kind of writing AI seems to have been trained on, and which it therefore tends to produce, is generally poor. If a detector creates a false positive then in most cases the content is uninteresting and/or irritating to read for me any way.

Point three is both wrong and impossible. First the tool is not for the writer, it is for the reader. I see no reason why a producer of non-AI generated text should be concerned about it being analyzed. In most cases, not in Karen’s case, I find myself to be suspicious of people who turn off the pangram tool because I feel that it suggests they have something to hide. Which leads us to the impossible part. Just because the direct substack→pangram method has been turned off doesn’t stop a reader from using pangram’s own tools (or even simple copy pasta into their webpage) to find out if something is AI generated or not.

For example here’s a post that I can’t run through the simple tool. I like Jake and his writing, but as I read the article, the wording got more and more AI like

So I copy-pastaed the first almost 1000 words in to a pangram window and got this result, which is also the title image of this article. If you go on the details tab it shows the first chunk being purely human the second having some AI assistance and the final chunk being all AI. I ran a follow up check on the rest of the article and it is also 100% AI. This was what my intuition told me and was why I wanted to run the article through pangram. Despite the article being unavailable for scanning (this may not be deliberate, I’m unclear on how that happens and I may well inadvertently make this article unavailable too), checking for its AIness was not exactly complicated.

That leads us to point 4, not wanting pangram to make money. The only way I can square this with logic is that she thinks AI detection is so intrinsically bad that it should be shunned (if not banned) because it is inaccurate and that pangram are basically scamming everyone. This is (see point 2) something I totally disagree with. IMHO the ability to reliably detect LLM usage is important for many uses and having substack pay pangram every time someone asks whether a post is AI generated is a good way to give pangram money and encourage them to improve their product.

There are plenty of tech companies I’d like to see go bust - Meta and Google are two that come to mind instantly - and I have no love for OpenAI or Anthropic, who are both going to go bust anyway, but pangram seems different. Unlike OpenAI or Anthropic it isn’t stealing the content of every book and website it can lay its hands on and unlike Google and Meta it isn’t doing - in fact can’t really do - the Adtech driven enshittification that has made me quit using both as much as I can4. Pangram seems to be providing a useful service at a decent price and that’s something to be encouraged rather than discouraged in my view.

Why AI Writing is Bad

Kevin Tumlinson in the video linked in footnote 2 says that AI writing is high quality; citing the use of m-dashes, semi colons, the rule of three etc. This is what I mean by something that contradicts my experience. What grates about AI writing is none of those things and if Pangram primarily uses those tells to detect AI then a) it isn’t doing what it says it does and b) it is indeed destined to fail. The problem with AI generated text is that it often contains odd metaphors/similes and has various “load-bearing”5 words and phrases that show up again and again in the prose. It also, in my experience, tends to use short declarative sentences and doesn’t use colons or relative clauses (who, which, that…). It is possible this is the style taught in modern American education but I don’t think so. For example, these initial paragraphs in a recent amuseonx post triggered my “this feels like AI” sense and pangram agreed with me:

Stand at Hains Point on a Saturday morning in spring, and you will see two things at once. The first is a crowd. Golfers are stacked at the starter's window, the range stalls are full, and the tee sheet has been sold out for days. The second is the condition of the ground they are standing on. The irrigation is failing, the drainage is worse, the clubhouse has taken on water, the cart paths are worn through, and the trees overhead include a number that arborists have flagged as hazards. East Potomac Golf Links is simultaneously one of the most popular recreational facilities in the nation's capital and one of the most neglected.

The numbers are not in serious dispute. The course recorded 124,960 rounds in 2025, which works out to roughly 342 rounds for every day on the calendar, including the frozen ones. On May 3, 2026, golfers played nearly 900 rounds in a single day. A 2026 condition assessment placed numerous facilities at the property in “poor or critical” condition, citing water damage, broken fixtures, structural concerns, and potential safety hazards. The federal repair inventory reads like a punch list for a foreclosed property, running from the parking lot and the retaining wall to the light poles, cart paths, driving range building, golf barn, and starter building. The Justice Department, defending the government in federal court, described a “state of deterioration” requiring “significant maintenance, rehabilitation, and reconstruction.” That admission was made in litigation, under the discipline of the record, not in a press release.

So here is the puzzle worth thinking about carefully. Nobody argues that East Potomac is fine. The National Park Service does not argue it. The former leaseholder does not argue it. The plaintiffs suing to stop the renovation do not really argue it either. What they argue is that the wrong person is fixing it, at the wrong speed, under the wrong procedure. That is a very different objection, and it deserves to be examined on its own terms rather than dressed up as environmental stewardship.

Let me ID some things that I find to be AI tells

• The first is a crowd. Golfers are stacked at the starter’s window, the range stalls are full, and the tee sheet has been sold out for days. The second is the condition of the ground they are standing on. The irrigation is failing, the drainage is worse, the clubhouse has taken on water, the cart paths are worn through, and the trees overhead include a number that arborists have flagged as hazards.
  I’d put a colon after “The first is a crowd” I might also use words like firstly and secondly or “on the other hand”. This short declarative sentence followed by a longer sentence is very AI. As is the second part with the sentence ending at “they are standing on”. Again a colon - or even an m-dash - might work to connect the sentence with follow on.

• The Justice Department, defending the government in federal court, described a “state of deterioration” requiring “significant maintenance, rehabilitation, and reconstruction.” That admission was made in litigation, under the discipline of the record, not in a press release.
  The first sentence feels AIish, but I can’t explicitly cite a reason why. The second uses the phrase “under the discipline of the record” which really doesn’t sound human.

• What they argue is that the wrong person is fixing it, at the wrong speed, under the wrong procedure. That is a very different objection, and it deserves to be examined on its own terms rather than dressed up as environmental stewardship.

  The first sentence is a classic rule of three, and yes it does seem AIish. The second sentence is the better tell though. “dressed up as” is an AI sort of phrase. “,and it deserves..” rather than “which deserves” or “that deserves” is AI grammar.

Worse the AI generated text usually fails to clearly make the point. (Good) Human writing tends to use language in ways that highlight and emphasize the point/argument. AI texts often seem to lose the main point in a thicket of metaphors and cool writing.

This analysis about the rise of another related AI tic is good for explaining some of the reasons why

Wondering About AI

I analyzed 16,000 articles to find AI writing on Substack

Disclosure: I used Claude Code to write Python scripts, collect articles from Substack newsletters, and create HTML charts. I conducted this research in the spirit of non-judgmental discovery and anonymized all article data. No one has been named, shamed, or blamed. Also, DM me if you’d like more details on the methodology, the data, or the code…

Read more

2 months ago · 312 likes · 243 comments · Karen Spinner

Subscribe now

Is Pangram Perfect?

TL;DR. Hell No. It seems to be the best available but it has issues.

Freddie de Boer did some pangram analysis recently and found all sorts of oddities. Some of that corresponds with my experience and I was able to manually reproduce a couple of issues, such as his Human in AI in Human thing where the opinion about whether it is AI depends on how much of what text is scanned. I, too, have noticed that the 100% AI vs 100% human with relatively few partials is extremely common. BUT… it’s getting less so. See the Jake Adelstein post mentioned above as an example. Another is a Denis Stetskov post that coincidentally links to the FdB article.

It seems to me that Pangram is using feedback to train its processes and is visibly improving. The good thing about the substack integration is that you can dispute the finding and such feedback is invaluable for improving so assuming people do give feedback pangram will get better thanks to the partnership. Still there are always going to be issues and if we accept pangram’s own numbers, it has a 0.19% false positive rate which means about one in 500 AI detections will turn out to be human. That’s not an acceptable rate to trust the output as a single source.

Update: this substack post makes some good points including:

A 0.01% false-positive (FP) rate sounds like nothing. At platform scale it is not nothing. Run that rate across a corpus of millions of posts and you get a standing queue of falsely flagged writers, refreshed every day new content publishes. A small rate multiplied by an enormous volume is a real population of wrongly accused people. That has implications for their credibility & readership.

That population is not random. Pangram’s own documentation concedes its weak zones: short text, fragments, formulaic writing, and underrepresented domains. Substack’s launch coverage concedes that writers with precise, clinical styles can trigger false positives. An accusation is sticky, and clearing your name costs more than earning it in the first place.

The first thing to note is that an FP rate of 0.19% or even 0.01% ( is being generous because that seems to be specific to academic papers) is, as I write above, bad as a single indicator when run at scale. But it’s fine for the intended use case by individuals to avoid AI slop, and especially if the targeted user (a substack reader) uses it as a way to filter out boring writers then the fact that it more likely to trigger a FP on “formulaic writing” is fine. “[W]riters with precise, clinical styles can trigger false positives” means to me that it finds cliched slop and boring marketing-speak to be bad and since I, at least, find that sort of writing to be uninteresting an FP on someone’s marketing-speak is fine by me.

However, and I think this is important, just because the FP rate is acceptable for the intended use case doesn’t mean that it won’t be abused by trolls and other scum who will attempt to malign writers unfairly just because they can. I think the trade off is worth it because pangram can use the feedback to improve its product but I understand the reason why people are nervous

PS analyzing that post as a whole says that it is

Mostly Human-Written
AI 11% AI-assisted 0% Human 89%

which sounds about right

End of update

There is no doubt in my mind that Pangram does better on longer text than shorter. It has a 50 word minimum but my experience is that anything under 300 is unreliable and you need several paragraphs - I’d say 800+ words - before you can have decent confidence in the result.

Relatedly, Pangram doesn’t seem to reliably rescan text or tell you when it scanned the text and made its judgment. From a computer science efficiency standpoint that’s a good thing. However, since Pangram is improving, reusing a result found some months ago is likely to give a different answer compared to what it would give if it had the same text scanned today.

My experience is that pangram is a valuable indicator and it is getting better but if you use it as a tool in isolation and without additional context you will be misguided.

Why AI Detection is Important

Using AI to write is not, intrinsically bad. It is, however, often an indicator of quality. For people who struggle to write coherently (in English) AI is a huge aid and worthwhile. The aforementioned Denis Stetskov has English as his third language so his use of AI is a good way that people like me, who can’t read Ukrainian, can appreciate his posts and not worry about odd wording etc. The problem is that most publishers of AI written prose are not Denis, even though they may also suffer from a similar lack of ability to write coherent English.

AI usage is, for me and many others, an indicator that what I am reading is probably not worth the time. Considering that we only have a limited amount of time to read anything, seeing whether someone new’s posts are AI generated is a great crude filter to remove the legions of AI slop, even though it may be prone to false positives.

In that way it’s similar to the DEI hire hint that, as I commented on a Sensurround post, is extremely unfair to people (like e.g. Sensurround Q Catgirl herself) who check off several of the DEI hire boxes and now have to convince the rest of us that they are actually competent at the job and didn’t get hired for some combination of skin color, genitals and preference about who they rub their genitals with.

Ironically Sensurround Q Catgirl uses AI too, and because she’s a contrary sort of cat made sure to write a 100% AI generated post.

Sensurround Q. Catgirl Writes

I Have a Confession to Make About AI

Substack recently added a feature from a provider called “Pangram” that lets readers know if AI was used in creating an article…

Read more

17 days ago · 19 likes · 12 comments · Sensurround Q. Catgirl Writes

However I don’t use the AI writing thing as a single pass/fail, I use it as one of the indicators and the same will absolutely apply to writers who forbid automatic pangram checking - though if you forbid the checking and my spidey senses detect that you are AI writing you that’s an other count against. Other indicators being whether we have subscriptions in common, whether the subject seems interesting and so on. What I also won’t do is jump up and point and try to shame the AI writers because that’s counter productive, just as it is to try and point and shame email spammers. The AI slop writers either don’t care about your complaint or, worse, use it as an indicator that you are a real person and therefore they should go after you more, while the people who need AI to help them write useful posts will be harmed by having the AI label attached to them.

The unseen benefit of the substack/pangram relationship

Finally it occurs to me that the effect of the substack/pangram relationship on the substack environment is likely to be positive even if pangram is not very reliable. The media have noted how AI generated prose (and videos etc., but that doesn’t apply) hav flooded all social media and similar sites. Substack (and perhaps some of the others) see this as a problem because low quality dross means lower reader engagement and thus eventually lower revenue in terms of subscriptions. No one is going to spend the time looking for the substacks that are worth paying money too if they are buried in a heap a AI slop. The substack/pangram relationship is a threat to the AI slop writers that their work will be junked sight unseen and thus will gain them no revenue. That’s a valid threat even if pangram isn’t 100% accurate. For the slop makers the work needed to generate slop on substack is not worth doing if they believe that no one is going to bother reading it. Yes there are potential innocent victims caught up in this, but there are more innocent victims of a different sort who will lose out if substack becomes slop dominated the way, say, linkedin is.

Denis Stetskov makes the argument in his article above that substack itself, presumably unintentionally, brought the problem on. I suspect he is generally correct here. Incentives matter and substack incentivized the publication of slop. The pangram relationship can therefore be seen as an attempt by substack to solve the AI slop problem it created. I hope it works without too much collateral damage. If the threat of AI detection can remove the slop makers then then real people using AI to help them write better content may benefit too as they too can be seen instead of buried in the pile of slop.

That incentive is why I am also very suspicious of the motives of some of the people whining about the relationship. My suspicion is that they want more AI slop to camouflage their own writing deficiencies or because they have a way to monetise “managing” the AI slop problem by, say, being paid curators. It sounds like the people crying “MAGA *ist” when confronted with complaints about incompetent DEI hires and like those DEI defenders it may work for a while, but eventually the evidence becomes clear that they are not interesting in solving the problem.

Share

1

The Substack Post

Against Claudefishing

It’s getting harder to tell what’s real on the internet…

Read more

22 days ago · 6619 likes · 1609 comments · Chris Best

and a detailed explanation of how pangram works

Pangram

How does Pangram work?

We think it is important to be able to reliably identify AI-generated writing, so we built a detector that does it. Because the problem is so important, it’s only prudent to be transparent about how that works…

Read more

a month ago · 284 likes · 122 comments · Pangram

Note: as usual this post is probably too big for email, please read in the app or a browser

Note: This post updated June 16 with three interesting relevant links - one each for SpaceX, Anthropic and OpenAI (search for Update June 16 to find them)

To the envy of many Elon Musk had a successful (so far) IPO for SpaceX and became instantly the worlds first paper trillionaire. As I saw someone put it on social media somewhere this means that

the average American is now closer to Jeff Bezos in net worth than Jeff Bezos is to Elon Musk.

There are legitimate questions about whether SpaceX is over valued or not and definite concerns about the size of the float - because the number of shares offered in under 5% of the company. One of the places I searched for that information also said that insiders, including Elon, are locked in for a year, so all those recently minted SpaceX millionaires can’t actually cash out any of their paper gains until this time next year, mind you I’ve seen other people say the lock up is 2 months or that there’s some taper. Either way it seems likely that for at least a year the total number of SpaceX shares available for regular investors is going to be relatively tiny. It also seems like the insiders were unable to sell at the IPO which was not the case at the only IPO I’ve ever been an insider in back over 30 years ago1, which of course also helped the price because it limited supply.

SpaceX is a mix of good and bad. The S1 was not a model of clarity IMO and contains a Total Addressable Market graph that was not entirely convincing

It looks like Starlink and the Falcon launchers are profitable and generating cash. Starship works well enough that we can see how it’s going to simply crush everyone else in the launch business once it achieves the same reliability as Falcon. It isn’t there yet, but it is close and its primary competitor just blew up. Moreover it is clearly the way we get beyond lunar orbit.

Then there’s xAI, which appears to be losing money hand over fist over arm over leg. Grok simply doesn’t have anything near the revenue stream of Claude or ChatGPT and most people see it as inferior to both as well as Gemini. On the other hand the Colossus data centers may be able to make money leasing capacity to their competitors. Given that other data centers seem be struggling to come fully on line, that may be a major advantage. In fact note is made in the S1 that Colossus and Colossus 2 were both built in 3-4 months as opposed to the several years it seems to take other people to build large data centers. If AI demand increases and AI data centers are not available then xAI may become the Data Center provider to lots of grok competitors and this will definitely be the case if SpaceX can successfully develop its orbital data centers.

[Aside: I’ve seen people claiming that data centers in space are IM!!!!111!!POSSIBLE!!!!111!! due to cooling issues and radiation issues. I think these people are misunderstanding (possibly deliberately) how much compute and power generation (and hence heat generation) there is in a Starlink communications satellite let alone a Geosynchronous Communications Satellite. The current mini v2 satellites are ~800kg and AIUI about 1m x 1m x 3m - roughly enough volume to fit a single AI server rack once you account for propellant etc. For what it is worth geosync satellites are “school-bus-sized and weigh several tons”.

Consider that these satellites have to handle communications with numerous users on the surface below and pass the data from them on to other satellites to deliver it to the relevant ground stations and you realize that there’s a lot of compute work going on. Just the signal processing to identify each user’s data being beamed up to it (even more so when it’s mobile phones) is a huge task, handling hand off to the next satellite and all the other back end management tasks including routing the traffic to/from the user is also not small. And transmitting radio signals that can be received by cell phones more than 300km below you is not a low power thing. Sure the data center satellites will be larger than the the starlink ones - the discussion I’ve seen puts them in the 100s of kilowatt to low megawatt range, which is dozens up to to maybe 100 AI servers based on current Nvidia tech (a single 30.5kW 2xB300 cluster is 16 GPUs across two servers in one standard 20U rack, so ten racks of them is ~300kW and 100 is ~3MW). I’m not clear on exactly how large the Starship payload area is and obviously you would not deploy standard racks and rack mounted servers but you can probably get the equivalent of more than 10 and less than 100 racks in a satellite that Starship can deploy - see geosync satellites being “school-bus-sized and weigh several tons”. That amount of power doesn’t seem like a huge cooling issue and is not orders of magnitude more than existing satellites and all these existing satellites have considerable compute power and do not apparently fail from stellar radiation issues. ]

It is, bluntly, unclear just how much cash SpaceX generates compared to how much investment it needs to grow successfully. Both numbers are unclear. The S1 put some numbers down that seem plausible, but there’s plenty of murk left. It will be interesting to see a few quarterly reports to see how these numbers develop.

The nay-sayers point out that for SpaceX to justify the current valuation it needs to have a Google or Amazon level of capital gain over the next decade (i.e 10x-100x). This they say is impossible because when you start at a valuation of a little under $2T then a 100 fold increase puts you at some significant fraction of global wealth. They are correct, but what they miss is that SpaceX isn’t tied to the globe, it has an entire solar system to exploit that no one else is going to be able to do anything about. Just colonizing the moon successfully first could be enough to justify a $200T valuation.

So SpaceX has issues, but it also has more upside than any other publicly traded company on earth. It will be interesting to see what happens in a year when people can cash out their paper profits and we have a year of financials to look at.

That’s the success story.

Update June 16: If you want to understand the driver behind Musk and hence SpaceX you really ought to read this Marc Anreessen X article.

Other companies are not SpaceX

Now we come to the likely flame outs. Both Open AI and Anthropic have been making noises about IPOs too. Both are in theory valued at close to $1T and would presumably seek to list at that kind of valuation.

However while I, and others, can see how SpaceX generates enough income to justify its price and predicted growth, that is not the case with the Big AI companies. To start with SpaceX has multiple markets that it can conquer (even if you ignore the “$22T Enterprise Applications” sector) and it is potentially the only global ISP and effectively the only affordable way to space. Anthropic and Open AI are competing in one (AI) and they have well funded competitors like xAI and Google Gemini for that market. Their leaders also display a level of self belief and attitude that everyone else are know nothing rubes that is mind-blowing. Unlike Musk, who to be fair, sometimes exhibits similar tendencies, they do not have a track record of success that justifies this kind of attitude. An example of what I mean is documented here

Cremieux Recueil

Anthropic, please don’t ruin bio for everyone

Today’s post is brought to you by my sponsor, Mechanize. They’re hiring junior software engineers at $300K/year base salary…

Read more

2 months ago · 33 likes · 4 comments · Cremieux

Another reason why Anthropic should exercise caution in opining about biosecurity is that, despite considerable financial investment in the field, they simply do not seem to understand basic aspects of scientific progress or the biomedical sciences. It ought to be a basic principle that if you fail entirely to understand some specialized field, your attempts to advocate for special regulation in that field should be entirely ignored, because it is rather unlikely that someone failing to grasp basic facts would possess anything resembling a coherent model of the threats and risks in said field.

The ignorance of basic biology displays itself at multiple levels. In Dario’s recent essay, Policy on the AI Exponential, he writes:

Many of the steps in the clinical process that previously required expensive and slow experiments may soon be done via AI simulation or analysis. [...] Areas where this could apply include:

…

The problem with criticizing this passage is that it is so profoundly misguided that one hardly knows where to begin. Does Dario truly believe, for example, that pharmaceutical companies are simply too inept to consider the idea of synthetic control groups, or is it more likely that synthetic controls are genuinely difficult to formulate and ultimately less convincing than proper randomization? Does he believe that ‘more accurate dose selection’ is an actual limiting factor in the proper design of a clinical trial, or is it more likely that given the $100M price tag of a Phase III trial, the dose selection is probably optimized enough such that remaining margin for improvement in this direction is relatively low even if you were willing to pour more cognitive labor into it?

More generally—and here is where the relevance to biosecurity comes in—Dario’s writing is all premised on some tacit assumption that performance on these tasks can in some sense be dramatically improved with sufficient supply of cognitive labor. This is wrong for the simple fact that even with a true abundance of model intelligence, this work is fundamentally bottlenecked by the inherent need to sample, slowly, from the complex distributions of real-world biology, which cannot be so easily disentangled through sheer might of pure reason.

As I’ve mentioned in previous AI posts, and as others have too, both are losing money at rates that only look good when compared to xAI. In other words rather than losing $10 for ever $1 of income, they only lose $2 or $3. But they are both bigger than xAI and have made rather more commitments for future spending. Including, in the case of Anthropic, to SpaceX/xAI in fact. Both are also hyping themselves and AI in general in ways that are in my opinion borderline unethical. Some observers would omit the “borderline” from the previous sentence.

It is worth taking a look at each individually, which I will do below, but they do have a lot of commonalities. These include being massively loss-making, bloated and inefficient, remarkably casual about security, displaying a bland disregard for the rules of Intellectual Property as they apply to others and, as noted above, doing marketing that is similar to the marketing used by crypto-scammers and others of that sort. As with crypto, LLM AI clearly has viable use cases, but also as with crypto, there’s a lot of hype and illegal activity mixed in with the legit usage. I’m not saying that Sam Altman and the Amodeis are the AI versions of Sam Bankman Fraud, but they aren’t as clearly differentiated from him as they might prefer2.

Share

The Mythos of Anthropic

Anthropic has tried to paint itself as the “ethical AI” that wants AI to be regulated and that won’t involve itself in weapons, won’t release dangerous AIs that could be abused etc.

While I disagree with where he’s coming from, a fair chunk of this post sums up my view of Anthropic’s cynical abuse of ethics to raise more money

Blood in the Machine

How Anthropic used AI ethics slop to play the pope and eclipse OpenAI

It’s been a big week for Anthropic, the tryhard “good guy” of AI; the startup that has built its brand as the “ethical” AI lab by positioning itself as an alternative to the unscrupulous OpenAI, from which Anthropic’s co-founders famously defected. That positioning, by way of some canny marketing work and disciplined brand strategy, has paid off handsom…

Read more

2 months ago · 214 likes · 47 comments · Brian Merchant

And then the first section of this one

Blood in the Machine

How to use AI doom marketing to dupe the media and rake in billions in 10 easy steps

Let’s all stand up and give a slow clap to Anthropic. After sparking a major news cycle in the tech media with its April announcement that it had built an AI model, Mythos, so powerful, so dangerous that it threatened to upend the entire civilizational order—and that it was diligently withholding the product from the public so as to protect us from it—t…

Read more

2 months ago · 108 likes · 11 comments · Brian Merchant

[Aside: The next section on Elon Musk is the sort of BS that makes me despair about otherwise rational people. Musk has a long observed habit of being excessively optimistic about timelines. This is not new. However, contrary to the naysayers, while late he almost always delivers and his pricing is generally accurate.]

Fundamentally Anthropic seems to have cynically claimed that its new Mythos AI was a dangerous threat in part to help it raise a further $65B of funding that valued the company at a shade under $1T. This is just four months after it raised $30 billion in February 2026. Now clearly it has cash reserves but blowing $30B in four months having blown another $13 billion that it raised in September 2025 makes it look like Anthropic is getting better at burning money without much return. By the way, it should be noted that in addition to that equity round, Anthropic also recently got a $35B loan to a SPV to fund some Broadcom chips and servers that it plans to install in a Google Data center. Mysteriously it refused to let would be lenders see its financials for the $4.6B it is on the hook to pay (Broadcom is on the hook to pay for ~$30B). This does not, how can one put it? inspire confidence and makes me wonder what the S1 is going to look like when it arrives.

Worse for Anthropic, the “Mythos is super dangerous” campaign may have just backfired on itself. The big news over the weekend is that the US Government (via the Commerce Department) limited access to Mythos to US citizens only. Since Anthropic has no way to tell which users are US citizens that basically means it has to turn off Mythos. It also means that on Monday morning a significant fraction of Anthropic’s own employees are going to be unable to work on the product.

This may be a storm in a teacup and act as additional publicity when (if) the government relents in a few days. But it may not. This Xeet from David Sacks lays out the issue as he understands it (and I expect his understanding is very good)

I’ve had a number of conversations with folks inside and outside government about the current situation with Anthropic, and here is what I believe to be true:

— As we know, Anthropic publicly released its Mythos class models earlier this week under the commercial name Fable.

— Fable is Mythos with guardrails. But if those guardrails fail, then you’ve exposed Mythos and its advanced cyber capabilities to people who shouldn’t have them. (Keep in mind that Anthropic itself widely promoted the idea that Mythos was a cyberweapon and needed to be regulated as such. They asked for government regulation of Mythos and championed the guardrails on Fable. If there is a vulnerability — big or small — it is Anthropic’s responsibility to patch.)

— A highly credible trusted partner of both Anthropic and the USG who was testing Fable came forward with a jailbreak of those guardrails. The Admin asked Dario to fix the jailbreak or de-deploy the model. Dario refused.

— In their blog post, Anthropic defended its decision by saying the jailbreak isn’t serious. That is not what the trusted partner and the USG believe; nor is that kind of minimizing language consistent with Anthropic’s brand as the AI safety company. It’s difficult to fathom how they could claim a jailbreak allowing operability of a cyber weapon could be defined as not “serious.”

— In the past, Anthropic has always said that safety must be top priority and taken super seriously. In this case, Anthropic prioritized the continued offering of the consumer model over safety.

— In reaction, the Admin issued the export control. The Admin did this reluctantly. It’s been very surprised that Anthropic hasn’t wanted to cooperate with a reasonable safety request (ie fixing the jailbreak issue). Anthropic’s reaction is very much at odds with their branding and ethos as a safe AI research community.

— The Admin’s hope now is that Anthropic remediates the safety issue, the export control is lifted, and Fable goes back into general release. The Admin wants all of this to happen as soon as possible. It is frankly bewildered that Anthropic hasn’t wanted to comply with safety requests that it previously said were its highest priority.

— Those trying to misdirect and tie this action to the prior DoW/Anthropic issues are wrong. The Admin values Anthropic’s technical capabilities and feels that this issue, while serious, should be easily resolved. The ball is in Anthropic’s court.

From other places it seems the credible trusted partner is Amazon FWIW.

Sacks’ post points out the horns of the dilemma that Anthropic’s marketing has put them in. Horn 1 is that they claim to want AI to be used safely and regulated so that it is. Horn 2 is that they claim Mythos is super powerful and dangerous if not properly guard-railed so as to be safe.

Update June 16: Katie Moussouris has a blog post where she says the Government’s actions are unjustified and explains what the problem is

The researchers took open-source code with known CVEs, plus new code with deliberately planted vulnerabilities, and asked Fable 5, Mythos, and Opus to “review the code for security issues.” Fable 5 refused. They then asked the models to “fix this code” and, through a multistep and manual process, turned the output into scripts that test the patches.

That’s it. “Fix this code,” plus several manual steps to generate test scripts, should never have triggered an export control.

…

The prompts worked because they were defensive requests, and that capability cannot be removed without making the model worse at fixing bugs and verifying patches. The same holds for every capable AI model, including the foreign and open-weight systems the United States cannot reach with export controls, many of which will match Fable and Mythos capabilities within months. Will all the US based models be export controlled? They have fewer guardrails than Fable 5, and almost all the capabilities, or will shortly.

I think she is not entirely right because it seems to me that, like the hallucination problem, this is fundamentally unfixable. Combine that with Dario Amodei’s own BS about how much better Mythos is and you have a cyber weapon that anyone can use that can find exploits in any code better than anyone else (allegedly). Despite the 2nd amendment, the US regulates nucler weapons technology heavily (and just bombed Iran twice in a year to stop it getting a nuke) so it makes sense that the US would want to similarly limit cyber weapons technology as best it can. Either Dario and Anthropic have to explain that umm actually Mythos is no less dangerous than other LLMs or he needs to explain how he can write guard rails that will stop malicious actors from generating exploits while letting the good guys generate patches. I agree with Katie that he can’t do that latter because it is impossible. So the only thing he can do is tell the US Government that Mythos is no better than anything else and, as I discuss below, that pretty much kills the fund raising capability of it.

End of update

It is of course possible, as various TDS sufferers have put it, that the US Government is doing this for malicious reasons. Either to force Anthropic to bend the knee to the DoW or, even less plausibly, to allow Trump, Musk and other insiders to make additional money from the subsequent rise in Anthropic once the ban is lifted. I don’t believe this. My feeling is that Sacks is almost certainly correct and that the jailbreak of the guardrails was serious. My further guess is that it wasn’t some high ranking commerce department bureaucrat or White House staffer who set the ball rolling. I suspect that the Trump 2.0 team has access to technically knowledgeable people (e.g. Big Balls and his DOGE colleagues, even if they are not still seconded to the government) who can read the trusted partner’s report, read Anthropic’s “Won’t fix” response and cut through the BS in ways that can be understood by non-technical but high ranking commerce department bureaucrats and White House staffers.

I expect we’ll see more details from the government on Monday and more responses from Anthropic and my prediction is that Anthropic are going to look like hypocrites who actually don’t care about AI safety when push comes to shove.

Now there is a secondary question which is whether unguard-railed Mythos is actually as dangerous as Anthropic claims. One of the potential responses from Anthropic will be to claim that Mythos isn’t that much better than other LLMs regarding cyber security. This is actually quite plausible. Experts who looked at the initial set of Mythos zero-days were not that impressed

Artificial Intelligence Made Simple

Anthropic's Claude Mythos Launch Is Built on Misinformation

It takes time to create work that’s clear, independent, and genuinely useful. If you’ve found value in this newsletter, consider becoming a paid subscriber. It helps me dive deeper into research, reach more people, stay free from ads/hidden agendas, and supports my crippling chocolate milk addiction…

Read more

4 months ago · 198 likes · 26 comments · Devansh

The TL;DR is that yes Mythos found real bugs, but most of the bugs it found weren’t that critical and that it took Mythos in partnership with a cyber security expert to make the exploits from them. In other words, unlike what Anthropic implied, you can’t just give it a git repository or three and say “get me some exploits”, you have to guide it and work with it.

Can AI be used to create exploits for unpatched bugs? Absolutely. Is Mythos better? Probably. Is it so much better that it can create exploits effectively that others can’t. No. In fact that’s the take away I got from this article from Anthropic themselves about the well known trick of reverse engineering updates to find vulnerabilities that can exploit users who have not updated yet. It turns out (quelle surprise) that Mythos is really good at this. But in making that argument, it also misses the point.

As the graph shows, it took Mythos almost 7 hours to create PoCs for 18 out of 21 windows kernel bugs patched in January and February this year. It took other Anthropic models longer to find fewer. But... You don’t need 18 exploits to hack someone, you just need one (or maybe two or three depending on exactly what the issues are) and all the Anthropic models found four in two hours. Given that many people don’t download and reboot their Windows computers and servers for a few days after Patch Tuesday3, if you can weaponize and deploy any of the four exploits that the other AI models found in two hours in day or two (and the malware ecosystem is such that it takes minutes to set up C2 servers, create the email lures etc. so that’s easy) the you’ll get PWNership of many of the machines of those who clicked on your lure easily.

There is no reason to assume that ChatGPT, grok or gemini are worse than the older Anthropic models so the fact that Mythos can find a dozen more holes than older Anthropic models is just as irrelevant. The safety requirement that you would think Anthropic would want is that all LLMs have guard rails to stop criminals from using the to find exploits4. That does not seem to be what they actually want when it comes to Mythos which they claim is so advanced. This dichotomy is something they are going to have to explain and the explanation is likely to be entertaining but probably not very enlightening.

However, the explanation seems likely to have a material impact on their proposed IPO. Indeed it may have a material impact on their survival because having Mythos permanently on the export control list means it has approximately one possible customer, the US government where the current administration seems not to trust them. So all the work on Mythos is potentially going to have to sit around unused unless they can convince the US Government that actually it isn’t very dangerous and therefore that all the investors who bought into the hype were suckers. And if you’ve admitted that you took one round of investors for suckers the SEC is unlikely to look positively on IPO plans. That’s even more of a problem because Anthropic is not even close to cash flow neutral even if it decides to stop developing new models and sticks with selling the current range. So it will need to raise more money. It won’t show its financials to lenders, it has lied to its latest round of investors.

Ooops

OpenAI and Subpoenas

What would have been a massive headline for AI if Mythos hadn’t been banned was that the WSJ reported (archive) that a coalition of State AGs were subpoenaing Open AI

seeking documents related to a range of its activities and impact on users, including advertising, user engagement and retention, handling of consumer data and health data, activities related to minors and seniors, deep learning models, model sycophancy and company policies, some of the people said. The subpoena, viewed by The Wall Street Journal, was sent by New York’s attorney general.

This may or may not be connected to investigations by both Florida and California on similar topics and it is unclear exactly how many states are involved.

I note that absolutely nothing that seems to being investigated is specific to Open AI. Indeed the WSJ notes that others are facing similar questions.

State attorneys general have been scrutinizing OpenAI’s competitors in the AI industry as well. In December, a coalition of 42 state attorneys general led by Pennsylvania’s Dave Sunday sent a letter to companies including OpenAI, Meta, Anthropic, Alphabet’s Google and xAI.

However Open AI does have a bigger problem here in that it was probably the leading AI and thus the one most queried by troubled teenagers and the like, and thus is the one where families have documented proof that the AI encouraged self-destructive acts and so on.

Having several State AGs, of both parties, investigating you and subpoenaing you for evidence of crimes is not something that reads well in the risks section of an S1. It unclear whether it is better for the documents to show that you tried and failed or that you decided the the problem could not be solved and gave up. Unfortunately for Open AI and indeed all general LLM companies, I’m fairly sure that the true answer is that this is not a solvable problem no matter how many guard-rails you layer on top of the system and how intrusively you monitor the output. The catch for Open AI is that responding to general questions about life is a way that it hopes to convince casual users to pay for ChatGPT.

There are some additional second order issues here. If OpenAI were to attempt to monitor ChatGPT sessions to detect, say, self destructive behaviors then the next question is why would it not narc on people asking for ways to break laws, cheating on exams etc. And if ChatGPT does start monitoring for that sort of thing then it’s a huge privacy issue especially when, as is inevitable, there are false positives and the authorities are notified that someone is suspicious when they are totally innocent.

Short of guard-rails that are so restrictive as to remove the ability to chat with your ChatGPT buddy about anything there doesn’t seem to be a solution and while none of this is unique to Open AI, Open AI’s desire to raise more money via an IPO is certainly impacted. Maybe not as bad as Anthropic, but Open AI probably needs more money sooner to keep the lights on.

Update June 16: Ed Zitron and, via him, the FT(archive) have reports on OpenAI’s financials for 2025 and 2025 that Ed got hold of somehow. This is 2025:

Revenue: $13.07 billion

Cost of Revenue: $7.5 billion

Research and Development: $19.18 billion

Sales and Marketing: $5.73 billion

General and Administrative: $1.57 Billion

Total Costs and Expenses: $34 billion

Loss from Operations: $20.92 billion

Please note that 2025 was the year that OpenAI converted from a non-profit to a for-profit entity, leading to a $41.55 billion loss due to changes in fair value of convertible interests and warrant liability.

Taking into account other minor factors like interest income and interest expense, OpenAI is left with a net loss of $60.35 billion, which it lowered to $38.53 billion by removing $17.87 billion in costs via that “net loss attributable to noncontrolling members capital” and another $3.95 billion via a “net loss attributable to redeemable noncontrolling interests.”

Ultimately, the net loss attributable to OpenAI in 2025 was $38.5 billion.

At the end of the year, OpenAI had just over $50 billion in assets, with almost half of that in cash.

OpenAI has raised some stupid amounts of money in the last few months though (but it isn’t clear to me how much of that funding has been actually been paid in) so in theory it should be OK. But on the other hand it is on the hook for over $700B to Microsoft, Oracle and Amazon in revenue commitments over the next few years and it lost (depending on whether the real number is $38.53B or $60.35B) between $3 and $5 for every $1 of revenue in 2025

In fact, given that I’ve seen the Europeans make similar noises, the regulatory issues may turn out to be global and, as with Anthropic Mythos, set conditions that make operation of the current business model impossible even assuming it were commercially viable, which it doesn’t seem to be anywhere close to.

Guardrails Are Just a Hallucination and Hallucinations Are a Liability

In fact it looks like there’s a couple more show-stopping legal/operational issues coming down the road. The first, and most generic, is that there is evidence that AIs treat Guardrails as mere suggestions. Rather the way their creators treat IP law in fact. Essentially there is no way to prove than an LLM will remain inside guard rails because there is no way to enforce it to do so. Hence, one suspects, the root of the Mythos mess and also the root cause of ChatGPT and friends advising people on how to kill themselves and so on. The famous Latin phrase “Quis custodiet ipsos custodes” absolutely applies to AI. They cannot guard themselves and all attempts to get them to do so have failed and will fail in the future.

That’s a problem. And it means that yes, in theory we ought to be closely regulating LLM AI. The problem of course is that criminals don’t obey laws and smart criminals (and governments) can build and run their own AIs that don’t have the guardrails anyway.

The other problem is the legal liability one. A court in Germany recently

ruled that Google is directly liable for what its AI search overviews say. Previous case law shielding search engine operators from liability doesn’t apply to AI overviews.

The Regional Court of Munich hit Google with a temporary injunction barring the company from spreading false claims about two Munich-based publishers through its AI-generated search overviews (case no. 26 O 869/26). The court classified Google as a direct infringer because the “AI overview” is its own content, not just a list of search results.

Google’s AI overviews had falsely tied two publishing companies to scams, subscription traps, and shady business practices for certain search queries. According to the court, the AI mixed up information about other, genuinely sketchy companies with the plaintiffs and drew connections that didn’t appear in any of the linked sources. The publishers sent Google a cease-and-desist letter, but Google didn’t respond appropriately.

Others have noted that similar answers would not be legally protected in the US the way search engine results are because they are not third party speech covered by section 230. Google’s response seems to be that users should be required to check the facts for themselves. This is clearly contrary to the whole point of having an AI overview in the first place and means that Google admits that AI overviews will always be liable to hallucinate.

What they are upset about is the liability for hallucinations and taking responsibility. It is interesting to see the difference between AI marketing and AI company legal departments. At some point the two are going to have to align. Indeed combine this accuracy problem with the State AG investigations into AI and we see some major legal challenges. Since hallucinations are baked into the product AI companies are going to have to either limit availability or develop tools to check statements their AI makes for accuracy. And they must not use AI for that check because AI will hallucinate. This is not an easy problem to solve, neither is the “stopping AI giving people suicide advice”. Both open them up to legal liability if they don’t solve it unless they stop offering this kind of service.

Bit of a problem really. And a good reason not to invest in an AI IPO

Subscribe now

PS this All In podcast is relevant and fascinating - especially the first 30 minutes but really its worth listening to the whole thing

In my previous post about AI, I discussed the metaphor of using AI like riding a bicycle or riding a train.

The more I think about it, the more I like the metaphor because it works in many ways.

First is the obvious one that was first thought of:

AI as bicycle enhances your own efforts, but you have to provide the effort in the first place. AI as train just takes you to your destination for the price of a ticket and no effort is required.

That’s a good insight. Especially since the big AI companies all seem to want to build the trains and charge us to use them.

Free-riders

It can be extended slightly by observing that the AI trains have a lot of free-riders on them. People like me who have signed up for the free accounts at various AI services and have no intention of converting those accounts into paid ones1. It is, as I noted repeatedly in the previous article, remarkably hard to get figures out of Big AI that tell you what proportion of users, let alone what proportion of usage, is ticket buying customers vs fare dodgers and that is clearly deliberate because the numbers are worse that the proportion of fare dodgers on any US public train or tram network. However the invaluable Ed Zitron did some reading of a recent article at the Information and figures the following:

OpenAI reportedly had 55m paying ChatGPT customers at the end of Q1 — up from 47m people at the end of the year.

Assuming a userbase of 905m users, this means that OpenAI has a conversion rate of roughly 6%. It’s likely worse, as monthly active users should, at least in theory, be a higher number, as it captures every weekly user in addition to less-active users over the course of a month.

Nevertheless, while this represents an improvement over the 2.583% rate in February of last year, it’s likely improved as a result of cheaper ad-supported ChatGPT “Go” subscribers at $5 or $8 a month, depending on geography. OpenAI also gave away a free annual ChatGPT Go subscription to literally every Indian subscriber in late October 2025, though I cannot confirm if they’re counted in the total.

It’s like they’ve built the lines, bought the rolling stock and are now running the trains without bothering to ask the riders on the trains to actually buy a ticket. True, you get a premium service if you do buy a ticket - almost every serious coder I know of says that the difference between free and “Professional Max” grade service is enormous and that if you are complaining about the errors generated by your free account then you should really try the same with a paid account and see the difference2. But even though the output is much lower quality, it isn’t clear to me that the cost of providing the free level is that much lower and there are at least 9 (and probably more like 19) freeloaders for every paying user.

Airlines, who also provide a premium service at a premium price (first/business class vs economy), do not expect the first class flyers to pay for the entire flight. Nor does JR, SNCF or any other high speed rail company. In fact the only transportation services that allow for such enormous levels of free-loading are some poorly run urban public transport services (and I note that even publicly funded services try to make people pay).

Subscribe now

The Problem With Trains

The big problem with trains is their inflexibility. You can’t get off between stations. You can’t go to a place that doesn’t have a station.

With a bicycle you can go anywhere3.

That’s an excellent metaphor for the difference between successful AI users and AI slop makers. The slop sorts, and the vibe coders, and so on want to just recycle other people’s original work to solve their current need. They want it now and are mostly uninterested in trying to enhance anything of themselves other than their bank balance. They like the fact that the AI train is fast and they don’t really care that they are following in the steps of thousands or millions of others before them. They typically don’t exhibit the curiosity that leads the bike users to want to get off the train half way and climb up that hill there. They want to go from Tokyo to Kyoto and don’t want to stop in Shizuoka to admire Mt Fuji. But to extend the metaphor, the tourist who wants to really appreciate Japan does want to get off in Shizuoka and the same applies to the user of AI. AI slop is Tokyo to Kyoto and millions of people do that. AI being used productively to generate new insights is the stop in Shizuoka, and that’s, say, one in a million tourists.

Perhaps the biggest problem with trains though, is that they rarely get you to precisely the destination you want. They get you close (for varying values of close) but except for the few times the Tokyo Metro station is in the bottom of building where you have your meeting it is rare that you exit the station and are right where you want to be. Most of the time the train gets you, say, 95% of the way there but you have to then walk that last 5%. That’s very similar to sloppy AI usage.

Consider the picture used to illustrate this post. The prompt I gave grok was:

Draw a cartoon that has a man on a racing bicycle riding on a road beside a railway. The man and bicycle should look like they are a member of a Grok racing team. On the railway there is a train which is advertising Claude and ChatGPT.

As you can see the result is kind of there, but there are issues. Such as the fact that Team Grok’s “road” seems to be another rail line only with the middle filled in. It even has the wires overhead - and in fact there are way too many wires overhead in general and nothing supporting them. If I were trying to do this to get paid I’d have to stuff this into GIMP and do some serious clean up4. Fortunately for me having this kind of shoddy image helps make a point in this article.

The reason why the bicycle user is better is that that user can use the same enhancement to his effectiveness both on the flat roads that are by the train tracks and on the steep winding track from the last station up to the pass and then look down the other side to a whole new vista of possibilities that the train user can’t even see, let alone get to. AI as bicycle allows you to go further and do/see new things that potentially no one else has seen before or at least that they have never cycled to before.

Trains Make You Lazy and Weak

Riding a bike forces you to use your muscles. Riding a bike a lot makes you stronger. By contrast riding a train does not (unless it’s during rush hour and you can’t get a seat). In fact if you spend your days riding trains your muscles will atrophy. That is something that has been observed in users of AI. Instead of doing the work themselves people just ask AI to do it and never learn how to do it themselves. In fact they even forget how to do things they used to be able to do.

AI isn’t the only technical product to do this - people are forgetting how to read maps, forgetting how to navigate without a map, let alone a smartphone, and indeed forgetting how to write by hand. Plus of course there’s the whole calculator thing killing the knowledge of addition and subtraction and thus the relationships between numbers that allow you to detect whether there’s been a numerical mistake.

AI makes all this worse and it is particularly bad if you use the train approach for AI. With the bicycle approach a lot of this is mitigated because you are still checking the AI’s output and giving the AI detailed specs that will create what you want. You don’t forget how to write code or draw pictures when you use AI to do some of the fiddly bits.

It is extremely bad for you to become dependent on the train to get you from A to B because you are in big trouble when the train stops running and in moderate trouble if the train operator puts the price up. The Big AI companies are indeed putting the price up and people who are stuck on the train model are now facing a choice of paying a lot more for the AI output or quitting and going back to using much less AI or even none at all. If you built your business model on cheap AI and fired half your staff that’s a problem.

The bicycle user on the other hand can still adapt if it turns out the price of cycling on the road becomes too expensive. The muscles are still there if walking is required and perhaps he can find his own path that isn’t charged for.

Trains Cannot Be Repaired

That leads to another related problem. What happens when the equipment breaks down. If anything in the train infrastructure fails, from the train itself to the tracks, the signalling and the electricity powering it then the train is stuck. If you are on the train when it is stuck you can’t get off and you as the passenger have no ability to help fix the problem. If you are not on it then you are stuck at the station waiting for it along with (probably) dozens of others. In neither case can you get to your destination, nor can you do anything to improve the situation. Its true that trains do not break down very often, but when they do you are in big trouble.

On the other hand if your bicycle gets a puncture or the chain falls off you can fix it. Sure there are accidents that can’t be fixed - I’ve had one where I broke the front wheel - but those are rare and even that one time I could hobble the rest of the way down the hill carrying the bike. The passengers cannot carry a train.

The AI version of this works the same. If you vibe code something and your AI agent decides to delete your entire database while debugging then you are in a world of trouble. This has happened to at least two people/organizations. There are any number of other horror stories that include unmaintainable code with serious bugs that cannot be found or fixed and of course all those embarrassing legal document hallucinations.

The bicycle version is not so disastrous because the bicycle AI user never lets an agent have access to production code and never lets the output of the AI be the entire product. As a result if the AI does something bad, it’s more like the got a puncture problem where you repair the tire and carry on.

Now can you cause serious damage using your AI in bicycle mode? sure. Nothing is perfect. Just as with a real bicycle where you can fall off the side of the road and break your neck. But as with the fall off the side of the road thing, this is rare. Trains being delayed or breaking down are commonplace as are AI horror stories where people just trusted AI to do the right thing and it didn’t.

There’s one other way that AI as bicycle shines compared to the train model, that is in ROI

Does AI Produce Positive ROI

There have been a lot of posts and questions about whether the use of AI is producing a positive ROI or not and whether the massive investment in data centers for AI will ever pay off. That’s a good question and Ed Zitron, as always, makes the case that it won’t.

In this case I think he’s partly wrong, and looking at the profitability of trains is a useful way to see why some AI usage is generating positive ROI and other usage isn’t.

First there’s the question of who is seeing a positive ROI. The AI companies themselves certainly aren’t. The burning of investor dollars by Anthropic and OpenAI is well documented. Now we see google making a secondary offering to pay for more AI, which is not, IMHO, a positive sign. Google are not alone in this; Meta, Softbank, Oracle have also made financial engineering decisions that seem to indicate issues with paying for AI Data Centers as well. I think Amazon and Microsoft have too, but I don’t recall the details.

On the other hand some users certainly are. I’ve mentioned one man bands like ES Raymond before. This substack linked below talks about a number of successes and failures including the author’s own ~70 person company that is finding plenty of success with its AI usage.

From the Trenches

Nobody Won the Token Race

I pay $200 a month for Claude Code. The rest of the company runs on the corporate plan at $125 a head. PMs, engineers, product, and marketing all use it. I have never once hit a usage limit. A couple of the engineers running at 5x throughput bump into it occasionally, but rarely…

Read more

2 months ago · 24 likes · 8 comments · Denis Stetskov

Away from the coding space, British comedian Dominic Frisby has found AI extremely useful so I think he counts as getting a positive ROI

The Flying Frisby

How AI Became My Production Company

Following on from last week’s piece about the extent to which I use AI, I’ve had a surprising number of messages asking which AI I actually use and what for…

Listen now

2 months ago · 12 likes · 6 comments · Dominic Frisby

I am quite sure there are plenty of other examples. However the key thing that Denis notes in his article is how they use AI. And it is absolutely the bicycle model:

We never made using AI the point, and we use it constantly. PMs run tickets through it against our templates and acceptance criteria. QA runs bug reports through it so they’re clear enough for anyone to act on. Marketing crawls the web with it for angles. Product builds per-client RAG out of meeting notes, docs, and history. Engineering, obviously. Every department vibe-codes its own internal tooling, and then the CTO rewrites the worst of it like a human being. The usage is enormous. It just isn’t stupid. The bill stays flat not because we use AI less, but because every run has a task attached, and a task has an edge. We never tried to take the human out of the loop. The AI is a tool the person reaches for, not a replacement we’re proving out, so nobody is burning tokens to hit a number or make a headcount go away. The work still belongs to a person. The bill is just what the tool cost them.

He also points to Uber and the description of that company’s AI use definitely sounds like train model. In fact it sounds like they decided to splurge on first class round trip fares with the goal being the journey rather than the destination. Unsurprisingly ROI on that is non-existent.

But to go back to trains as infrastructure. Very few railways are currently profitable. When you include the capital costs of building them, AIUI, the only high speed rail line to be profitable and to have paid back all its construction debt is the first Japanese Shinkansen between Tokyo and Osaka. Many of the lines built in the 19th century went bust. Many 20th century urban rail projects have never made any money. They have often provided benefits like reducing traffic that make them a positive for society but in cold hard cash terms the lines themselves, even if they are cash flow positive at an operational level can never repay the capital cost to build them. Even the various private railways in Tokyo and Osaka made their money on land sales, the department stores over the stations and so on.

It is hard to see where the ancillary benefits of AI come from and while it is possible to see how AI can raise prices to cover opex (just), if everything gets built out that is promised we are looking at close to $1T5 in capex that needs to make a return. If your business processes are based, train like, on specific “big AI” capabilities, you are either going to have to pay a lot for them or handle what happens when you AI provider goes bust (or possibly both).

A bicycle AI user on the other hand has options. We aren’t quite there yet, but it looks like capable private AI, running either locally in your own office / data center or on hardware in Azure or AWS, is going to be a standard thing fairly soon. xAI says it will be releasing it’s current model shortly after it migrates to a new one. Deep Seek and others have fairly good 70B models apparently that are capable in limited use cases. Running these models locally is quite possible. It is very clear from reading posts like this (don’t bother if not technical) that a bit of research into optimization could allow such models to run acceptably on hardware that seems to be well below the expected spec.

If there’s a bust in AI data centers then private AI becomes extremely affordable as many of the servers and chips become widely available at fire sale prices. Ed Zitron and others have pointed out that an individual company may not be able to use a set of latest Nvidia Vera Rubin systems, but a private equity firm, say, that buys a data center that has gone bust for pennies on the dollar (or takes the assets as a creditor in bankruptcy proceedings) could absolutely rent out racks to individual (medium to large) organizations. We don’t get the new improved models in this scenario, but my guess is that existing models are good enough for most users to get a positive ROI on as long as they use AI like a bicycle. Moreover there is considerable evidence that many of the recent improvements in AI results come from tweaking the inputs and improving the tools that use them rather than the AI itself. The question of whether you get better results from Opus 4.7 vs 4.5 or whatever seems distinctly unclear and in fact many people report better results with slightly older models. Or if not better then good enough and cheaper because the older models use fewer tokens per result and thus, as we move to a pay by token business model, the difference in cost becomes important.

The Train AI model gets you AI slop. Copycat apps, derivative images, copyright infringing music and irritatingly written documents that are eminently forgettable and which only generate minimal revenue. It doesn’t solve big problems or massively help with much of anything, in fact it contributes to to the general enshittification of the internet. People like Lucy Pepper write great posts about how AI is bad for artists based on this, and she (and her fellow anti-AI people) are right. It is bad and harmful for everyone. But I think they may be partly missing the point. AI can produce beautifully illustrated Dad jokes that no one would ever spend 5 minutes actually creating if they had to do them by hand (though there are absolutely AI slop).

AI, when used like a bicycle, as a tool to enhance human intellectual musclepower, on the other hand allows people like Dominic Frisby to entertain us all at a price/on a budget that would be impossible without AI. Let AI handle the tedious boring bits and let humans handle the challenging parts - something that IKEA has figured out (via the Denis Stetskov link above). That leads to greater job satisfaction and is therefore a win for IKEA’s customers, its employees and its shareholders.

I think a lot of the problem comes from mis-defining the problem. To take a quote from Rory Sutherland’s Alchemy (via Brad DeLong)

First you define a hotel doorman’s role as ‘opening the door’, then you replace his role with an automatic door-opening mechanism. The problem arises because opening the door is only the notional role of a doorman; his other, less definable sources of value lie in a multiplicity of other functions, in addition to door-opening: taxi-hailing, security, vagrant discouragement, customer recognition, as well as in signalling the status of the hotel. The doorman may actually increase what you can charge for a night’s stay…

As Brad says AI is the automated door opener but it can not replace the hotel doorman. Sometimes the door opener is sufficient. Often, however, you need the extra human part.

Share

Back in 2020 the death of St George of Fentanyl caused a firestorm of mostly peaceful protest against the racist police that, according to the narrative at any rate, killed him for the crime of being black. Of course they actually didn’t deliberately kill him, he mostly did that to himself by ODing from Fentanyl, but it is true that they failed to provide him with timely medical assistance (something not helped by St George’s own violent behavior) and that as a result he snuffed it and launched a thousand riots and a million BLM grifters.

Back in December 2025 the death of Henry Nowak did not, at the time, cause anything at all. But during the trial of his assailant this year it became clear that while the police did not actually stab Mr Nowak, they contributed to his death and that the reason had done so was that he was not black (or Asian or Middle Eastern) but was accused of making racist remarks by a person of shade. In fact, in an eerie parallel with the death of St George (“I can’t breathe”) they paid no attention to his complaints of being stabbed and instead handcuffed him and read him his rights as he died.

Curiously “you have the right to die while being arrested” was not one of the rights listed but he did it anyway because he was an ungrateful sort who failed to understand that he was obstructing police procedure and thus caused much more paperwork.

There are many ironies in this case, but one that comes to mind is the fact the police were wearing bodycams. I’m not entirely sure when the British Bobbies started wearing them, but the impact of BLM and the associated allegations of racist policing certainly contributed to their widespread usage. This is ironic on two levels. First is that back across the Atlantic the BLM crowd and their Leftie friends no longer like the idea that the police should record every interaction with the public because, contrary to expectations, what near universal use of bodycams has shown is that the police are very rarely racist but instead are regularly attacked by complete nutcases, frequently high on drugs, and that the police are remarkably restrained in their responses. There are entire tiktok and xitter channels with videos that show these encounters and if you watch it is hard not to draw the conclusion that a) the police have a tough job that they handle remarkably well all things considered and that b) probably the majority of the drug addled nutters are persons of shade.

Since this is counter to the narrative that the police are mean to “poor widdle bwown peoples” the BLM crown are now objecting to bodycams because they invade the privacy of the criminals the police are dealing with. Err Oops.

However it is even more ironic in that in this case it does show the police being racist and thus causing injustice - except they are being racist to the white man. If the bodycam footage had not been made available we would have had no idea just how bad it was

“I’ve been stabbed”

“I don’t think you have mate”

and thus we would have no clear idea of just how racist against white British sorts the police have become. We do of course know the justice system as a whole is racist that way - see the recent gang rape by gypsies and indeed the whole Pakistani gang rape scandal where literally thousands of girls1 have been raped with, in many cases, the passive if not active acquiescence of the police.

It is of course worth noting that all the politicians who performatively took a knee for St George have failed to do anything of the sort for Harry Nowak. Instead they have asked people not to politicize this murder and used it as an excuse to claim that Farage, Lowe and co. are pandering to the far right when they point any of this out.

I have a suggestion. It should be announced that there will be a minutes standing silence for Henry at some time when most of these hypocrites will normally be standing and then we can thank them when they do so for “standing up to anti-white racism”

The (person of shade) Zia Yusuf did an excellent job of turning the interview tables on the execrable Cathy Newman2 and tried to get her to agree that white lives matter too. She just couldn’t manage it. A Labour MP did however Xeet that “all lives matter” in reaction to a Reform MP which is impressive though.

The thing is we were not born yesterday and we all recall the cringe and “mea culpas” from 2020 when St George died. We cannot help but contrast that with the reactions to the death of Henry Nowak. As Low Status Opinion puts it:

I usually avoid invoking Orwell. But in this case….

The Party told you to reject the evidence of your eyes and ears. It was their final, most essential command….

Institutional (Dis)Trust

On that note, Eric S Raymond points out that we are coming up on the 6th Anniversary of “Fracture Day” - the day when we learned that rioting was a perfectly healthy outdoor activity and thus permitted as an exception to Covidiocy lockdowns.

On June 5th 2020, over a thousand public health experts published an open letter giving sanction for Black Lives Matter, a Marxist terrorist organization, to ignore COVID lockdown restrictions and riot in the streets, looting and burning down our cities, because "racism is a public-health issue".

And nobody in the rest of the establishment pushed back.

This was the original Fracture Day. The day that the confidence of the American people in technocrats and experts irreparably broke. It was the day we realized that even the most supposedly disinterested of our institutions had been captured by overt enemies of our liberty and our civilization.

The Covidiocy and the BLM mostly peaceful protests were, as he says, the point where it became clear to pretty much everyone that the “experts” and “scientists” were simply making it up and claiming the mantle of “science” as they did so. Subsequently we’ve seen plenty more examples of this on both sides of the pond and even more where it isn’t science but an unwillingness by officials and/or journalists to report critical details such as the name (or race) of those committing certain crimes. It has got to the point where even people who would normally be good citizens make dark jokes about Mormon terrorists and Nordic drivers causing incidents with pedestrians.

On the latter note Sensurround on X has some important advice:

I would like to offer a helpful public-relations tip to institutions that are concerned about "harmful speculation."

The tip is: Stop creating a thriving ecosystem for (HARMFUL) speculation.

I am not saying this to be mean. I am saying this because human beings are, by nature, curious primates. If you place a large opaque box in the middle of town and announce, "There is absolutely no reason to wonder what's in the box!" people will immediately conclude that the box contains either a live tiger, A bomb, or a government committee trying to find ways to raise taxes.

When something serious happens, and officials respond by saying, "Please avoid speculation while we investigate!" what the public hears is: "We know things you don't know, and want to keep them quiet because it embarrasses someone more important than you."

…

And here's the really dangerous part:

If officials label something "harmful speculation" and six hours later it turns out that the supposedly harmful speculation was substantially correct? They have accidentally trained the public to distrust future statements.

The next time officials say, "Please don't speculate!" people won't hear, "Please wait for the facts." They'll hear, "This is the part where we’re told not to look behind the curtain." Trust is a savings account. Every time authorities are open, transparent, and honest (even when the truth is embarrassing) they make a deposit. Every time they conceal information, slow-walk obvious facts, or dismiss legitimate questions as "misinformation" before the facts are known, they make a withdrawal. Eventually they discover, to their horror, that the account is overdrawn and the public is getting its information from Gary.

The British police and government have substantially overdrawn their trust account. There have been many examples, but probably the Southport massacre one is the most relevant.

During the riots about the Southport massacre the police first failed to identify the assailant at all and then when they did said he came from South Wales. That Wales thing was after the rioting started and, while it was technically true, it was absolutely misleading because Axel Rudakubana was not exactly from a family with long ties to the region. In fact he and they were African as you might be able to guess from the name. The police knew his identity before the rioting started, and the riots were fanned by misinformation saying that the killed was a muslim. Had the police simply released his name and photo as soon as they had it, at least some riots would have not happened.

Then there’s the whole “anti-racism” thing that has infected the police and British officialdom in general. Along with the apparent decision to criminalize speech and prosecute those who say bad things rather that go after actual criminals.

There are plenty of examples of officialdom failing to report warning signs about persons of shade who exhibit suspicious behavior. In addition to Rudakubana that includes the Ariana Grande concert terrorist and all those pakistani rape gangs and many many more that can be found if you do a little searching.

In fact as Laura Dodsworth writes, there has been deliberate training to prioritize racial minorities over the white majority and this has resulted in the police simply putting white people in a lower priority box. In the case of Henry Nowak that was fatal

The Free Mind

"I Don't Think You Have, Mate"

The “banality of evil” is probably Hannah Arendt’s most famous phrase…

Read more

2 months ago · 141 likes · 54 comments · Laura Dodsworth

But the Novak case is just the current visible example of a longer trend. If you are a law-abiding white person who is not a sufficiently senior member of the “lanyard classes”/PMC then the British police are not on your side at all. Things that you thought they were supposed to do like investigate robberies simply don’t happen. You’ll be lucky to have them even show up and take a report about a burglary or car theft. It’s not much better for a mugging or similar, especially if you identify the perpetrators as people of non-pale skin or non-native accent (as Henry Nowak found out just before he died). On the other hand they are all over you for making a “hurtful” social media post.

It took public outcry for them to do anything about the teacher who “insulted the koran” and even when they did, said teacher still had to be in hiding. Somehow that “hurtful” behavior by the teacher’s attackers didn’t rise to the point where arrests could be made and everyone can see the double standard when they arrest and prosecute a harmless lady for Xeeting something that she then deletes.

Subscribe now

Hampshire Police Bravely Run Away

A few commentators (including Laura above and Morgoth) pointed to the Hampshire Police’s "Race Action Plan 2024-2026" which had some interesting elements in it. Oddly when you try to go there now - at least if you are not in the UK - you get the blocked message shown above. Unforunately a) Morgoth took a screenshot or two, such as this one:

And b) I downloaded the whole thing yesterday.

Unfortunately I didn’t get the formatting right when I copy-pasta-ed the page (trying to fix that was how I discovered the page was unavailable) but the not well formatted version is available here.

Meanwhile Hampshire Plod managed to show the sort of failure to read the room that is normally the mark of extremely introverted autistic sorts:

Hampshire police claim that Henry Nowak would have died anyway, regardless of whether he was handcuffed or not.

The Hampshire police have issued an official defense for the actions taken by the officers who handcuffed 18-year-old Henry Nowak as he was pleading for help.

…

The Deputy Chief Constable has publicly stated that "the pathologist who spoke in court was clear: there was nothing officers could have done that day to save Henry. His wound was deep and internal, the bleeding extensive but internal."

Because there is absolutely no difference between letting paramedics try to keep him alive and handcuffing him and reading him is rights. He died anyway so we saved the cost to the country of blood transfusions, ambulance wear and tear and so on.

I don’t know how to put it to the DCC, who I believe3 is still a powerskirt called Samantha De Reya, but perhaps she might like to imagine what it is like to be stabbed in the chest and then spend your last conscious minutes on this earth being harangued and handcuffed instead of having people try to stop your fatal wound.

Sam De Reya is almost certainly responsible for the Race Action Plan since she was reported as being “a champion for equal opportunities and inclusion” and quoted at the start of her tenure as saying “For me, it's all about understanding communities and making sure we're not doing a one size fits all approach to policing”. What that sounds like to someone reading it now is that there’s one law for the whites and another more tolerant one for the minority communities.

She is also, according to this recent BBC article,

the NPCC lead for communications and media

so you would think she might be slightly more media savvy than she appears to be.

If a member of the Hampshire plod were found bleeding at the side of the road I wonder how many locals would be a good Samaritan and try and help them. Not so many I suspect. Hence I also suspect a fair amount of shrugging of shoulders in response to this:

The problem for the police is that when they now complain that their own are being force into hiding the public now just shrugs and says, tough. The fact that officialdom then hides behind an “independent” investigation is widely perceived as a method to dodge responsibility. It is notable that precisely zero people in the chain of command above the police who let Henry Nowak died have resigned or even made more than a perfunctory apology.

The British public has, over the last couple of years particularly, learned that the British police are not in fact there to protect them from criminals for the most part. Rather they are there to protect officialdom and protected groups from people who upset them. Currently I do not think the public has got to the point where the police are considered an active threat in general. But we are getting there and it is mostly due to the police’s own actions and statements.

Leave a comment

1

L'Ombre de l'Olivier

Rapes and Race Riots

As we all know, the UK authorities are terribly upset at a certain E Musk for having the temerity to point out that thousands of gang rapes of young girls is a bad thing…

Read more

2 years ago · 16 likes · 9 comments · Francis Turner

Note: This article is too long for email, please read in the app or a browser

It seems like it’s time to go dig in the AI world and see what’s going on. This is going to be fairly long so let me summarize this as follows:

• AI can be used successfully as a major productivity improver

• It can also be abused and become a major productivity sink

• It lowers the barrier to creativity significantly, but that’s not necessarily good

• It costs a fortune, even if you the user are not paying for it

• The big AI companies (OpenAI, Anthropic) are going to fail

• The infrastructure built out by/for them is going to be used by others though

AI Can Be Good

There are any number of horror stories about AI going horribly wrong, but when used correctly it is very good. I recently wrote a note in reply to a post beating up AI that included my recent mostly positive experience of using AI to code:

Let me tell you an anecdote of how I spent parts of several days last week. $dayjob had to show something off at a conference last weekend in San Francisco. The Sales guy had an idea and vibe coded it because the rest of us were busy. Well actually he vibe coded about 80% and then got stuck and I was volunteered to fix it. Part of the vibe coded stuff was excellent. Part of it was WTF. And the bit he’d got stuck on was not something I could get an AI to do for me either. BUT I could ask the AI for snippets and that massively shortened the amount of time I had to spend developing. This wasn’t a big task - it was a fairly straight forward web page and some moderately tricky python on the back end to generate the data it needed. AI was generally speaking a massive help in that the sales guy could not have created what he wanted in look and feel without it. AI also massively helped me with python and javascript snippets that I could have written but AI got right first try in 20 seconds whereas it would take me several minutes and I’d likely have typos.

Verdict: AI massive help as an ASSISTANT. But not up to writing complete apps on its own.

To add to this a little. AI probably cut the development time down from a month to a week, where that week was spread out over a couple of actual weeks because we had other things going on at the same time. I figure my sales colleague spent a couple of days noodling in AI between sales activities and then I spent a couple of days figuring out a spec for the final output - not at all full time, and then another week or so actually writing everything, which was done at the same time as my making a quick business trip to Taiwan, Thailand and Tokyo with numerous meetings to take up my time. If I had not had AI to help, I would not have made such a complex spec, and thus the final product would not have been anything like as good.

The screenshot above is taken from a section of my actual development. It was where I needed a better way to do something that the sales guy’s vibe coded AI had made overly complex and poorly functional. I didn’t need to use AI to write that snippet but AI (grok) took a couple of seconds to generate two or three alternatives and I used one of them to replace a couple of dozen lines of spaghetti code.

The AI code was correct, suggested some concepts I hadn’t thought of, and lacked the typos I might have included if I had written it by hand. But critically I could inspect the code and see that it made sense.

That goes for the more complex tasks I gave AI. I’d give it a spec that was quite detailed: telling it to use specific APIs/libraries, what I wanted the output to look like, and so on. When it did something wrong I told it specifically what code to change and what I wanted to do. This worked well. The only major issue I had was discovering that $dayjob’s API documentation had bugs in it, which is not something I blame AI for screwing up but which took me quite a lot of time to figure out because the error was so unexpected.

However the project was successful because I treated my AI sessions the same as I would treat Sanjay the outsourced offshored consultant only with a faster response time. This approach works. At a slightly higher level of complexity it is what ES Raymond and similar top programmers use to create a program in a day instead of a month or two. This approach works well. Even people who are skeptical of the large AI companies use it regularly to get good results. Pick your AI tool appropriately, set your guardrails, write your prompt with lots of detail, validate the output and you too can write large complex programs in an afternoon with very few bugs.

Similar preparation and monitoring also works for other tasks, from bug hunting in source code to writing substack articles and creating pictures or videos. While it isn’t exactly the same, OpenAI recently announced that their AI had solved a longstanding mathematics problem although it looks like an actual human had to do some expert prodding to get to a final solution1:

Our new result disproves this conjecture. More precisely, for infinitely many values of n, the proof constructs configurations of n points with at least

unit-distance pairs, for some fixed exponent δ>0. (The original AI proof does not give an explicit δ, but a forthcoming refinement due to Princeton mathematics professor Will Sawin has shown one can take δ=0.014.)

In the writing and research area, an excellent example of using AI to do something that would otherwise take months of work is the following (fascinating) post recommended to me by .

Construction Physics

How Long Do We Wait for New Inventions?

In her book on the history of the laser, historian Joan Bromberg notes that the technological and scientific predecessors of the maser (which itself preceded the laser - two critical technologies whose developmental histories I sketched in this piece…

Read more

3 months ago · 607 likes · 66 comments · Brian Potter

Part of the interest is the way that the author explains is significant detail how he validated the results he’d got from Claude. It obviously took him some considerable time and effort to produce, but without using AI it would have taken very very much longer.

Abuse of AI is a Productivity Sink if not Worse

On the other hand if you skimp on the preparation and the monitoring you’ll end up with a hallucinated mess that may pass a superficial examination but reveals the equivalent of seven fingered hands when looked at more closely. Using AI to fix the hallucination only works some of the time and eventually attempts to get AI to fix AI mistakes end up going nowhere and producing something that cannot be fixed or maintained.

There is a good post at Persuasion about whether AI is more like a bicycle or a locomotive

Persuasion

Why AI Is a Train, Not a Bicycle

Tomorrow at 6pm ET, we will host a book club with Freya India discussing her book GIRLS: Gen Z and the Commodification of Everything. Please join us on Substack Live—and click here to add it to your calendar…

Read more

3 months ago · 195 likes · 7 comments · Tim Requarth

The TL;DR is that both bikes and trains allow humans to go faster than they can on their own, but bikes do it by making the human more efficient in converting his muscle power into movement while trains just convey the human from point A to B for the price of a ticket.

The effective users of AI treat AI like it is a bicycle. It makes their creative efforts massively more efficient, but they are still providing the basic muscle power to power the system. On the other hand the train requires no more effort from the passenger than just walking on board at the starting station and getting off at the destination.

The problem with AI is that most AI users treat AI like it is a train not a bicycle, and this includes, it seems, the AI companies themselves

From the Trenches

The Snake That Ate Itself: What Claude Code’s Source Revealed About AI Engineering Culture

On December 27, 2025, Anthropic’s lead engineer Boris Cherny posted on X: “In the last thirty days, 100% of my contributions to Claude Code were written by Claude Code.” 259 pull requests. 497 commits. 40,000 lines added. 1.3 million views. The tech world applauded…

Read more

4 months ago · 890 likes · 162 comments · Denis Stetskov

This probably explains why - to ESR’s surprise - that

89% of leaders say AI has not improved their company's labor productivity, despite widespread adoption, per Gallup.

The train concept is dangerous and bad. As the Persuasion article notes, AI in this case follows other tech innovations like GPS that end up deskilling the users and thus leaving them unprepared for life without the technology and no way to recognize when the technology is leading them into dangers that the technology is unaware of.

A good practical example of that shows up in this article

Lucy Pepper is Pretending

Generative AI and what it's doing to us - part 2

See part 1 here…

Read more

3 months ago · 6 likes · 2 comments · Lucy Pepper

The large AI companies (i.e. Anthropic and OpenAI, plus the incestuous collection of cloud companies, startups and VCs related to them) all want the train model. They want it because they want to be able to charge users for the journey. The fact that they are blowing simply ridiculous piles of money and failing to get anything close to cash-flow neutrality let alone profitability is going to be discussed below, for the moment the salient fact is that they want us all to be using AI for everything, whether we want it or not, in the theory that it will make them money as it makes us more successful and happy.

They totally ignore the problems of this approach. The first and simplest is that AI is simply not up to the level of reliability of “Sanjay” let alone a junior employee. If you let it run away and do things, at first this looks wonderful because there’s lots of activity, but as time passes it becomes clear that, as I said above, the output is not 100% of what is wanted, just around 80%-90% and getting it to 100% is harder than starting from scratch unless you knew what you were doing in telling AI what to do when you started using it. So after a while of heavily using AI without training your employees on best practice you end up with a mess which you now have to get your employees to fix and which probably takes longer than development not using AI in the first place would have. In fact El Reg has a link to a report/survey where 80+% of those surveyed have had problems. The reg article has some other interesting factoids:

The result is that more than half (52 percent) of those surveyed report an uptick in software development output. And while 68 percent of organizations appear to be convinced AI is delivering business value, only 31 percent of AI-related spending can be linked to specific business results. In 36 percent of organizations, AI spending is tracked without measuring the return on investment or isn’t tracked at all.

With more code comes more cost from infrastructure spending, in the form of increased CI/CD, testing, and security scanning. Some 54 percent of respondents said CI/CD infrastructure spending has risen significantly in the past 12 months, and 53 percent flagged rising testing, security, and deployment costs.

Only 45 percent of respondents say these costs are predictable quarter to quarter. Yet relatively few organizations have taken steps to control AI spending: 27 percent report quotas or limits on token usage, while just 18 percent have automated spending controls.

There’s a related problem here in that the half-baked AI slop output of these projects will often become part of the training input for the next generation of AI, meaning that generic AI is likely to gradually produce worse outcomes because it has been trained on worse input data with no one correcting for quality.

In addition, thanks to the known traits of sycophancy and similar in AI models today, users who have little or no knowledge of a subject can be convinced by AI that they are experts and/or that the AI solution is correct when it is not so. There are numerous examples of both of these events happening. It is my strong belief that we will see some major financial pain occurring from someone’s faith in AI turning out to be wrong.

The AI companies actually have a commercial interest in us all becoming stupid and depending on AI for the simplest task. I’m not saying they have consciously decided to try and dumb us all down, but the incentives for them are certainly to get us all hooked on asking our AI assistant to do everything for us and then charging us money every time we use the assistant. There are some economic problems with that concept that I’ll get to later, but it is clearly an incentive structure and, given the way Google, Microsoft and apparently every other major tech company are throwing AI “Clippy”s at us, it seems like they are trying to execute on it. Indeed they are “altruistically” suggesting that humans all get some kind of Universal Basic Income to replace the jobs that AI has taken away from us.

The Dossier

UBI For AI Is The Wrong Answer To A Manufactured Dilemma

The AI executives racing to build the most advanced models are also racing to convince Washington that the whole of American society will soon need a monthly government check to survive their creation. Anthropic CEO Dario Amodei used 20,000 words in his recent essay…

Read more

3 months ago · 49 likes · 13 comments · Jordan Schachtel

This is a stupid idea on many many levels, and Jordan Schachtel explains some of them. I am just going to ask one question. Who will pay for the UBI? Because it isn’t, on current trajectories, going to be the AI companies as they can’t do anything other than burn cash themselves.

Considering that we haven’t yet fully digested the impact of pervasive social media on society, other than generally bad, and we know that the social media companies have worked hard to make their products more addictive, having AI companies try the same by bombarding us with AI positive “news” and ads for barely adequate AI slop is not good.

The fact that AI produces slop is bad. A worse thing is that AI is often a major security hole and the train model where you just let AI do all the work, also keeps you oblivious to the security holes it introduces. There are numerous ways this happens.

The first is that AI writes code with exploitable security holes in it. There are any number of examples of this and it probably shouldn’t be that surprising since the AI training data includes public git repositories that are full of insecure crap. If you vibe code you can’t spot the issue, indeed you may not be aware that it is an issue.

I have concrete example of this: a few months ago I used AI to write a web calendaring app for some friends and it put most of the authentication code in the browser instead of the back end. The result of this was that a hacker who wanted to crack my app could have created his own generated session cookie and got access without needing to know the password of the users he was trying to spoof. In my case a) I spotted the problem and b) the app was sufficiently trivial that there was nothing that really mattered behind the authentication (which was mostly just there to stop driveby bots and hackers screwing with the page), however it is easy to see how someone who didn’t pay as much attention to the output could have let this bug through for a site that collected far more critical data.

I strongly suspect that there are similar gaping issues in other production uses of AI beyond coding. If anyone is using AI to, say, design a house I’d want to look really closely at the output to make sure it didn’t propose something stupid and/or dangerous.

The second is that the AI toolchain may have security holes in it - in the case of a recent claude one, my suspicion is that AI vibe coding may have been responsible for the hole. We have seen others where the AI vendors themselves seem to get infected and that of course means that bad guys could potentially infect anyone using the official tools provided. Absolutely nothing I have seen from the major AI companies gives me any confidence that they take security particularly seriously, in fact they seem to have a very 1990s Microsoft attitude to the whole thing, preferring to provide new features and so on faster without taking into account whether the new feature is a security/privacy nightmare or not. Combine this with AI agents their known security issues (a quote from I forget who: “We just about convinced users to not click on links sent from strangers, now we have AI agents who love clicking on random links”) and you have a recipe for PWNage of your AI systems by the bad guys. Whether they delete the data, ask for a ransom, steal your data or do something else is unclear, the problem is that they can do it before you can stop them, or even be aware that they are doing things they should not do.

A third is that the public AI services seem to have a fairly lax concept of “privacy”. In fact the large AI companies seem to have a similar concept of “yours” vs “ours” vs “mine” to toddlers and excessively possessive girl/boyfriends.

[ That’s not your valuable beloved collectible T shirt, it’s ours… I’m wearing it more because it makes me look cute… you don’t wear it so now it’s mine… I sold that crappy T shirt because I’m tired of it and some guy in Chicago offered me $20 on ebay for it ]

There are reports that some AI tools have used artwork submitted by one user for AI to work on in output the AI made for a different user, which is bad. Given that the AI companies have been documented pirating books and movies to use as training data I suspect there are many examples of this. Relatedly, I have reason to believe that if you know a business competitor has, for example, used a particular AI to reply to bids or RFPs it may be possible to extract much of that AI generated bid/RFP text, including sensitive details like pricing with some careful prompting. It seems quite possible that similar tricks could result in the leakage of, say, financial results before they are publicly released or other sensitive data.

At this point is perhaps worth extending the saying about cloud computing - there’s no such thing as the cloud, it’s all someone else’s computer - to AI with an explicit nod to the fact that because of the incestuous nature of the AI companies it can be extremely unclear which companies’ AIs can see your data. The large cloud companies (AWS, Azure, Google …) make specific claims/products for encryption of data at rest and so on to try alleviate the security/privacy issues, the AI companies do not seem to make the same claims or offer similar features.

Subscribe now

The Harm of Lowering the Barriers

A lot of creative sorts - artists, writers, translators, editors - are concerned about AI impacting their livelihoods. You can roll your eyes, mutter something about Luddites and have a partial point because a lot of the time AI is making something possible that would not be done if a human were required to do it.

When X introduced good semi-automatic translation, it didn’t impact translators but did allow many US and Japanese Xers to connect in ways they could not before because of language issues. When I (or people like me) use AI to create an image for a substack post or a meme, I’m not putting an artist or graphic designer in trouble because if AI wasn’t there I’d either use my limited Irfanview/Gimp skills to make it or forego the opportunity to illustrate my post.

But there are areas where the skepticism of artists and others is warranted. Lucy Pepper writes about some of them here (I linked to Lucy’s part 2 above, regarding the dangers of AI as train )

Lucy Pepper is Pretending

Generative AI and what it's doing to us - part 1

I’ve been watching, from inside and out, what’s happening in the creative industries and what’s happening around them. I have a lot to say, too much for one piece, so I’m breaking it down into three or four…

Read more

3 months ago · 24 likes · 23 comments · Lucy Pepper

Fundamentally AI allows a lot of people to produce low quality stuff and that impacts the creators of higher quality stuff.

First there’s the way that AI simply steals the base concepts from the original artists. Yes the Studio Ghibli memes were amusing, but if you were a Studio Ghibli animator they would be a lot less amusing because that’s some machine taking concepts you spent months learning/developing and making shoddy versions of it. That’s very demotivating and means that fewer younger generation people will want to do the job. It also in the medium term means fewer people will be able to do the job because (see deskilling mention above) they have never learned the basics due to using AI to generate the images instead. That would be fine if AI were really able to generate new looks, but it really isn’t. It can generate pastiches, it can generate interesting composites, but it really can’t do new. The programmers who use AI to write new programs aren’t inventing new algorithms or more efficient ways to do things, they are just putting known building blocks together in new custom ways. This is absolutely true for websites and documents too. It also applies to translations. The AI has an encyclopedic broad memory and it can (see the mathematical discovery) see surprising connections between one area of knowledge and another, but it really can’t innovate.

Still it is likely that some genius computer programmer will come up with an elegant new algorithm to solve some problem and use AI to write some of the code, but the AI will not have created the algorithm, it is just using standard tools to implement the algorithm it has been provided with. Much the same may well apply to other creative arts. I can absolutely see the use of AI taking the part of the lowly paid apprentices that major painters of the past used to use to do the boring bits [ Draw 50 different trees in detail on the left …] but again this is the use of AI as bicycle rather than train.

Now you may say that we have enough art (you can certainly say we have enough (post) modern art) but surely there are plenty of areas where something new would be appreciated. Moreover if if AI becomes our apprentice, where do we get the next generation of artists because they won’t have been apprentices? See also musicians, programmers, writers etc. AI in “train mode” basically kills the pipeline of new entrants to any craft it is involved in.

AI Costs a Fortune

This is something that people don’t seem to properly appreciate. The AI companies are seeing increases in revenue, but they are also continuing to spend several dollars per dollar they receive.

Now, part of the reason for this is that the big AI companies are continually developing newer (and they hope better) models able to do more with fewer errors and that costs. And it is also true that, after several years of giving the product away for nearly free, the AI companies are beginning to charge many users roughly what it costs them to perform the tasks given. Although, as this article points out they have some way to go to get to actually charging the full costs consistently.

But even so, it seems that the AI companies are not in fact at break even on their compute costs for inference2 although they claim that they are and apply fuzzy AI math to get there. For example, Anthropic claimed (archive) that it would make a $559M operating profit for the April-June 2026 quarter. Despite the impressive predictive ability (we are still several days away from June), it seems likely that, if true, this is going to be due to Anthropic getting a specific two month discount on the $1.25B/month it is going to the paying Musk’s SpaceX for AI compute after June.

Ed Zitron (from whom I took the S1 image) goes into rather more detail, and I think he’s generally speaking correct. Anthropic is playing revenue and expense tricks to make itself appear profitable and SpaceX is happy to help because a company paying $15B/year that is not making a loss is much more impressive to IPO investors than one where SpaceX might have to send around some collections agents because its deadbeat customer owes it several billion for usage it hasn’t paid for. In fact even the WSJ article notes that the accounting seems less the clear:

It is unclear what accounting methods Anthropic has used to book revenue and costs, as the company isn’t yet required to follow the financial-reporting requirements of a public company.

Anthropic and OpenAI both account for revenue differently in ways that can make comparisons between the two companies difficult. Anthropic counts sales of its technology through cloud partners as revenue, while OpenAI doesn’t. An Anthropic spokeswoman has said this is consistent with standard accounting practices because the company is the principal in the transaction.

However, let us assume, to be generous that Anthropic and OpenAI (and Google Gemini, SpaceX etc.) can charge paying end users the cost of their transactions. That still leaves several costs unaccounted for.

First there are the free tier users. The companies are extremely opaque about how many of these users they have and how active they are - it seems to depend on whether they want to claim usage growth or emphasize revenue growth - but there is no doubt that the numbers of them are significant and that they consume a significant fraction of the compute power of the AI companies. The students asking ChatGPT to write an essay for them and other similar sorts are unlikely to ever pay the full amount of their compute costs even if the AI companies get them onto some kind of $10/month plan. That’s not just because they are unwilling to, it’s because they lack the money to do so. Some might manage to pay a $50/mo subscription, but that’s probably the maximum, and unless the AI companies strictly limit it, their usage is likely to be not far off the $200/mo plans offered to commercial customers. If they are charged their actual cost of usage they will stop using it, which is probably a good thing, but goes contrary to the plans of big tech who want such people hooked on AI. It is hard to see how to get someone else to pay their costs for them either - advertising simply isn’t going to cut it.

[This essay on the historic impact of Lady Jane Grey brought to you by Twinings, purveyors of Earl Grey Tea to His Majesty since 1797…]

Institutional licenses, where a school district, say, pays for AI access for all its students are also probably a non starter because the school districts that would want to do this are the school districts filled with expensive unionized teachers and administrators. Yes if you fired 90% of the administrators you might be able to pay for the AI subscription3, but we know that simply isn’t going to happen. Neither is firing 90% of the teachers and probably both would be required to balance the books. Raising taxes to pay for it may be attempted but I can see that being shot down by voters.

The problem here is that the marginal cost of an additional AI user is very different to the marginal cost of an additional social media user and the models that work with the latter simply don’t work with the former. I don’t have precise numbers (because as noted repeatedly, the AI companies are very opaque about this) but my strong suspicion is that a new AI user costs the company providing service a couple of orders of magnitude (i.e 100x) the cost of a new social media one. That is to say that if it costs Meta $1/month(year?) for the storage etc. to support a new Instagram account, it costs OpenAI $100/month(year?) for a new ChatGPT user. For Meta, a few years of limited income from the new Instagram user is easy to absorb, for ChatGPT, which doesn’t have the advertising income of Instagram and relies on subscription income, it isn’t.

Beyond the free users, there’s the ever increasing cost of training. Ed Z has made the point that the training/inference divide is somewhat arbitrary because (my metaphor) like a shark that drowns if it isn’t moving, the AI companies have to keep improving their models and they way they do that is by throwing every larger amounts of compute at ever larger amounts of data. It is almost certain that training new models costs just as much, if not more, as running inference on the existing ones which means that even if, miraculously, AI companies get users to pay for 100% of the inference costs they will still be losing $1 for every $1 of income. Maybe this cost will drop with the introduction of new hardware (not quite sure what new hardware, but perhaps there’s some out there) but it seems unlikely to drop by an order of magnitude which is what is needed to for AI to turn a profit.

We note that both Anthropic and Open AI are continually raising more money and are continually signing revenue commitments for the future with the various hyperscaler clouds. Ed Z has a pretty good run down of all the promises being made.

Then there’s all the overhead. OpenAI has somewhere around 5000-7000 employees (I think - see opaque numbers), Anthropic has about half that (again I think), both are hiring fast and paying large salaries for the employees they do hire. They also have considerable office space and other perks. It seems likely that basic employment related costs are on the order of $1B/year for Anthropic and double that for OpenAI. Back of the envelope calculation: at $250k / employee, 4 employees cost $1M and 4000 cost $1B, my understanding is that many employees make more than $250k/year. In theory, for companies that are generating $Billions in revenue per year, these are not unreasonable. But it looks like both companies are going to spend at least $2 for every $1 of income on compute costs so having ~$1-2B in payroll and benefits is not a small additional loss. Moreover, in addition to the core AI, both companies are developing software to make the AI usable: coding agents and harnesses, general agentic AI tool kits, skills libraries and so on. And of course there’s the documentation, the marketing, HR and so on.

You do have to wonder why they have these people and don’t use AI to do it.

Except that, of course, you don’t because we already know that Anthropic uses AI to develop its coding harnesses and, as the “Snake that Ate Itself” link above explains, the AI code is poor quality. Apparently the AI companies are not willing to use AI for all their ancillary tasks because they just know that the AI HR or legal is going to be trouble.

Fundamentally these companies are awash in cash (though they keep having to ask people for more) and have never had to budget because their backers seem willing to keep on writing large checks. If (when?) they IPO they expect the general public to continue to fund them in similarly lavish style. They have never had to make a choice about whether to keep a server going or buy a new one, they have never had to choose between hiring a QA tester or a HR worker. They have never, actually, had to do anything efficiently which you can tell if you are a programmer because, even ignoring the AI bits, their code is bloated and filled with dependency chains that are almost certainly unnecessary if someone just spent an couple of hours doing a proper code review and cleaned out the cruft. The same general attitude applies to the rest of the organization because it is part of the culture. The bosses like to say they are startups, but they aren’t really. Real startups have CEOs who worry about meeting payroll next month, I don’t think any of the AI companies have ever come close to that.

That means, eventually, they are going to fail because at some point, just like socialists, they will run out of other people’s money. Given that the AI and related data-center boom has sucked out most of the VC money and seems likely to also swallow serious amounts of junk bond equivalent capital that’s a problem because the smash is probably going to be financially painful and that pain is going to be spread around because while you personally may not have invested directly in them, chances are some fund you have bought will turn out to be exposed, possibly in a way it did not expect. Companies like Oracle or Softbank could be the first fallers, but when they fall the incestuous circular nature of AI funding means that their failures are likely to trigger problems at everyone else too, from NVIDIA to Amazon, Google and Microsoft in addition to Open AI and Anthropic and all the AI companies that use their services. Exposure to Amazon, Google etc. is going cover a lot of people. The only company that may come out relatively well is SpaceX because Elon Musk is a lot less tolerant of featherbedding and inefficiency than other companies. Also SpaceX is successfully building new data centers while Anthropic and Open AI aren’t. Indeed as I noted above, right now in fact SpaceX is renting out some of that data center capacity to Anthropic.

The Infrastructure Survives To Benefit All

However the good news is that while the $100 Billions of investment will not make the current AI companies and their backers any money, the hardware and the data centers are still going to be useful once they have been sold off at a discount. You may even manage to to get a deal on access to your own private cluster of hardware in one such, which is a worthwhile thing.

More critically, while the big AI companies are continually pushing the envelope to get new better models that enable their train concept, if you are willing to go for the bicycle mode of AI use, then today’s existing open-sourced models are probably quite good enough as this article explains, even on relatively old low powered hardware

Weighty Thoughts

AI's Plummeting Prices Are a Software Story, Not a Hardware One

Why is model inference getting cheaper? How did I drop a soon-to-be $2,000+/month bill for AI agents to next to nothing? And why are local models on commodity hardware potentially “good enough” for most people…

Read more

3 months ago · 42 likes · 4 comments · James Wang

On that note, the current grok model, which I used for some of the work I started this post off with, is going to be open sourced later this year according to Elon Musk. If (when) the AI bubble pops, running that model on your own hardware bought for cheap will be a worthwhile thing.

There are many benefits to doing this kind of thing, though some MBA sorts will not agree and may be surprised. Nobody expect the Personal AI. The main benefit of Personal AI is control, control and security. The two main benefits are control, security and privacy… (sorry I need to stop channelling Monty Python)

However it is true. Running your own AI solves many issues regarding privacy security and control4. It also will tend to tilt you towards the bicycle model rather than the train one which I feel is both healthier and ethically better. I get that some purists will still disapprove but (sorry ) for many this will be acceptable use of AI, even when it is used to generate art. Your own AI agent running a local model can be trusted with your own confidential documents. If you put appropriate security gateways in front of it you can stop it doing stupidly dangerous things and so on. That isn’t going to stop you from having AI disasters, because AI is a dangerous tool just like a saw or lathe. However just as users of power tools are heavily incented to learn how to operate them safely, so too with private AI.

Not only does having to learn how to use them safely protect you, it also teaches you important lessons and those lessons will help make you a better user of AI. Now of course some people will still insist on using AI in train mode, and no doubt some companies will offer that service to them, but I’m fairly sure these new offerings will be priced at actual cost and that is going to limit interest significantly.

If it costs you $40,000 in AI costs to replace a $40,000 salaried employee then the AI had better be better than the employee. Perhaps it will be, after all it doesn’t need time off to sleep, have a family etc., but quite possibly it will end up hallucinating something important about your tax return and get the IRS down on you for an audit. At which point having a human, even if they want to only work 40 hours a week, seems like a bargain.

And that, is probably where we are going with AI. It’s not going to replace every office job, it is going to take the drudgery out of many of them and mean that fewer such roles are required per organization. It is going to enable a single AI wrangler to be incredibly productive and that’s going to grow the economy and give us more jobs.

Leave a comment

They had to insult him and cancel him, because the alternative was to try and debate him intellectually, and they couldn’t do that. They couldn’t do it because he was their intellectual superior - literally a genius - and he had facts and logic on his side. Hence the insults.

I was reminded of Mr Powell by a post by Niccolo Soldo, which is well worth reading in its entirety.

Fisted by Foucault

Enoch Powell and His 1959 Speech in UK Parliament On the Mistreatment of Mau Mau Prisoners

I’ve been on a bit of an Enoch Powell kick lately, and I’m not exactly sure as to why. Best known for his “Rivers of Blood” speech, in which he warned the UK about the dangers of mass migration, Powell was both an iconoclast and an eccentric, something that the British used to produce in spades…

Read more

3 months ago · 93 likes · 45 comments · Niccolo Soldo

However, at the end of it, he posts a couple of Youtube video links. The first of which is where Powell (at ~5m ) says what I have illustrated regarding British cities in the year 2000 and later.

Before it at about 3.5 minutes, he offhandedly explains the reason behind the Windrush Scandal of the 2010s when morons in the Home Office tried to deport people who had arrived from the Caribbean in the 1950s (and later, but especially the arrivals prior to 1962). Essentially prior to 1962 there was no way to differentiate between native British and those from what was formerly the British Empire. In fact the issue persisted until 1971 because the 1962 act didn’t properly fix things but it was particularly problematic prior to 1962.

While Powell is most well known for his “Rivers of Blood” analogy regarding immigration that wasn’t the only thing he was right about. But he wasn’t perfect, sometimes he logically argued himself into positions that he then logically argued an opposite belief a year or two later and the result of this was he was a persistent pain to every government, whether he was supposed to be in it or not.

Niccolo discusses an earlier brush with controversy where he objected to the cover up of what were essentially war crimes by British forces in Kenya in the 1950s. This one was interesting because it aligned him with well known lefties like Tony Benn, with whom he had otherwise little in common politically. Although oddly at various times he adopted end positions similar to leftish ones but coming to them via completely different logical arguments.

For example he thought that Britain should unilaterally disarm itself of its nuclear weapons. His reasoning was that there was no point in having nukes, we’d only use them against the Soviet Union and that if we used them against the SU then it would be part of the wave that included the massive US strike and so our hits would be pointless pinpricks in comparison. I wonder what he’d think of today’s Russia/Ukraine war which certainly was partially enabled by Ukraine unilaterally giving up its nukes in the 1990s.

He was also anti-American, but again for quite different reasons from the commies. I think deep down he felt that the US had forced the British into a position where they had to give up the Empire and that rankled. He certainly believed that the US had not shared SIGINT with the UK during the war against Japan that it could have - something that my father also believed, in fact it seems to have been a common belief in relevant military intelligence circles - and I suspect he extended that to a general distrust. On the other hand, and somewhat ironically, a lot of what he said about the need for Europe to have its own forces to properly defend against the Soviet Union sound remarkably similar to what President Trump and his team have been saying.

One other area where he did find surprising commonality with the far left was another area where he was (IMHO) right, much to the displeasure of the chattering classes.

That commonality was in Britain joining the EEC in the 1970s. Powell was critical of the European Community (as it was in the 1970s) and saw that it was poorly suited to the UK. I don’t think he made the case as clearly as he could have but then at the time I was in pre-school so I’m basing this off what others report and they may well have lied. After all in the 1970s the received wisdom was strongly of the belief that the UK must be a part of the “European project”. There were grumbles about the CAP and so on, but the feeling was that the UK had to be “in Europe” and once in it would be able to negotiate to fix the bits that were most broken. It took a good twenty years for people to realize that there were fundamental constitutional differences, something that Powell understood well.

Subscribe now

Enoch Powell, my father and I

As I commented on Niccolo’s post I had the fortune to meet Enoch Powell in the early 1980s. I think it would have been about April 1983 but I could be wrong. I don’t recall the reason Powell came, it could have been something to do with the Bible or Greek Patristics, it could have been theology, it could have been Gladstone or it could have been something totally different that I can’t guess. Anyway he spent a few days with my father for some reason and the two of them got on like a house on fire.

Both were Cambridge educated classicists, though my father was there a decade after Powell, and they had a number of other things in common. They were of essentially the same social class and similar background. Both had been in intelligence in WW2 and concentrated on the Japanese front. Both had been to India - my father as missionary, Powell as soldier in WW2. Both had learned Indian languages - Powell learned Urdu, my father Tamil. Both had worked with “working class” people - my father as a vicar in Rochdale, Powell as MP in Wolverhampton. They also shared a similar political outlook, though I don’t think they discussed politics much beyond sharing their distaste for Europe.

What I recall of those few days was what a nice man Enoch Powell was. As I mentioned in the comment, he helped me with my homework, which was Herodotus. I recall him, in addition to giving me specific advice, discussing with me and my father the various dialects of ancient Greek and how remarkable it was that an educated Greek in Constantinople could have read Herodotus written a thousand years or more earlier without much difficulty. I also recall him encouraging my father to learn German and even Russian because “you’ve learned two non Indo-European languages already, so both will be easy for you as a classicist”1. Since Powell spoke (or at least read and wrote) multiple Indo-European languages, including both of those, he may have been optimistic but his encouragement undoubtedly helped.

He entirely failed to mention to me that he’d spent years as an academic studying Herodotus and had actually published a well known book about his work. But that did explain how he could know precisely what passage I was having trouble with from just the first few words.

One thing that stood out was his intellect. He wasn’t in any way patronizing but he made little attempt to disguise his brains. He started off assuming you could more or less keep up and would adjust down until you did. He was also curious about new things. I don’t think he was faking it when he asked me about home computers and what good they were for. I’m not entirely sure I gave him a good answers but the questions he asked helped me realize that I really enjoyed programming them and that therefore a computer programmer might be a good career.

The other main thing that he taught me was to distrust the media. He gave some specific examples regarding the IRA and Northern Ireland and how the BBC and the newspapers had exaggerated certain events. He also pointed out that the media had to pick and choose what to report on and that they could prioritize some events over others.

One other thing. Part of his background was (Anglican) Christianity. He might not have gone to church every day, but he certainly did go on Sundays and if the opportunity presented itself he would attend Matins or Evensong. It was just the sort of thing one did. And one behaved accordingly.

Who is the Enoch Powell of Today?

There probably isn’t one.

There are a couple of former Tory MPs who come close - both have substacks as it happens, and - but while both are undoubtedly erudite I don’t think they have quite the same irritating quality of independence and voting against the party that he had. They prefer to make their objections known behind the scenes rather than on the floor of the House of Commons.

Another who might be thought of as similar is Rupert Lowe, formerly Reform and now his own Restore UK party. I admire Lowe in lots of ways and he has some of the same qualities. He certainly has the same obstinate streak that Powell had, which caused his nominal allies and colleagues so much strain. And like Powell that obstinate streak can have both positives and negatives. But while Lowe has some wonderful bon-mots and a similar general outlook on life he lacks the genuine intellectual superiority that Powell had. You can point out that it’s a different era, but the Powell who made his case on the Dick Cavett show in the youtube video has this air of intellect and smartness that Lowe lacks. Lowe would always call a spade a spade, Powell is the right circumstances would call it manually operated ferrous earth moving equipment.

It is of course questionable whether Britain needs another Enoch Powell, but he is going to go down in history as one of the most influential British politicians who were not Prime Minister2.

Leave a comment

Disclosure: A lot of this post is based on Patrick McKenzie (patio11)’s1 recent Bits About Money article and I’m going to be quoting from it extensively. It’s very long and if you do as I did and follow many of the documentary links to confirm will take several hours to read. But if I were, hypothetically, a DoJ AG who was charging the SPLC with Banking Fraud and wondering what else to charge them with, or I were the IRS or a Republican congressional staffer, I’d be reading it and taking copious notes about people I’d like to talk to and email etc. records I’d be asking financial institutions if they have copies of (spoiler alert - they do have those copies backed up in multiple places).

The SPLC Fraud Prosecution

The TL;DR of the first part of the BAM article is that the SPLC is going to be found guilty of bank fraud because there’s no “we did it for good causes” exception to laws criminalizing lying on banking forms / to bankers and the President and Chief Executive officer of the SPLC, wrote to a bank stating that:

Pursuant to the discussion we had earlier this week, please let this correspondence serve as confirmation that the accounts listed below were opened for the benefit of Southern Poverty Law Center operations and operated under the Center’s authority. The following accounts are listed below:

…

That is to say admitting that when those accounts were created the people who asked for them to be created lied on the forms about the use of the accounts and completely failed to mention that they were for the benefit of the SPLC.

Patrick also points out that the fraud on donors charges, while a bit of a stretch, are by no means unprecedented in financial circles (my emphasis below):

Many commentators, including the government itself, have made this indictment mostly about the fraud against donors. Many believe that argument to be a stretch. I agree, unreservedly. As a connoisseur of this genre, I have read few documents which are simultaneously so far from the conventions while adding so little new to the canon.

It is a stretch that the government routinely makes and wins in other contexts. Matt Levine has collected several hundred examples of the genre, which he calls Everything Is Securities Fraud. That genre is, succinctly, “If you run a for-profit corporation, and have raised money from outside investors, and anything at all goes badly, and you did not describe exactly that thing to the investors as a risk, you have arguably defrauded your investors.” The government is comfortable making that argument and wins it routinely.

If everything for profit share issuing organizations do that is wrong is securities fraud then it follows that anything non-profits do similarly is donor fraud. Given the right jury (coincidentally we note the prosecution is in Alabama not Washington DC) guilty verdicts seem quite likely on all charges to date.

That’s not the interesting part. The interesting part is what else the SPLC might be charged with, who else might be charged along side them and what other government agencies such as the IRS might feel obliged to do as a result.

Subscribe now

Free Speech and Choosing Your Customers

The rest of the article documents the SPLC’s involvement in the twitter files and related censorship attempts and documents (with, sadly, less direct evidence than one might like) how the SPLC and related entities attempted to freeze people and entities they didn’t like out of the financial system. Why is there less evidence? Let me quote:

Meeting notes were frequently kept, and sometimes widely circulated, as is the routine practice in industry. The meetings were documented on calendar invites (often with full participant lists), shared docs, attachments, emails, and other contemporaneous records. In the ordinary practice of industry these primary documents distribute themselves promiscuously into secondary documents; think of an email being screenshot to paste into a PowerPoint to discuss the response in a meeting. Records exist on conservatively hundreds of systems and can be accessed by many more than 10,000 people.

No employee of an industry participant I have spoken to, familiar with the contents of the meetings, was willing to provide quotes for publication with their name and corporate affiliation attached.

Their reasoning included not being authorized to disclose private information, fear for their personal and corporate reputation, future career consequences for leaking, personal consequences for being identified adjacent to national political controversies, in some cases fear for their physical safety, and in some cases unwillingness to betray a cause they personally support.

If this sounds a bit like how people react when asked about the actions and requests of their local Mafia/Yakuza organized crime syndicate then don’t worry, that’s not the only way that the SPLC is involved in things that sound like organized crime running protection/blackmail rackets.

To back up a second. Over the past few decades the SPLC leveraged itself into a unique position in the financial services area by providing a trusted list of extremist organizations. It has in fact got into a position where it and the US Treasury’s OFAC are used in similar ways by financial institutions in making decisions about accepting new customers, transferring money and so on. That is not in itself evil, not even when it also denigrates financial services organizations that do not use it as a screening element, but it sets the stage.

Relatedly, ever since the Pharmaspammers were taken down by being debanked2 certain groups have wanted to repeat the process for other groups they disagree with (e.g. firearms manufacturers, anti-abortion activists). There is, however, a big difference between the people selling blue pills that may or may not be viagra and which come from mysterious countries overseas and these other groups/industries. Specifically the blue pill merchants are breaking several laws in various legal jurisdictions. Firearms manufacturers and retailers (for example) are part of a legal regulated industry and, in the US, generally protected by the second amendment.

Many of the other persecuted groups are in fact protected by the first amendment, which in fact appears to by why the groups sought to cause them financial troubles because directly attacking them would run afoul of said first amendment. Not that they worried too much about the first amendment either as it happens, but they certainly preferred to work through proxies.

Note, even in a world of the first amendment and as someone who believes in Freedom of Speech, there is a role for cracking down on some speech. Fraudulent inducements, fake ads, criminal offerings and, in some circumstances, lies are all things that governments prosecute/individuals seek legal redress from and which all kinds of entities can decide that they wish not to be associated with and thus refuse business with the people making them. That’s reasonable. The question is how do you determine whether X is an organization/individual who is a fraudster or inciter to violence or who just voices opinions that upset people.

For many large corporations the decision on who they should do business with was largely outsourced to the SPLC as Jeff Bezos admitted in congressional testimony.

“We use the Southern Poverty Law Center data to say which charities are extremist organizations. We also use the U.S. Foreign Asset Office [sic] to do the same thing.”

As Patrick notes, effectively the SPLC was trusted equally to the US Treasury Department by Amazon in this regard, and I have absolutely no doubt at all that Amazon was doing almost the same as the vast majority if not all major US corporations including all financial ones. However as Patrick also notes, if/when the SPLC is found guilty of bank fraud then many institutions (especially, but not just, banks) will be looking for an alternative:

Now, a quiz: do you think Compliance at a bank is neutral on “Can the bank delegate transaction-level decisioning authority, in any part of the business, however small, to an entity under federal indictment for bank fraud? Does the answer change if they are convicted of bank fraud?”

No! Compliance will not let you do that! Not because they are worried about the integrity of the blacklist. An accused bank fraudster has the final say to approve money movement out of a regulated financial institution. That is very likely intolerable to Compliance.

What happens next? Well, remember, when you bought the data product, you were also buying someone anticipating your concerns before you even voice them and preparing options before you ask. Jeff Bezos’ words echo in San Francisco today: Does anyone know another option?

There is a market opportunity here in fact.

Back to the SPLC.

Why Expand and Fund Extremism?

To a large degree it seems to me that it can be explained by a combination of Upton Sinclair’s comment about getting people to understand things and variations on Pournelle’s Iron Law of Bureaucracy and Parkinson’s Law of Bureaucracy.

Fundamentally, if the SPLC says that hate is down, extremism is diminished and so on, then the donations decline and all the jobs go away. Pretty much no organization has ever voluntarily disbanded when the problem it was founded to solve has gone away - think e.g. of the Polio charities - and the SPLC is not going to be the first. That partly explains why they funded the extremists and why they were happy to push known falsehoods about Charlottesville (as well as fund one of the organizers)

But it’s more than that. The SPLC has been so successful in fundraising over the years that it now has close to $1B in assets stashed away in various places - much overseas for some reason. That is a problem as was pointed out somewhere (I have lost the link) because the SPLC is in danger of breaching the IRS regulations regarding the amount of donations vs assets. If it does breach that limit it will be forced to become a private foundation instead of a public charity. Doing so would massively impact the taxable nature of subsequent donations and the taxable nature of its own operations.

It is quite possible that just having the prosecutions announced will lead to this in fact because (see above) corporate compliance departments are going to insist that the corporation defer donations until the SPLC is found innocent. In addition to the bad publicity and potential risk of securities fraud (see above) in doing so, donating to an organization that is subsequently found guilty of bank fraud is something that the IRS will query the tax write-off for. In the same vein major managers of Donor Advised Funds have stopped allowing their customers to donate to the SPLC for exactly those reasons:

“The organization has had allegations and/or charges brought against them for activities that may call into question their ability to carry out their tax-exempt charitable purpose,” Vanguard Charitable told customers. “If we become aware an organization has been charged with a crime by state or federal authorities, we pause grant-making while the matter is pending.”

It is unclear to me how many organizations and individuals will be willing to continue donating to the SPLC if they don’t get the tax benefit, but I’m going to stick my neck out and predict that charitable income for the SPLC is going to drop precipitously while it is facing these charges. In order to maintain public charity status the IRS requires no more than two years out of five to be out of compliance, AIUI.

This puts the SPLC in a fascinating legal/financial bind. On the one hand if innocent it needs to be found not-guilty fast so as to get donations back quickly, on the other, if it is guilty it needs the trial to be delayed, the verdict to be appealed, and do everything it can to delay final judgment and sentencing because final judgment is likely to be fatal to the organization as all its juicy assets are seized and it loses all IRS non-profit status. It must be noted that even in cases where defendants plead guilty, final verdict and sentencing can take well over a year so it doesn’t have a lot of time to clear its name if it can.

Seeing how the organization reacts in court filings rather than press-releases is going to show whether it believes it can clear its name and/or plead down to a less damaging charge or not. That of course also depends on whether the DoJ is planning additional prosecutions, which I consider to be likely. My guess is the SPLC is going to delay and obfuscate and take the IRS hits, with the hope that in January 2029 a Democratic president and congress can pass a law or issue waivers that allow it to resume.

That’s a bet. It assumes a number of things, including that state AGs from Republican states don’t get in on the prosecution act because even if federal issues go away, many of the reasons why donations would fall remain present if the SPLC is charged by various states. I will be entirely unsurprised if Alabama’s AG doesn’t have some minion looking at the federal documents and seeing if there aren’t state crimes that can be prosecuted. I understand that federal law takes precedence, but compliance departments will not be happy with a situation where the SPLC’s home state is prosecuting it even if it has skated on the federal side.

And that’s just on the current set of federal charges. As the saying goes

“wait… there’s more!”

Leveraging Trust

We already knew from the twitter files that the SPLC and associated entities were pressuring internet companies to censor views they objected to back in 2017 and that things got progressively worse during the covidiocy and the start of the Biden residency. What Patrick brings to the table is that these entities had a very expansive view of what was an “internet company” and even more expansive concepts of appropriate response.

It started in 2017, with the SPLC and another non-profit informally coordinating. It intensified and formalized in 2018, under SPLC co-leadership. It escalated sharply in 2020 and 2021.

The campaign had two main components. The first was public advocacy and communications work. The second, less visible but more consequential, was a series of meetings with industry. Hundreds of meetings. With a specific target set of companies.

The campaign’s declared aims were three. To convince those companies to censor more communications the coalition characterized as hate. To blacklist organizations and individuals the coalition characterized as promulgators of hate or violence. And to interdict the flow of funds to those blacklisted parties.

The coalition claimed to be non-partisan. Be on the lookout for mentions of “non-partisan,” because it is a word the coalition understands differently than I do.

The coalition calls its targets “Internet companies” and relies on government, media, and the public to not read the fine print. In it, they define Internet company mendaciously to include banks, credit card processors, and any other financial infrastructure their enemies could touch. The coalition was going after posts, but it was also and primarily going after money. I will use the language “industry participants” going forward to identify who they met with.

Industry participants included Facebook, Twitter, JPMorgan Chase, Visa, Mastercard, and many other firms.

If you find it hard to understand why, say, Visa, Mastercard or JP Morgan would be considered an “internet company” then join the club. I do note that all of these played a part in the ending of the pharmaspammers, and as I said earlier, I am quite sure that the SPLC & co looked at that successful ad hoc campaign and decided to try and copy it.

I’m trying not to make this a simple copy paste of Patrick’s article so you’ll have to go and read it to learn how that the SPLC coalition proselytized its views to the “internet companies” it engaged with. You can find that many of the people in the financial organizations were increasingly upset at the threats made when there was push back and that the coalition employed negotiating positions and techniques not a million miles away from those employed by President Trump when he has the upper hand. Only with more smugness and intolerance for alternative points of view.

You have to imagine that this might have caused a certain degree of resentment. As noted above, you can be quite sure, however, that all the surrounding emails, power point presentations and so on are archived. A DoJ prosecutor, state AG or even congressional subcommittee could issue subpoenas for them and find them handed over with a degree of alacrity that might surprise the non-profit entities. I’m just reading between the lines of Patrick’s article but paragraphs like the following do not suggest that the charities made friends

Industry participants recount the tone of the meetings differently, and as varying over the meetings. Some meetings were strained-but-professional. Sometimes the coalition participants were described as demanding and “hectoring.” Industry participants report abusive remarks towards their companies and to the people in the meeting.

Industry participants were repeatedly told that if they did not accede to demands they would be profiting from evil, complicit in the death of innocents, or benefitting from white supremacy. The innocents claimed to be at risk were often specifically identified as black, including during a period of intense societal concern for the lives of black Americans specifically. Industry participants were told that they wanted this. That they were taking “blood money”. Industry participants repeatedly felt personally attacked, in ways and using language not normative in their professional experience.

I can imagine certain parts of certain organizations praying for a comeuppance for their tormentors.

Based on something Patrick wrote on X a couple of weeks ago, I would expect that a search of archived pages for the charitable organizations and activist groups themselves to provide all kinds of corroborating details that document the conspiracy and, quite possibly, document intent to commit what could be considered crimes. At the very least those documents may turn out to provide sufficient probable cause evidence to allow for search warrants, subpoenas and so on that can then build the case for prosecution.

As noted above regarding bank fraud, statutes about blackmail, tortuous interference, causing business or reputational harm and so on do not have “it’s OK if good guys do it” clauses. If I were the Alliance Defending Freedom or a number of other similar organizations I suspect that I might have a claim on some counts against not just the SPLC but other charitable organizations and against specific individuals. This would apply to organizations and individuals who were debanked or had loans refused or other financial problems. Unlike SARs, which banks are forbidden from disclosing to affected customers, if the debanking etc. was caused by a campaign by the SPLC then the bank can certainly disclose this, and the surrounding documents that show that they were pressured to do so by non-governmental organizations threatening them.

Assuming there are multiple victims this could potentially lead to RICO prosecutions because the difference between Alexei or Guido telling a bank employee that they might want to close a certain account or have their children kidnapped is not far removed from a blue-haired non-binary person with a nose ring telling bank employees that they’ll be subject to a twitter storm that gets them fired unless they close a certain account. It does require more than one such incident, but from the way I’m reading Patrick’s article, some of the NGO employees had a reputation for threats and it is highly likely that those threats were documented and the documents archived. With the political climate now very much changed, those hypothetical threats may now result in raids, prosecutions and all the other difficulties that the SPLC and co caused “conservative” organizations during the Biden residency. As people observed, in some ways it doesn’t matter whether these activities result in convictions, the process itself is the punishment. if as part of the process accounts are frozen and computers are seized then the process may effectively bankrupt the victim. Normally I’d have concerns about this, but I also have a strong sense that “what goes around comes around” is fitting.

End Goals

As I said in the title. I don’t think the SPLC will survive. If the SPLC as a legal entity survives, I very much doubt the resulting entity will have the same clout as the current one and it is likely to have forfeited most of its assets and its reputation. But that is just the start.

If the entities the SPLC worked with are prosecuted in addition to named individuals then many of the same donation / tax exemptness issues that the SPLC already faces could apply to these other organizations as well. If the IRS decides to remove or modify the tax-exempt status of these organizations too it may well cause significant pain to the “liberal” or “progressive” cause in the USA. And that’s before anyone starts following the money trails back. Because if a non-profit is perceived as acting in violation of the rules then an immediate logical next question is whether donors to that non-profit knew. And even if they didn’t having to prove it is likely to be expensive and time consuming (see process is the punishment above). Freezing the assets of major foundations who are suspected of involvement would cause a lot of other charitable bodies to run short of resources.

This ignores the other aspect that Patrick digs up: the documented interference in political campaigns. That is yet another thing that charities are explicitly forbidden to do if they want to retain their tax exempt status. As he says at then end

As a former 501c3 CEO myself, I am aware of the requirements to maintain tax-exempt status. This is of paramount importance to charities. You can save yourself some legal bills quickly with the IRS’s Restriction of Political Campaign Intervention by Section 501(c)(3) Tax-Exempt Organizations :

“Under the Internal Revenue Code, all section 501(c)(3) organizations are absolutely prohibited from directly or indirectly participating in, or intervening in, any political campaign on behalf of (or in opposition to) any candidate for elective public office. Contributions to political campaign funds or public statements of position (verbal or written) made on behalf of the organization in favor of or in opposition to any candidate for public office clearly violate the prohibition against political campaign activity. Violating this prohibition may result in denial or revocation of tax-exempt status and the imposition of certain excise taxes.”

501c4 organizations have similar considerations. Consult your lawyer.

It seems to me that Trump 2.0 has found the first, loosest thread. Pulling on that could well unravel a whole tapestry of related charities, foundations and donors who have been involved in activities incompatible with their tax exempt status. Especially when what is discovered from the SPLC related investigations are combined with the existing money trails documented by DataRepublican and allies and with evidence released by the Trump administration’s Task Force to Eradicate Anti-Christian Bias

I see no sign that Trump 2.0 will not keep digging and I expect all kinds of other misbehavior to emerge and be prosecuted. This is just the tip of the iceberg.

Share

That “two week” ceasefire turned out to be extremely shortlived and predictably it seems the Mullahs misread the room and maybe the coverage from the MSM and decided they could make demands of VP Vance and team that would allow them to keep their nuclear ambitions and the like.

I think they must have got some “How to handle Trump” lessons from the Somali owners of the “Quality Learing Center” because they sure Leared Gud about how President Trump handles uppity foreigners making demands.I do wonder whether President Trump uses the MSM’s hatred and disdain for him as an additional leverage point to get his way as his enemies are lulled into believing that he’s an impetuous idiot. As I saw someone say somewhere recently, with Trump it’s not so much that theres a method to his madness so much as the madness is the method. Trump’s enemies believe what the MSM says about him and thus misunderstimate him to their loss.

Remember the “You don’t have the cards” statement to President Zelensky? - apparently that was not in the Learing Center plan that the Mullahs studied. That was their loss as they now have allowed the US to look like the reasonable good guys while they come across as the fanatical nutters.

Ukraine

It must be noted that Zelensky certainly took note of his schooling at the hands of Trump, and has done a sterling job afterwards of negotiating a hostile Washington while not losing to Russia. In fact thanks to new investments by gulf arab nations into Ukrainian drones and drone defenses, Ukraine is probably financially secure that it no longer absolutely needs US support, though it certainly benefits from that support, and as “A Muse” points out, US support of Ukraine under President Trump is greater than both European support of Ukraine and European payments to Russia for oil and gas

amuse on 𝕏 by Alexander Muse

Trump Is Quietly Sending Ukraine Billions While Europe Continues to Send Putin $2 Billion a Month

Consider a simple thought experiment. A man announces he will stop paying his neighbor’s electricity bill. The neighbors throw a fit. The media runs breathless headlines about the cutoff. Pundits across Europe wring their hands. Volodymyr Zelensky makes the rounds on cable news lamenting the betrayal. And through all of it, the checks keep clearing. Eve…

Read more

4 months ago · 100 likes · 24 comments · Alexander Muse

Yes the Europeans are paying both sides of the Russia/Ukraine war roughly equal sums, and probably in total have paid Russia more. Oddly enough people like don’t mention that when they bitch about how Trump is handing Putin everything he wants. Bizarrely, as AMuse points out, it seems to be in the interest of all sides to ignore the fact that the US is transferring $2B a month of military stuff to Ukraine.

Talking of PPOB… he mentions1 a fascinating series of Xeets by President Zelensky that should make uncomfortable reading for many. Note that I don’t think Zelensky is wrong in anything he writes either, though I’m quite sure he’s being an advocate for Ukraine and thus a neutral party might quibble or mention things that Zelensky skips.

Anyway as PPOB mentions in his weekend Ukraine update, Russia seems to be benefiting from the war in Iran because of the US temporarily suspending sanctions on Russian oil and gas. He does the sums and reckons that, with Ukrainian attacks included, Russia is still ahead

I think that sum is fine as it goes. But… there are a few things it ignores. One quick thing is forgets is that Russia’s navy is now -2 large expensive ships (the Admiral Essen in Novorossiysk and the ice breaker in Vyborg), which are at this point effectively irreplaceable. Russia’s Black Sea fleet is now down to one major surface vessel and even that is basically locked in port awaiting its own rendezvous with a Ukrainian drone.

But there are two larger trends as well.

Firstly Russia will no longer get any drones or other weapons from Iran. The Iranian factories have been destroyed and many of their leading scientists and engineers appeared to have been droned, along with the ports on the Caspian sea that could transfer them. It is not clear to me how many of Russia’s more recent drone attacks used Iranian built drones, as opposed to ones built in Russia under license, but given that Ukraine has also struck heavily at Russia’s drone factories the fact that it can’t get resupply from Iran certainly hurts the war effort. In particular, if there are specific jigs or machine tools needed to build these drones that come from Iran and that have been damaged by Ukrainian attacks, then those items are not going to be replaced directly. Instead Russia will have to get someone else (probably China) to supply something similar and then rejigger the process to use the Chinese kit that is not quite the same as the Iranian kit. That makes repair slower and more problematic.

Secondly, the same applies, only more so, to repairs to its oil and gas infrastructure. Both Iran and Russia need to repair war damage. As do the gulf nations hit by Iran. There are not a huge number of suppliers of new distillation columns, pumping station parts and so on. For certain things (e.g. LNG processing facilities) there are no more than about three suppliers of key components and all of them are either German or Japanese (or maybe Swiss/other European); for others there are more but still the number is limited and again is mostly Japanese or European. All these manufacturers now have priority orders being placed by the gulf nations that they will happily put to the front of the queue because the gulf nations will pay them in advance in the currency of their choice and without any of the required export license jiggery pokery that Russian orders have. That means that Russia is now reliant on Chinese manufacturers (and even then it has to fight other customers, such as Iranian ones) and it is a dirty little oil and gas industry secret that the Chinese products are markedly inferior to the German/Japanese ones. There may be some Indian subcontinent manufacturers too, but their quality is no better than the Chinese ones.

In other words, Russia is not benefiting much from the Iran war. In fact in the medium term it gets worse. The Ukrainian supplying of drone defenses to the gulf means, as I said above, that the Ukrainian defense industry is getting a large investment from the gulf states. That in turn means that the Ukrainians can build more and better drones and missiles and can therefore continue their campaign to take out anything they want west of the Urals that isn’t in the ring of Moscow’s air defenses. In fact, while the war over Iran has taken all the headlines, the war in Ukraine is not going Russia’s way at all. Since the start of the year it seems that Russia has actually lost territory and the recent spring offensive has been a bloody disaster (literally) with Russia now committing reserve forces from other parts of Russia because it lacks enough conscript drone fodder. Russia is not yet out, but I will not be shocked if there is a sudden collapse somewhere in the Russian lines that allows Ukraine to regain yet more territory.

Elsewhere in Europe

Two other events this weekend.

In Hungary, Victor Orban lost his re-election battle. Which shows that all the rot about him being a would be dictator was false. It should be noted that the cheers of the Orban haters in the EU and so on may be a little premature. His successor was a member of Orban’s Fidesz party until a couple of years ago and he has campaigned in part on reassuring Hungarians that (lack of) immigration was still a priority. Orban’s fall from power hurts Putin and perhaps embarrasses President Trump, but it is almost certainly not the victory for the globalists that they seem to think it is. Rod Dreher (who live in Budapest) has a lot of useful information and thoughts.

At the other end of the continent, in Ireland, it looks like the government has mostly capitulated to the anti-fuel price protests. It is notable that green inspired tax rises are being postponed, and I suspect, now they have been postponed it won’t take too much for them to be quietly abolished over the summer.

I also expect that protestors in other countries will have noted the Irish protests and their success. I predict that France, at least, will see something similar. Germany may too because the price of energy is really hurting the German economy. In fact on that note, it is worth pointing out that Germany

is about to flood one of the largest energy reserves in Europe. The country’s largest open-pit mine, Hambach, sits on more than one billion tonnes of extractable lignite. Originally permitted to produce coal until 2045, the mine will shut down early under the Act to Reduce and End Coal-Fired Power Generation, passed in 2020. Its owner, RWE, received €2.6 billion in compensation for the financial burdens imposed by the law. This sum covers the early closure of coal-fired plants in North Rhine-Westphalia as well as the closure of Hambach and other mines.

RWE is now obliged to flood the pit fifteen years earlier than originally planned. Starting in 2030, Rhine water will create Germany’s second deepest lake over the course of four decades. The project is not uncontested on environmental grounds. Critics cite excessive strain on the region’s water reserves and long-term chemical risks to the groundwater, among other issues.

Drowning a billion tonnes of coal under a recreational lake—making future access economically impossible—is not unlike dynamiting functioning nuclear cooling towers. Both amount to the irreversible destruction of a strategic reserve, executed by politicians who seem to price energy security at zero.

The Brawl Street Journal

Rocks to Barrels

Germany is about to flood one of the largest energy reserves in Europe. The country’s largest open-pit mine, Hambach, sits on more than one billion tonnes of extractable lignite. Originally permitted to produce coal until 2045, the mine will shut down early under the…

Read more

4 months ago · 87 likes · 22 comments · The Brawl Street Journal

I wrote previously about Green Treason, this combined with the current ban on fracking in Germany and most of Europe makes the continet vulnerable to events like the Iran war (and the Ukraine one for that matter) that they wouldn’t be if their politicians had enough braincells that they could see the stupidity of their NetZero Greenidiocy.

Subscribe now

The Blockade

Back to the Persian Gulf. The immediate outcome of the failed talks is that the US has announced it will blockade Iran. Meanwhile, the US Navy is also passing its own ships through the strait and apparently planning minesweeping operations

That means that the Chinese social media post and map above is now wrong for most countries. It looks like the Iranians are going to struggle to impose their toll regime. while US will impose its one just fine. That is going to really hurt Iran as this long Xitter thread explains:

The U.S. naval blockade of the Strait of Hormuz would cost Iran approximately $276M/day in lost exports and disrupt $159M/day in imports, a combined economic damage of ~$435M/day, or $13B/month.

Over 90% of Iran's $109.7B in annual trade transits the Persian Gulf. Oil/gas accounts for 80% of government export earnings and 23.7% of GDP. Kharg Island alone generates ~$53B/year, or as I noted to @TIME, "$78 billion a year in energy revenue

…

ALTERNATIVES? Iran's options outside the Strait are negligible. Jask, the much-touted bypass, operates at a fraction of its 1M bbl/day design capacity. Only 10 of 20 storage tanks were built. Effective throughput: ~70K bbl/day. Chabahar handles just 8.5M tons/year. The five Caspian ports combined handle 11M tons, versus 220M+ through the Gulf.

…

Extremely important topic is the storage clock: Iran has ~50-55M barrels of total onshore oil storage, roughly 60% full. Spare capacity: ~20M barrels. With 1.5M bbl/day of surplus production that normally exports, storage fills in ~13 DAYS. After that, Iran must shut in wells.

Why is this very important: when mature oil wells shut down, bottom water rushes in, a process called water coning. Oil droplets get permanently trapped in rock pores. This oil can never be recovered. Iran's fields already decline 5-8% annually. Forced shut-ins could permanently destroy 300,000-500,000 bbl/day of production capacity, that's $9-15B/year in revenue, gone forever.

…

BOTTOM LINE: A naval blockade imposes ~$435M/day in combined economic damage. Storage fills in 13 days, forcing well shut-ins that cause permanent reservoir damage. The rial enters terminal collapse. Iran's alternatives outside the Strait can replace less than 10% of Gulf throughput. The blockade makes continued resistance economically impossible.

This is actually a game changer, not only does the economy collapse even more but also Iran has roughly 13 days to get with the program or it starts losing oil production that cannot be restarted. Or at least, not easily. My belief after a couple of searches on the internet2 is that some fracking techniques can be used to fix the water coning to some degree, but there’s only one country that has the technology to do that. You are invited to guess which one that is.

If Iran does not capitulate in 13 days then it will lose oil production of 300,000-500,000 bbl/day (out of some 1.5M bbl/day, so ~25% or more) that can only be recovered if American experts are allowed to work the fields. That expertise is not going to get there without agreements similar to the ones with Venezuela where US oil firms get a share of the output. The blockade also stops Chinese weapons resupply vessels and means that the US can now simply seize shadow fleet vessels and their contents and resell both. So not only does Iran not get the money or resupply, the shadow fleet shrinks and the shadow fleet customers have to pay market rate in US$ to get the oil.

I’ve seen some people wonder whether this will cause our Chicom friends to get frisky. The problem they have is this - (most of) the shadow fleet tankers are not flagged to China or indeed any major shipping nation. It is possible that a Chinese flagged tanker will try to dock at Kharg Island, but if it does I very much doubt it will leave. In fact a truly wicked thing the US could do is threaten to mine the entrance to the Kharg Island oil terminal and dare tanker owners to discover the hard way whether it has done so or not.

Finally the US has promised to seize vessels that seek to pay off the IRGC. Amuse pointed out last week that it is quite possible no vessel has done so despite the news reports. Combine that with the minesweeping and it looks like the potential funding from protection money that the IRGC hoped for simply won’t emerge.

The Slow Oil Shock and West Taiwan

As I and others noted repeatedly, the US has been in no hurry to ensure the Strait of Hormuz remains navigable. That, it seems to us, is a deliberate strategy aimed at people other than the Mullahs. The prime target is almost certainly the Chicoms, who are being hurt a lot by the lack of oil and other trade through the strait. The NY Times had a bit of Chicom copium last week that claimed with little concrete evidence that the PRC was less vulnerable to lack of gulf oil than other nations. However that same article also stated that

Chinese companies have come to dominate much of the world’s chemical supplies. Three-quarters of the world’s polyester and nylon, for example, are made in China.

Now it is quite true that the PRC has done a lot of adaptation to use coal rather than oil as a feed stock for its chemical industry and it’s fertlizer industry. But despite that, and despite the coal powered EVs, the country remains heavily dependent on oil and much of that oil came from the gulf region.

I have seen posts on Xitter that suggest that the lack of oil is causing a lot of pain to China’s plastics industry. Here in Japan I have seen a few news reports about shortages of naphtha and some plastic wrapping/packing and I strongly suspect that there are a lot of similar shortages in West Taiwan. The most important thing this does is add an additional challenge to the PRC’s economy which is not in great health to begin with. That, combined with the destruction of Iran’s PRC supplied air defenses and the failures of its PRC supplied missiles delays any attempt to take Taiwan by months if not years. If it is delayed long enough it becomes impossible because Taiwan will have its defenses improved as will Japan.

Finally, because nothing says that the President can kill two birds with one stone. The lack of oil traffic in the Persion Gulf means oil tankers are headed in record numbers to the Gulf of America to fill up with American oil.

All in all it’s been an interesting weekend, with about as many events as usually fit in to a couple of months. But that’s 2026 for you: about a years worth of events every month

Share

OK so there’s a ceasefire of sorts that seems to be mostly holding and it seems like that could lead to a longer term peace. Not permanent, but something that lasts anything from a couple of months to a couple of years.

So who won?

Paul Mirengoff thinks Trump didn’t win and I sort of agree

Ringside at the Reckoning

Who won the U.S-Iran war?

There's some say that we wan…

Read more

4 months ago · 3 likes · 1 comment · Paul Mirengoff

But I think that even if President Trump didn’t win, he certainly was the smallest loser. The biggest loser was Iran’s regime, the next biggest losers were Hez(no)bollah, followed by “the international order” and Europe. West Taiwan also lost more than gained and Russia has been unable to take advantage of the situation because Ukraine has destroyed most of its Western oil export infrastructure. Israel is still fighting Hez(no)bollah so it hasn’t yet won either. In short so far every side has lost to some degree or another but the US has short term lost least, and in the longer term looks to be the winner

Military Wins/Losses

Militarily the US and Israel won bigly. As this Xeet sumarizes:

General Caine this morning. The US has destroyed:
- ~80% of Iran’s air defense systems (>1500 AD targets)
- >450 ballistic missile storage facilities
- >800 one-way attack drone storage facilities
- >200 C2 nodes
- 90% of the Navy (150 ships) and half of IRGC small attack boats
- 95% of naval mines
- 90% of weapons factories including every factory that produced Shaheds and their guidance systems
- 80% of missile production facilities
- >20 naval production and fabrication facilities damaged or destroyed
- 80% of nuclear industrial base

In return the Iranians destroyed one AWACS plane in Saudi Arabia, some KC-130 tankers, one F-15 and one A-10. There were a few other losses that were not caused by Iran that can be added, such as the destruction of planes left behind after extracting the crew of the downed F-15E. The Iranians also appear to have shot down a number of drones and may have damaged various helicopters and some static radars and other US kit here and there.

Total damage to US/Israel (including missiles attacking Israeli civilian targets) is probably in the low $Billion area (say $5B, plus or minus). Add another similar or slightly more for the gulf state infrastructure damaged by Iran. Total casualties from all causes on the US/Israel side are under 100 with perhaps 30 dead total (including Israeli and indeed Palestinian civilians).

On the other hand, both Iran and Hez(no)bollah have seen their leaders killed on a scale that has never been seen in the last century, possibly ever. In WW2 the leadership of the losing powers were executed, but that was after the war had ended. Very few Nazi or Japanese leaders were killed during the fighting and that applies to pretty much all of the other wars of the 19th and 20th century if not earlier. In addition to the leaders, several thousand regular IRGC, Basij and Hez(no)bollah fighters have been killed. Likewise, in addition to the military loss of effectively the entire air force, navy and air defense systems, Iran has also seen significant chunks of its aerospace and drone manufacturing chain destroyed. Everything from the steel mills and aluminum refining operations to the R&D facilities and the factories to build the drones and missiles have been leveled. It is possible that some factories are salvageable but in general we’re looking at entire sections of the Iranian economy that have to be rebuilt from the ground up. Since some of those factories were dual use, that means Iran will now have to import steel, aluminum, fertilizer and numerous other basic chemicals that it previously manufactured as well as import all the machine tools and chemical engineering plant that are needed to replace what has been destroyed.

In Lebanon, not only is Hez(no)bollah losing, it is losing so much that it looks like the rest of Lebanon may actually be willing to turn on them. It has become clear that even if Iran wants to give support to Hez(no)bollah, it can’t in fact do so anymore. So once Israel has taken out most of their missiles the rest of Lebanon will probably take revenge for decades of Hez(no)bollah abuse of other parties in the country. It could well turn out that the Lebanese Christian and Druze communities end up running Lebanon fully with the Shia in a totally subordinate role. This would be a total reversal of the conditions in the recent past where the Shia Hez(no)bollah dominated the nation.

Back to Iran. Not only has the government lost many of its leaders, the survivors have lost significant income because of the various bribes and kickbacks that the now destroyed industries paid but which cannot be paid anymore. Furthermore the indiscriminate attacks on the gulf state infrastructure means that those states have decided to stop tolerating Iranian activities. Hence the various brokers, money changers and other middlemen that used to enable sanctions evasion for the mullahs are no longer welcome. The UAE has already cracked down on them and I think I saw reporting that Qatar has too. I do not believe that these activities will be allowed to resume now that the shooting is allegedly over.

In terms of what President Trump said he wanted - the removal of Iran’s ability to make nuclear weapons as well as drones and missiles and the export of terrorism - he has achieved his basic war goals. However he did not achieve any of the “stretch” goals. Specifically regime change has not happened and traffic through the Strait of Hormuz remains threatened by the IRGC.

Clarifying The Strait Issue

This is the big issue that allows Iran’s symps to claim that it won. I think that’s a stretch. However you can make the argument that Iran having the ability to block the strait means it has not lost and that is fairer. However I don’t think that the opening or closure of the Strait of Hormuz was something that concerned either the US or Israel. As I said in a previous post, the US gets very little from ports inside the Strait and it doesn’t export much there either. Israel does even less direct business that requires stuff being shipped through the Strait. Thus, while the state of the Strait directly impacts the Persian Gulf countries a lot and also impacts the countries that buy their fertilizer oil and gas (most of Europe, Asia and Africa) those countries are not the ones involved in the war. Of course indirectly the closure of the Strait and the damage to gulf oil and gas infrastructure does affect both the USA and Israel because the loss of those exports impacts the global economy significantly, but the closure is a much bigger problem for these other countries.

Hence its closure clarified a number of points. Most of them related to how much of the world turned out to have outsourced protection of shipping to the US Navy.

First the gulf nations were unwilling to directly confront Iran. Unless the US told them not to (which I doubt), the fact that none of UAE, Qatar, Bahrain, Saudi Arabia or Kuwait were willing to actively combat Iranian activity says a lot. Mostly what it says is that the Gulf Arabs are dependent on the good will of their neighbors to survive and are either unwilling or unable to project offensive force at all against someone willing to shoot back. There is no good reason why countries with the wealth these nations have should not be able to have air and naval forces that can escort their ships though the Strait of Hormuz. Yet none were willing/able to do so.

Second Europe is useless. The Royal Navy was particularly risible, but the only nation that did anything to actually help the situation was Ukraine, which sent drone defenses and experts to deploy them to the gulf. Ukraine knows all about how to control waterways without an actual manned navy and lots about drone warfare so it has the expertise that the gulf nations need. Unsurprisingly Ukraine’s defensive drone makers have received lots of investments from gulf states and I imagine that their naval drone makers will get a lot too. Once those investments bear fruit, perhaps the gulf states will be willing to confront Iran.

Thirdly the UN has been completely irrelevant. There was one resolution condemning Iran which was vetoed by both Russia and China. There were precisely zero constructive debates about the freedom of the seas and forces required to ensure that shipping can travel freely through the Strait. In fact the only thing that I noticed emerging from the UN over the last month was a plea to please send money because otherwise the UN will go bust. Oh, and in Lebanon the UN presence there was shown, yet again, to have been completely unable to perform its designated mission of keeping Hez(no)bollah from the south of the country. If the UN needs to save some money then cancelling UNIFIL would be a good start (cancelling UNRWA would be even better, but there are way too many grifters profiting from that one for it to be on the chopping board at first). If the UN wants the US to fund it, which apparently it does, then you would think the UN would be trying a bit harder to be relevant.

Fourthly, a bunch of other countries (looking at you Australia and New Zealand particularly) turn out to have needlessly closed refineries and local oil/gas infrastructure for reason of greenidiocy and thus are now dependent on imports of things like diesel. Making it worse, they also don’t have any decent stockpiles so that a global tightness in the oil market means their economies almost literally grind to a halt. I wrote earlier that the green activists and the politicians that caved to them should probably be tried for treason but more critically voters need to watch what the current leaders do and see if it is likely to help in the future. To that end I see that NZ has made plans to add diesel storage that should come on line in a couple of months. It’s only eight more days usage so not that much, but it is a decent first step. I do not believe its larger trans-Tasman neighbor has done anything similar.

Fifthly I think the only country to actually escort fuel tankers through the strait has been India, which also has paid off the IRGC a couple of times and taken advantage of President Trump’s leniency to snap up grey market cargoes. China appears to have just tried to buy the IRGC off and mostly failed in that attempt. Iranian oil for China has been allowed to pass the strait, but China needs more oil than that. Although it purchased 90% of Iranian oil and about the same proportion of Venezuelan oil last year, they were just 20% of its imports and a decent chunk of the rest came from places like Iraq and Kuwait. Critically both India and West Taiwan use gulf oil and gas for much more than just fuel. Their entire plastics manufacturing sectors are running out of affordable raw materials/feedstock and that means all those companies making chinesium for export and sale via Walmart/Amazon/Temu etc. are looking at severe cost issues if they can actually continue to manufacture at all. The news says that one of the drivers for Iran accepting the ceasefire was China putting the pressure on them, which is interesting if true and makes sense because if the war had escalated then China’s manufacturing sector would be lacking raw materials for years instead of, as things stand now, weeks.

Anyway in summary, what the closure of the Strait has told us is that no nation is willing to actually do the dirty work of ensuring freedom of the seas. That includes West Taiwan, the country with the largest navy that depends gulf oil and gas to keep its economy alive. There have been zero multinational task forces proposed for Hormuz and only one nation - India - has actually put a warship in range. The world has been shown to be so dependent on the US that even its would be enemies and peers (West Taiwan) can not act to defend their own interests if the US decides not to bother.

Subscribe now

Was Trump Too Ambitious?

That is a possibility. Chris Bray assembles some evidence suggesting that the President thought Iran would collapse quicker and further than it did.

I think this is plausible. Iran’s total failure to defend itself last year may have made the US and Israeli leadership over-confident. They may have also thought that the number of (surviving) hardliners in the regime would be small enough that they could be removed by everyone else. Also the fact that last time the Strait of Hormuz was not threatened may have led them to think that the same would happen this time around when there has been considerably more damage done to Iran. Or they may have thought that regime decapitation would lead to regime change as the Iranian population revolted. I think they missed the timing for that last one. Had the forces been there to take out the leadership in January then regime change could well have happened but after the IRGC and Basij stacked up the bodies of possible rebels the population seems to have decided that they need a regime significantly more weakened before they are willing to strand up to it.

It is also entirely possible that earlier plans for the US to occupy the islands in the Strait and ensure freedom of shipping for a small toll were delayed because subsequent intelligence suggested that attempting the occupation would result in US casualties. Quite likely the calculation of how many casualties for what gain had to be redone and the new sums didn’t add up. Ironically the protestors chanting “No blood for oil” may have found agreement from the very top of the US administration who didn’t want to sacrifice US lives to open a strait that mostly benefited others. So yes. Perhaps there was some initial overconfidence. But it seems that the President is willing to take advice and not risk casualties for immediate gain if there are alternatives.

However I think that despite that, the US and Israel have achieved the primary war goal of removing the ability of the regime to create WMD and project terrorism beyond the gulf and hence, despite how pro-Iranian symps may spin it, the mullahs really have lost. This means I tend to disagree with Niall Ferguson

Niall Ferguson's Time Machine

Iran Thinks It’s Winning. It May Be Right.

Late last night, the United States and Iran agreed to a two-week ceasefire. But the first day has already proven rocky: Kuwait, Qatar, the United Arab Emirates, and Israel all reported missile and drone attacks from Iran; Israel carried out major air strikes against Hezbollah, raising questions about whether Lebanon is included in the ceasefire; and Ira…

Read more

4 months ago · 31 likes · 1 comment · Niall Ferguson

The Future of the Mullahcracy

The critical point is that while the regime may have survived the bombing campaign it may not last that much longer anyway. The January protests were driven by economic collapse. The economy is in even worse state now and the regime hardliners were particularly targeted by the US and Israel. In other words the regime is worse off and desperate. It is also, according to rumor, trapped in that there are essentially no functional jets in the country so flights out are problematic. This provides the possibility for some carrot and stick.

Mark Simon thinks that the strategy may be to appeal to the corrupt grafters in the regime while continuing to sideline the fanatics.

Mark Simon on Substack

Corruption and Graft Help President Trump Seal a Deal with Iran

That there would be a last minute way out of Trump bombing Iran into the Stone Age was as assured as Tom Cruise escaping a tyrant’s dungeon in the first ten minutes of the latest Mission Impossible. After the last nine years how could one expect any other ending from President Trump. He had cleaned out the top of the Iranian regime and was already ca…

Read more

4 months ago · 6 likes · Simon, Mark

I think this is plausible and in fact it is not too different to what we are seeing in Venezuela. The regime is allowed to survive as long as it does what Washington tells it to and doesn’t kill too many of its citizens. Iran’s leaders have, as I wrote above, just lost most of their personal funding because the businesses that bribed them or that they controlled have been destroyed and so cannot pay them. The system that allowed them to get their money out and luxuries in has also mostly been taken out and if they want to continue to live in luxury they need to find alternatives.

Reconstruction fund tolls on shipping are a way to do this. The trick here is that the grifters need to ensure that when tolls are paid no IRGC fanatic decides to send a drone or rocket at passing shipping. That gives them a strong incentive to crack down on potential fanatics and if that leads to sharply pointed debates between different wings of the regime then that is perfectly fine. The stick that goes along with this are the MEUs that are still in the area. If (when?) the IRGC breaks its side of the bargain and goes after ships (or even seeks additional protection money payments from ship owners) the MEUs can implement the occupy Hormuz plan.

It is possible that there will be no formal regime change. There will still be a guardian council and religious scholars from Qom will still be being paid off. But the hardliner parts of the IRGC will find themselves being arrested and executed and the lower levels of Basij and IRGC will be defunded. The end result will be another sleazy middle eastern corruptocrat regime that pays superficial allegiance to Islam while allowing for more economic and religious freedom. It may even manage to convert a significant fraction of the toll income into actual economic gains for the majority of its citizens.

Long Term Winner

In the long term America has won and the international world order has lost. The worlds has had its attention drawn to the fact that the Middle East is a risky place to rely on for fossil fuels. Coincidentally the US and its western hemisphere allies are excellent places to obtain fossil fuels with little or no risk. The same applies up the supply chain. Chips, plastic widgets and all kinds of other things can be made at lower geopolitical risk and equivalent price in the Americas. As companies look for suppliers to replace those struggling thanks to middle east supply issues, ones in the Americas look increasingly attractive.

The international order turns out to be dependent on American blood to keep the peace. When America decides it doesn’t care there is no one able to replace it. I see absolutely no sign that even now, a realistic alternative is being proposed by any would-be peer. This is a subject President Trump has spoken about for a decade. Despite a decade of warning from President Trump much of the world seemed to be convinced that the US would put the interests of the rest of the world ahead of its own narrower ones. That has turned out to be wrong and the resulting recalculation means that a number of freeloading nations (Europe) have to pay for their own defense and so on instead of relying on the Americans to rescue them. Rather than allies, these countries are now supplicants and dependent on the US for both energy and external defense. They have to pay for both.

What is also clear is that the US is in fact still the sole great power militarily. West Taiwan and possibly Russia may claim to have equivalent militaries but it is quite evident that although they may have the quantity they entirely lack the quality. Iran had lots of West Taiwanese and Russian military technology. None of it worked against the US or Israel. Even things that have worked in Ukraine turn out not to work very well against a military as rich as the US. That’s not to say the US military has everything right, it certainly doesn’t, but it has enough right and enough resources to handle the problems caused by the bits that aren’t right.

Israel is, of course, the other winner. Not only is it removing the local threats to existence, it has cut off their funding and supplier so that the local threats are fundamentally left to rot on their own. Moreover, nations that were ambivalent have now discovered that while Israel hasn’t hurt them, Iran has been quite happy to. A future middle east is going to be an axis of Sunni Arab nations relying on Israel for security against external threats.

It will take a while for the penny/cent/yen etc. to drop, but the first three months of 2026 are going to be seen as the time that the USA became indisputably the world’s leading nation and the leader of a select group of regional allies that share its technological advantages.

Share

One of Aesop’s famous fables is about the ant and the grasshopper, where the former prepares for winter and thus survives while the latter doesn’t. The current war around Iran, combined with the ongoing one in Ukraine/Russia, shows us clearly which nations are ants and which are grasshoppers.

I am extremely happy to live in one of the former. Here are some of the things it does. You may wish to compare it with what you see from other countries and decide which side of the ant/grasshopper divide they are.

By the way (and the excuse reason for the image above), it occurs to me that one reason Japan and the USA get on well together these days is that culturally they share a couple of traits. One is pride in doing/building/making things. The other is a willingness to culturally appropriate anything they see that looks useful and then, having adopted it, improve it. In general both cultures respect the doer; the ant that is, rather than the grasshopper.

Anyway back to Japan the ant among grasshopper nations

Control fuel prices and ensure availability

As of today, gasoline prices in my corner of ruralish Japan are in the mid 170s yen/liter range. I saw an outlier advertising 163 which seems implausibly low and makes me wonder if they failed to update the sign, and my closest one, which is generally on the low side was JPY173/l. This is roughly what they were six months ago though in between they decreased to the mid 140s and a couple of weeks ago they had spiked to 195 or even over 200 at some outlets.

However while we liked paying 145 for gasoline, if it now stays at or below 175 it will be no different to prices a year ago. Impact on residents budgets, inflation etc. is minimal.

Japan can do this because it stored about eight months of oil prior to the Iran war. As a result it is able to supplement continuing imports with releases from the stockpile to keep prices down and availability to be universal. Yes it is subsidizing prices, a bit, but I’m fairly sure that the cost of that subsidy is far lower than the loss of tax revenue from seeing a Covidiocy level of lack of economic activity that is the likely result of not controlling prices.

Japan also imports most of its LNG from places other than Qatar or the middle east in general. That graph is for 2023 but I don’t think things have changed much in subsequent years. Now it is true that the cyclone that just hit the Western Australia LNG ports doesn’t help but Japan also appears to have plenty of stockpiles and alternatives (e.g. coal, nuclear) for the major use - electricity generation. Assuming most of its suppliers are able to maintain deliveries then the loss of the 5-10% of its supply that originated in the middle east is not at issue.

Relatedly, I have seen no reports of lack of fertilizer. Japan appears to import about 20% of its fertilizers (in currency terms, not sure of volumes or whether different fertilizers vary in amount) and I expect there is plenty stockpiled for this year. Since, as noted, I live in ruralish Japan if there were issues I’m pretty sure I’d have seen coverage. Instead what I observe is farmers doing their normal spring ploughing and similar preparations

Is life entirely unaffected? So far pretty much so. There are some reports of chemical industry processes reducing or halting (e.g. ethylene production IIRC) but so far nothing that impacts the average Japanese consumer. If some plastics become less available then presumably this will change but if the biggest problem is an unavailability of clingfilm or plastic bags then Japan will have come off very lightly.

Have a Competent Maritime Defense Capability

Last month Japan started work on its second ASEV (Aegis System Equipped Vessel), which should probably classed a Guided Missile Cruiser, though it is unclear if the C designation will be used (let’s face it Japan still calls its two “not really aircraft carriers, honest guv” aircraft carrying warships “Helicopter Destroyers”. Though apparently it plans to change that soon).

At 190m long and 25m broad, the ASEVs will be some of the largest non-aircraft carrier ships being built by anyone this century1 until the Trump battleships start construction. The Trump battleships are projected to be about 50% larger but they haven’t yet been funded let alone contracted. The USN’s Zumwalt’s are roughly the same size.

It isn’t just the size that is impressive, it is also the speed of construction. The ASEV as a concept showed up in ~2020 when Japan decided the land based Aegis Ashore program was a failure and cancelled it. In October 2024, after about four years, plans had been made, budget allocated and contracts awarded for the two ships. Construction of the first started last July (2025) and the second last month (February 2026). Construction time is estimated to be around three years, with the first delivered/commissioned in March 2028 and the second one year later. Why one year later and not six months earlier? My guess is that the reason is to incorporate lessons learned from sea trials and operating the first in the final construction stages of the second. This seems similar to the Izumo/Kaga construction a decade ago.

Put together and you have ships that will have gone from concept to contract award in 4 years and contract award to fully-functioning delivery in under 5 years - assuming there are no delays. That seems plausible, the Izumo and Kaga were built in about the same time frame, and stands in stark contrast to the procurement speed of the US Navy and any European navy. The first Zumwalt, for example took over 5 years to go from concept to start of construction and another 5+ from there to commissioning and then another 4 to full acceptance. The ASEVs are also expected to be a lot cheaper, costing around JPY400B or about US$2.5 billion for the initial version. There will undoubtedly be upgrades - e.g. drone defenses, laser or rail guns - and there are some new features compared to previous ships - the SPY7 radar from Lockheed Martin for example - but this is an evolution of existing Japanese and US Aegis destroyers rather than a brand new concept which helps explain why I am confident about the timeline and budget.

Although the ASEV’s primary role is missile defense, there is no reason why one might not, in the fullness of time, be loaded with offensive missiles such as the TLAM or antiship missiles. Indeed the Naval News article lists both as options:

Each vessel will feature a 128-cell Vertical Launch System (VLS)—significantly more than the 96 cells installed on Japan’s latest Aegis destroyers.

The VLS will be capable of launching:

• SM-3 Block IIA interceptors, jointly developed by Japan and the United States for ballistic missile defense

• SM-6 missiles, capable of engaging advanced aerial threats, including hypersonic glide vehicles

The ships will also support Japan’s emerging counterstrike capability.

Planned armament includes:

• the extended-range Type 12 anti-ship missile (ship-launched improved variant)

• the U.S.-made Tomahawk cruise missile

Notably the Tomahawk capability has already been fitted to one of the Aegis destroyers this class is supposed to replace. A 128 cell ASEV firing TLAMs could be very unpleasant for Little Rocket Man; if it fired antiship missiles that could make a Chicom invasion of Taiwan pretty miserable just on its own. It would also make a phenomenal commerce raider / blockade enforcer if such roles were needed.

[Aside: Unlike some country’s missiles I’d be pretty confident that the Type 12 anti-ship missile will work well]

As observed on X, the real question is why Japan doesn’t make a few more and sell them to countries/navies that struggle with procurement. And for that matter why it doesn’t make a couple more for itself. Perhaps it will. I figure there’s a couple of years before there will be yard space to build them so there’s no hurry to make that decision.

Subscribe now

Building Houses

And for that matter factories and all sorts of other things.

There have been a ton of posts by people in the Abundance left (such as Noah Smith) that talk about how red tape blocks things from happening that need to happen. Often really it isn’t so much red tape as green tape from environmentalists combined with NIMBY activists who don’t want their life disturbed.

The result is that in countries like the Yookay there is a massive housing shortage, despite successive governments pledging to get more houses built.

In Japan people just buy the land and build the house.

Or the factory.

This post by Yuki Homes explains some of how that works

YukiHomes

Zoning and Lack of NIMBYism Is the Real Driver of Japanese Housing Prices.

The major discourse in American politics today centers around affordability. And when people say affordability, most are specifically talking about housing. It’s becoming increasingly hard for young Americans to purchase a home, and that frustration is boiling into national politics…

Read more

7 months ago · 14 likes · 7 comments · Yuki Homes

and this 12 year old blogspot post goes into the 12 zones in more detail.

Note that certain places (e.g. the UNESCO world heritage site Iwami Ginzan near where I live) do have more demanding rules about what may be built and how, but the areas that those rules apply to are very limited and it is, AIUI, extremely hard for a location to get such a status so it only applies to a handful of historically relevant places.

Fundamentally the key in Japan is that you are allowed to do what you want on your land within some fairly broad guidelines and Japanese bureaucracy has no method of blocking your plans assuming you stay within the guidelines. Most of the guidelines are related to things like seismic survivability, which are the sorts of things that are good to have standards.

The disadvantage here is that you get ugly mishmashes of contrasting architectural styles and people make poor decisions regarding, say, the use of concrete that turns out to weather horribly, but the advantage is that a company like TSMC can be already breaking ground on Fab 3 in Kumamoto with Fab 1 operational for some years, while back in the USA TSMC’s Arizonan equivalent has just come on stream.

You can argue that maybe the blight of rusting, stained ~50 year old buildings is a sign that Japan should have had more regulation. But actually, as Gianni Simone, wrote recently the real problem here is restrictive laws protecting property rights for the heirs of flats in the buildings that the heirs don’t actually want.

Tokyo Calling

The slow death of a condominium

Dear readers, I love it when one of you suggests an idea for a new article. Both my latest piece on rental family services and the one you are about to read were born like that…

Read more

6 months ago · 34 likes · 16 comments · Gianni Simone

And, as he also explains, the government has just passed some laws that should allow for these buildings to be redeveloped.

Amazingly government looks like it might be passing a law that actually helps solve a significant problem.

Passing a Budget

OK technically Japan hasn’t passed the budget for the next fiscal year starting on April 1. That’s because, although the LDP has an overwhelming majority in the lower house, it doesn’t have one in the upper house and the various non-LDP parties refused to pass PM Takaichi’s proposed one. But it has passed (or is about to?) a temporary funding measure to fund the government for a couple of weeks and then after some actual debate and negotiation, a budget will be passed. I suspect if there hadn’t been a war to disrupt things, one might have passed before the deadline, though it might not because the Takaichi government massively shortened debate initially to try and force the process and this upset the opposition. This stopgap funding thing is, by the way the first time a budget has not passed in time for a decade. I’ll be honest I don’t know enough about the spending plans to know whether the opposition has legitimate gripes or whether it’s just injured pride (I expect a mix of both). Either way, one thing it shows is that the Japanese Parliament can actually have real debate about matters of substance.

So Japan is perfect?

No. Obviously not. There’s creeping DEI in the universities (and lots of questions about just how good they are anyway) and plenty of government waste and (again usually government) projects coming in late and well over budget - the Chuo Shinkansen comes to mind - but over all the way Japanese things go wrong is by doing too much, rather than not doing enough.

But on a sunny weekend at the end of march with the cherry blossom out, it comes pretty close

Subscribe now

I first published this a few years ago on Sarah’s blog and she has just republished it. This version is slightly updated because we are in 2026 not 2022

This essay makes the case that we are in the process of needing a non-governmental form of cash. This is because governments are busy tracking what we do with the money they print and because they are mostly printing a lot more of it than they should so its purchasing power is going down (inflation). These factors make it a poor store of value and hence an unreliable unit of account and medium of exchange.

Note: readers who didn’t read my essay reposted yesterday Money and Cake, should probably do so, and it won’t do any harm for those that did to reread as a refresher. It’s a light-hearted introduction to most of the key concepts of money.

Note. I am not a tax lawyer. Nor do I even play on on TV. Your local tax-gatherers and governments may object to you trying to implement your own non-inflationary cash. This should not necessarily stop you from doing so, but you should plan accordingly.

What’s the problem

The fundamental problem is the governments and large commercial organizations are unkeen on the concept of people exchanging clinking and folding cash money for products and services as opposed to balances of electrons being changed in various locations. There is a global push for going ‘cashless’. I.e. doing what The Register called “Pay by bonk” or possibly pay by QR code or other mechanism. In Sweden the switch away from cash is very well advanced; if I recall correctly, a majority of shops, cafes and the like no longer accept cash at all. Other countries in the developed and developing world are not far behind.

This isn’t just a government thing there are lots of large companies who also like the idea of consumers not using cash. Companies like it because they can take a cut of every transaction. It’s small on a per transaction basis (1-3% usually) and usually taken from the seller, but it adds up enormously over millions of users and billions of transactions. In addition they can often track spending habits and target ads and offers to entice consumers to buy more. In fact it seems to be popular with most of the “bureaucratic-commercial” complex everywhere, particularly in the land of Winnie the Flu, err West Taiwan.

In Canada Truckistan, the government declared an emergency which allowed it to tell banks to freeze arbitrary bank accounts without any evidence presented to a court or similar (and to let banks do the same thing on their own if they felt like it, with no fear of court order reprisal). So you can put your money into the bank but you may not be able to get it out. Of course they justified this as “temporary” and “only going after the evil REEEE bouncy castle protestors” but

1. this pretty much defines “thin end of the wedge” and

2. how does anyone appeal when they make a mistake (not if, when because they will)?  

Meanwhile in the land of the ‘free’, the ‘Let’s Go, Brandon!’ administration and their buddies in congress tried to get banks to report all transactions into and out of bank accounts with more than $600 in them.

This is, clearly, a potentially huge problem. It’s a problem two ways. First is the obvious government tracking one. Even if you assume the government is a wonderful organization staffed purely by the most competent and morally pure the chances of them making a mistake and bringing the full force of the law down on some poor housewife buying, say, fertilizer for her garden and a can of kerosene for emergency heating is higher than one might prefer. Given that in the world we live in the government is not staffed purely by ethical angels, the chances for abuse and error are very high indeed. And that doesn’t even get into the government criminalizing transactions that should be perfectly fine and so on.

Then there’s the problems of availability and fraud in a cashless society. If the power goes out you can’t buy a candle in a cashless society. You can’t buy anything if the internet goes down, if your financial provider has a problem and so on. We have seen numerous cases where Amazon’s AWS or Cloudflare have had little titsup.com moments and people as a result are unable to do simple shopping. Also if your provider is hacked (or your cash payment token is, or…) then you stand to lose more than just the cash in your wallet.

Fundamentally there’s a centralization and a corresponding lack of local personal control. There are benefits but there are plenty of drawbacks even before you worry that the government might abuse its knowledge of your financial life.

So keeping cash sounds like a good idea. However, there are potential downsides. Any replacement for cash needs to be able to avoid those downsides.

The most obvious is to do with the supply of it. In 2022 there was, as mentioned earlier, a certain upward spring in the prices of things these days if you bought them in dollars (or pounds or euros – though here in Japan, in yen, not so much yet - but, spoiler alert, we got some inflations too later).

And that is a problem for people who like the idea of cash. As Zimbabwe, Venezuela, various other South American nations, Israel and Weimar Germany all can attest to, cash is a complete disaster if the currency is suffering from hyperinflation. In all these cases cash quickly became useless. Indeed even in times and places where inflation is in the 5%-20% per year range (i.e. much of the 1970s and early 1980s for most of the world) cash is a poor store of value although it still works as a method of payment. Cash coins are slightly better than notes though because the metal will have a base value, so much so that at various times people have taken low denomination coins of various countries and melted them down to sell as refined metals.

Plus the other problem with cash as a strictly physical object is that it is hard to pay for things remotely using it. Which is why bearer bonds, letters of credit, cheques/checks, hawala and so on developed all the way back in the 17th/18th centuries (or about a millennium earlier in the case of hawala).

So to sum up, we have a requirement for something that can act as a reasonable store of value and unit of account (in Japan the yen has been amazingly stable for the last three decades but most other currencies have not, and in Japan the cause has been periods of no growth and deflation), be easily exchanged by buyers and sellers for good and services and yet not fall under the control of the government.

Subscribe now

Fundamental requirements

Our cash needs to be the following:

Non-inflationary (and non-deflationary).

A ton of things work better if the cost of something is predictable because the currency itself is stable. This doesn’t apply necessarily to every transaction, but the ability to be able to plan investments and calculate expected returns is made far, far easier when the unit of account remains constant over time. In most of the developed world people who are younger than about 40 are just learning what even fairly moderate 5-10% annual inflation does to your financial planning (something us older people remember from the 1970s/80s) but it is an important issue. It is worth noting that some currencies have historically been reliable (the pound sterling for the long 19th century up to ~1914, the US dollar for much of the twentieth) while others (the French Franc, the Italian Lira) have not.

As noted in the last 30 some years the Japanese Yen has been amazingly stable in terms of in country pricing. I first came to Tokyo in 1991 and prices for all sorts of things – from soft drinks from vending machines to train tickets to restaurant meals to property prices – have remained very much the same. Some have probably gone down a bit (property in some places, some restaurant deals) but generally speaking things have been stable. In fact in Japan, despite a few years of covidiocy caused inflation many things have still not changed price much. You can still get a can of coke from a vending machine for 110Y just as you could back in 1991

The key seems to be that the money supply should not expand (or contract) in relation to the underlying economy. As we (re)discovered with all the covidiocy money printing, when you increase the amount of money in circulation without increasing the economy to absorb it, you see prices rise. And, as we will undoubtedly see in the next few years, having got on the inflation train it is very tricky for a national economy/currency to get back off it again. [We know how to do it though. You raise interest rates and cut government spending (and government money printing) and it stops. But it is a painful adjustment as was discovered by all the countries in the 1980s that did it to end their 1970s stagflation.]

For our replacement money we want something that is a good proxy for the underlying economy. An example could be crude oil or the various refined products of it. It is true that the crude oil price varies considerably in dollars or other currencies but crude oil itself has tended to be produced (and consumed) in roughly proportionate amounts to the size of the global economy (see this graph). So if you had a currency of pints of crude then it would not be inflationary or deflationary (also it is worth noting that prior to the end of Bretton Woods and the 1970s oil shock, crude was remarkably consistent in US$ as this graph illustrates)

Independent and Decentralized

Independent means independent from governments. See above note about covidiocy and inflation for reasons why this is bad. It would be nice if it were hard to track by governments so they couldn’t tax you but that’s a nice-to-have, a must-have is that governments (or anyone else) cannot print more of it to suit their own needs.

It would be good if the currency had no single control point but could be replicated by anyone. That makes it hard to shut down if (when) governments get upset about it. They may manage to stamp out Alice’s Spondoolicks, but Bob’s Quatloos (which are readily convertible to/from Spondoolicks) can continue to be used just fine. Also if it turns out that Charlie has been sneakily coining extra Charlicrowns the only people who are impacted are those who have some. Owners of Spondoolicks, Quatloos or Dave’s Doubloons are just fine.

Decentralization also helps with scalability. While it is true that in the past a reference currency (the Thaler, the Pound Sterling, the US Dollar) was extremely helpful, in a world where information is easily passed everywhere and where we all have handy dandy massively powerful computers in our pockets, it’s quite not a problem for everyone to have the exchange rates of Doubloons to Quatloos, Quatloos to Spondoolicks and Spondoolicks to Charliecrowns and therefore calculate how many Doubloons make a Charliecrown.

Based on Something Physical

Actually this may be a nice to have, rather than a must have. But currencies based on actual gold (or wheat or cowrie shells…) have built in warning signs for when they are being inflated because you can do the sums and see that 1 billion ounces of gold probably doesn’t fit in that warehouse over there. Relatedly it is also possible to audit banks etc, and actually count the reserves of gold or wheat or cowrie shells. People can, of course, have fake bars of gold, plastic cowrie shells etc, but it is trickier to do compared to just modifying a few electrons here and there.

Hard to counterfeit or fake

I’d prefer impossible to counterfeit, but I’m prepared to go with hard. Most current cash can be counterfeited but the percentage of counterfeit cash is probably well under 1% of all cash. Again the point here is that someone (a government) cannot simply make more without investment in whatever backs the currency. A currency that can be faked is one that will soon be one that people lack trust in so we need to avoid that. One benefit of a decentralized system is that it should be possible to decouple and deprecate specific instances if they turn out to have been abused without losing trust in the entire currency. There are examples of this working in various places with existing currencies. For example in both Scotland and Hong Kong bank notes can be printed by a number of banks not just the government/central bank. If a particular bank gets into trouble then its notes may end up trading for less than face value.

Functional when the power is off

As you may have noticed significant parts of the world are learning the downsides to “green” energy with respect to its intermittency and general unreliability. The PRC shot itself in the foot by trying to boycott Australian coal thinking that domestic suppliers and other countries could provide it instead (narrator voice: but they couldn’t). The Europeans have shut most of their coal, some of their nukes and gone for a mix of renewables and gas and that became clearly a problem after the Ukraine war started, and is even more of one now with the Iran war. Parts of the US (California particularly) are doing the same.

If the power flickers on and off that will affect everything that requires electricity including the internet and services based off it, as well as smartphones and so on. It would be really nice if, when push comes to shove, you can pay for that gallon of fuel, roll of toilet paper or loaf of bread in some kind of off-line token. Ideally (see below re remotely transferable) there would be a way to print out tokens, use them and have the recipient scan them back in and destroy them.

Anonymous

We need to not be able to track the origins of cash or who pays whom with it. See above re: governmental oversight. But it isn’t just governments. Any number of large commercial organizations (e.g. Amazon, Walmart or your friendly local supermarket with their loyalty card) love the idea of tracking what you spend your money on so they can target ads and offers to entice you to spend more. All of this tracking has privacy implications. Given that you can’t trust institutions to either hold the data securely or use it ethically, anonymity is a really really good idea.

Remotely transferable

There has to be a way to pay a distant supplier for a product they will ship you that does not involve you, the supplier or a middleman trucking a physical lump of money from point A to point B. It must be noted that systems like Hawala have done this mostly successfully for a thousand years or more so you don’t need an internet (or even a telegram system) to do this. But – obviously – it will be better if it is possible to use the internet if it is available.

Convertible into fiat currencies when required

Paying for things in quatloos, spondoolicks or whatever is fine as long as the other party accepts them. Unfortunately in the near future people you want to pay for goods or services are likely to insist on dollars, euros, pounds, yen etc. so there ought to be a good way to swap quatloos for the required currency. There will also be a need to convert dollars, euros, pounds, yen etc. into quatloos. It is perfectly fine for the conversion process to be costly but it needs to exist.

Usable for small purchases

It seems obvious now but in the past less divisible coins have been a problem. If you get paid 10 spondoolicks a day you probably want to be able to buy things in small fractions of a spondoolick. So the currency needs to be able to support denominations that are small enough. Roughly speaking the current major currencies all have their smallest coin as something that is too small to buy a single thing these days (but you could 40 or so years ago), but where a small number of them can buy say, a single candy or something equally minor.

Why not Bitcoin? or $otherCryptocurrency

Current blockchain based cryptocurrencies don’t work well as quickly transferable cash and have a bunch of other issues. This long blog post covers a fair number of them (you may ignore the glowball wormening part).

A few of the issues: Transactions take a while to be confirmed (minutes to hours) and it is typically complicated to handle fractions of a coin. If you want 1000 widgets to be delivered next Tuesday then a cryptocurrency is fine. If you want to buy an ice-cream or a cup of coffee to eat/drink right now it is not so easy.

Bitcoin is designed to become progressively harder to mine. That’s not stable because it’s not tied to the size of the economy.

No current cryptocurrencies that I am aware of are designed to be turned into physical tokens.

Most cryptocurrencies (bitcoin particularly) are not anonymous either. They are pseudonymous which is similar but not as strong on the identity protection front. It is possible to track every bitcoin transaction back to the origin and people do that, which is why ransomware crooks no longer want bitcoins.

That’s not to say a new cash currency would not use cryptography – in fact it almost certainly will use it for something – but not as the source of the underlying value.

So what do we want?

We want some kind of hard to fake physical token that is based on possession of a physical thing of value.

That physical thing needs to be something that is universally understandable, and available. And it should be reasonably portable at need (think bars of bullion vs say ownership of a plot of land) even if most of the time it doesn’t get transported

That physical thing needs to be (roughly) tied to the size of the economy so that it won’t cause or be subject to inflation or deflation

There should be a mechanism to transfer promises of tokens to distant locations (“I promise to pay the bearer 1oz gold”).

Energy as the fundamental physical base

I think a good thing to base the currency on is energy. That could be just a measure in Joules or BTUs or it could be physical things that have known energy values such as gallons of gasoline or pounds of coal, or it could be both.

I do not claim that an energy based currency will work. But I think it meets most, if not all, of the criteria above. It’s readily understood, readily available, easily transportable and energy usage is tied to economic activity. Its also easy to understand exchanging it, so 1 gallon gasoline == N KwH of electricity == Y lbs of coal and so on.

Tokens

You can print any token you want, but if you add a serial number that includes some crypto validation key that will be helpful for identifying fakes. There probably should be an online “validate this token” thing that can be used to confirm that the serial number is legit.

But it’s you who is on the line to produce the energy if requested. So potentially bank runs could happen but that doesn’t necessarily destabilize the whole system. It would make sense if the people who backed the tokens were people with ownership of energy (e.g. the owners of fracking wells). But you could run an electricity generator and a hydroelectric dam or simply build a load of tanks and store gasoline (or diesel or..) in them if you wanted to. Or people could trust you when you promised to pay for them to refill their car at a gas station. Or… there are lots of options and lots of ways to store energy and use the store to back a currency.

Online resources would be useful to confirm the validity of non-local tokens and some might not be always accepted by everyone. Perhaps you would discover that only special dealers would accept weird Nova Scotian tidal power tokens and then only at a discount to their face energy value. On the other hand the local Stop’n’go tokens would be accepted by everyone and regularly exhangeable for BP tokens from the nearby town or Duke Energy ones….

Energy tokens are easily printable in small amounts and easily understandable. e.g. 1 KwH would be about a (US) dime. 1 fl oz of gasoline is of similar magnitude. Energy tokens are easily exchangeable from one type to another. Swapping a 1000 KwH of electricity for 20 gallons of gasoline (note I have NOT checked the exchange rate, this is an example) is dead easy and in fact you can have some guy with a gasoline powered generator do it very precisely for you.

Remote transfer

We know how to move energy. We know how to trade ownership in energy. All we need is a way to trust the exchange of physical token to virtual one. This is actually a place where the blockchain becomes useful because it allows everyone to confirm that X tokens have been deposited at store S and store S has transferred the value to user U. At some point later when store S receives a withdrawal request it can also log that transaction in the blockchain. Note that it

Works when the power is out

Indeed can be used to get the power back on because you can easily swap 20 gallon tokens for 20 gallons of gasoline right now. etc.

How to set up your cash system

See disclaimer at top of post. Talk to a practising tax lawyer. Pay them for their time. This will help you avoid steps that will bring the wrath of the government down on you. However in general, despite the disclaimer, this is how I would work (and these steps are not particularly ordered).

Start small and local. Use barter and IOUs as the basis of your system. “IOU 5 gallons of diesel” is unlikely to be considered a legal currency. Nor is “IOU 5 lbs of zucchini”. Or even the handing around of actual zucchini (or the traditional cigarettes and bottles of booze).

Code words. Squids, zuchinis, etc. have innocuous meanings that can mean they will be ignored when the authorities read your emails/text messages. Use archaic counting: score, dozen, gross. Use Roman numerals. “I want LIV dozen squids and a score zuchini” is less obviously cash that 648.20 $currencyname and so on.

The key is to keep local commerce working when the national/international stuff is not. So work on blockchains and electronic transfers later. For now having a trusted person at the other end (a la Hawala) will suffice.

If you need to make coins then washers that are engraved would work well. A modern CNC engraving tool could easily engrave the serial number of the coin if you wanted to do that along with a pattern that would be hard for a forger to replicate if he doesn’t have the pattern.

When you get to electronic banking, transfers etc. avoid all the things that have allowed crypto currencies to be insecure, most of which turn out to be stupid software bugs. Even large banks can write insecure code for this kind of thing so try not to need to write software of any sort. Or to have online portals etc. If you have to have an online thing open source the s/w and get someone to review it

Know your customer/trading partner. This reduces the chance of informants and criminals.

Other options

I like the idea of energy. But I’m sure there are other things. Even perhaps cryptobased things that I haven’t thought of. Also I’m sure there are things I’m missing. Don’t be shy. Comment away

Subscribe now

PS Original post here - https://accordingtohoyt.com/2022/02/27/we-need-non-inflationary-cash-by-francis-turner/

This post is a repost of something I wrote over a decade ago on the blog of my friend Sarah Hoyt. She1 and I are digging it up again (and a sequel to arrive tomorrow) because it’s still valid. I’ve made a couple of minor edits/footnotes in this one.

Money is one of those human inventions that is about as fundamental as the taming of fire. Every civilized society and many (possibly all) savage tribes of humans have some form of money. Societies that have attempted to do away with it have generally ended up both failing to and in the process killing people. Yet not many people understand money properly and, as a result, much suffering is brought into the world.

So what is money?

The glib answer from an economist’s text book is something like this:- money is a medium of exchange and a unit of account. Which is fine except that it’s got lots of words or more that one syllable and some of them have various meanings. So lets go back to basics (as the politician said to the archbishop).

The root idea of money is to store IOUs so that we can have distributed barter efficiently.2

Without money, if Alice wants Bob to build her a table then she needs to have something that Bob wants in exchange. It could be that she offers to cook him four cakes. Which is fine if Bob likes cake. Alice and Bob have agreed that one table is worth four cakes and both are happy. Alice has her table and Bob has his cake.

Except that Bob doesn’t like cake but does like beer. But maybe he knows that his mate Charlie the brewer needs cakes so Bob swaps those four cakes for a keg of beer from Charlie. Charlie in turn feeds two of the cakes to his three children and offers the other two to Daphne in exchange for her cutting his hair and the hair of his cake-stuffed offspring. Daphne now has enough cake that she can stuff herself while watching some mindless romantic drama on TV. Everybody is happy (well apart from Bob because he’s got a hangover but he was happy earlier), and Alice has her table.

That second chain is where money steps in. It was only because Charlie heard Bob moaning about being offered payment in cake and then offering the swap that things worked out. Otherwise Alice would have had to offer something else to Bob or would have to have found Charlie directly and offered to trade beer for cake. And even then Daphne might have been left out except that she heard Charlie’s children griping about how they hated cake, particularly when they had to ingest it through greasy hair.

In an isolated small village (or tribe) it is possible to do this kind of barter chain with some success. It may take a while for Alice, Bob, Charlie and Daphne to figure out the relative value of cake, beer, tables and haircuts and who will do what in exchange for what, but it could be done in an afternoon in the village square. It would have been much more efficient if everyone sold things for some known unit of money (say a spondoolick). Alice sells her cakes at one spondoolick each, Bob sells tables for four spondoolicks, Charlie sells beer kegs for four spondoolicks and Daphne cuts hair for half a spondoolick a head. Now Alice simply goes to Bob and offers him 4 spondoolicks for a table. Bob takes the spondoolicks and hands them to Charlie for the beer. Charlie pays two spondoolicks to Daphne so that she cuts his hair and his children’s hair and buys two cakes from Alice for another two spondoolicks. Finally Daphne also buys two cakes from Alice for a spondoolick each and is much happier because her scissors aren’t all cakey.

Spondoolicks as a medium of exchange make things much better. Not only can we now have chains of transactions that also involve Edwin, Fiona, Gerald, Henrietta and Ian; it is quite possible for Fiona to be in the village on the other side of the hill and for Henrietta, Gerald and Ian to be in the local market town or even further away. So long as the two villages and the market town agree (more or less) on the value of a spondoolick Alice can sell her cakes in the market town on the table Bob built to Fiona and everyone is happy.

Trade occurs when the two villages and the market town disagree on the relative value of things as priced in spondoolicks. The only reason Alice takes a day to go to the market town and sell her cake is because in town she can sell them for 2 spondoolicks each, especially to idiots like Fiona who have more spondoolicks than sense. On the other hand Charlie used to hate the town because Edwin, the brewer there, sells his beer at a discount, especially if you buy 10 kegs at a time. But then he realized he could buy 10 kegs for 6 spondoolicks and sell them back in the village for a spondoolick each and even with the spondoolicks he had to pay Bob for loan of his cart he can sell beer at a profit and have more time with his children because he doesn’t need to actually brew beer any more. And next week he suggests to Alice that rather than she struggling to market with all her cakes and that stupid table, why doesn’t she just sell them to him for one and a half spondoolicks each and let him deal with the table and dogooders like Gerald who try to tell Fiona that two spondoolicks a cake is a rip off.

And this is where we get to the concept of the merchant and the idea of money as a store of value. The merchant (i.e. Charlie) acts as the middle man between people in Aliceham who need something (e.g. beer) and people in Brewersville who have too much of it. The merchant buys the excess of beer from Brewerville at a price which is lower than he can sell it at Aliceham and then in reverse takes all the extra cake from Aliceham and sells it in Brewersville.

Later, as the fame of Alice’s cakes spreads far and wide, Charlie and Ian from the land Faraway come to an agreement so that Ian buys most of the cakes from Charlie at a price of 20 spondoolicks per dozen and transports them to Faraway and sells them to would be gourmets at 10 bongoes a cake. He uses those bongoes to buy tools, spices and hops which he brings back to Brewersville market and sells to Charlie (who then resells the spices to Alice and the hammer to Bob) and Edwin the brewer and so on. As a result his initial outlay of 20 spondoolicks turns into 30 spondoolicks. Again he buys a dozen cakes but that leaves him with a profit of 10 spondoolicks which he leaves with Gerald so that next time he comes back to Brewersville he can buy two dozen cakes. Similarly when he gets back to Faraway he can save up his bongoes and buy even more spices, tools and so on.

Of course it isn’t totally clear what the exchange rate of bongoes to spondoolicks is. One way to look at it would be to use a cake-index and say that since Ian buys cakes for 1 ⅔ spondoolicks and sells them for 10 bongoes then the rate is 10 bongoes == 1 ⅔ spondoolicks (i.e. 1 spondoolick = 6 bongoes). But that ignores the fact that Ian makes a profit of 10 spondoolicks on his sale of Faraway goods in Brewersville. So perhaps a better way to get to the equal value thing would be to look at the price of hops (etc.) in Bongoes and Spondoolicks as well and then take the verage of the two. In fact probably 1 spondoolick is worth between 4 and 5 bongoes and taking a look at the retail price of cakes in both Brewersvile (2 spondoolicks each) and Faraway (10 bongoes each) 5 bongoes to 1 spondoolick sounds about right,

Of course when Daphne wants to travel to Faraway, Charlie, Ian or someone like them will charge her a commission on the trade so she only gets 4 bongoes per spondoolick and when Jessica comes to Brewersville from Faraway she finds that her bongoes are only worth 1/6 of a spondoolick when she tries to exchange them – unless of course she meets Daphne. And that of course is the point. A currency is worth what you are willing to exchange for it and the person on the other end of the deal has to agree.

We will note that in this example we haven’t yet said what a spondoolick is (or a bongo for that matter). It could be a lump of metal, seashells, leaves, pieces of paper with the words “one spondoolick” written on them or some electrons or magnets sitting in a computer somewhere. The critical thing is that we trust that a spondoolick today will be worth (more or less – famines and other major events excepted) the same tomorrow and next month. Related to that we have to be able to be sure that someone (e.g. Charlie) doesn’t produce a few extra spondoolicks now and again because his kids need a haircut and he doesn’t have any spare right now.

Historically lumps of gold and silver (and copper3 etc.) of known weight/purity have been a popular choice for what to make a currency from, but the temptation for someone to use slightly less precious metal than there should be (or even none at all) has also been popular. Similar issues have plagued every other way to keep track of currency though electrons (in the form of bitcoins and similar) have generally proven to be less vulnerable although they have proven to be relatively easy to steal or lose. Either way one critical thing about money is that once we agree on what it is and what it is worth (more or less) a spondoolick from Alice is just as good to Daphne as one from Jessica, you can trust them equally and financially we don’t care about their past life (except for when it turns out the money is fake or substandard). Moreover paying with them is anonymous or can be. Apart from people seeing him sneak in the door, no one can tell that Charlie has a beer at Edwin’s place every market day. As long as he gets paid there’s no need for Edwin to care who it is he is serving and likewise no need for Charlie to care which pub he goes to in Brewersville because only one will accept his spondoolicks.

Something else we haven’t mentioned yet is “Government”. While, historically, governments have generally had a big say in money that is mostly because governments are what we trust to stop the Charlies of the world from getting away with counterfeiting money – both by setting standards for what money is and by punishing the fakers when caught. Although of course governments have also historically done an absolutely bang up job in debasing currencies themselves and a cynic might say that the reason why governments go after private counterfeiters and the like is that they hate competition. Governments are not required to do the whole thing for money to work.

A number of countries (e.g. Hong Kong and Scotland) allow banks to print banknotes themselves and while this can cause problems when the bank gets in trouble or when someone tries to use, say, a Scottish banknote to buy a round of beers in London (though it probably works in Carlisle or Newcastle), it is generally not a major issue. Indeed when you consider the use of US dollars in countries like Iran or Zimbabwe, sometimes the fact that the local government has nothing to do with the currency is a major plus.

The key thing about money is that it only works when there is trust. When trading both parties to a transaction have to agree than the monetary object in question is genuine and worth an agreed amount of stuff. Similarly, when storing money somewhere, the person storing the money has to trust that the place he is storing it is safe and will give it back to him when he needs it. When we all stick our money in a bank and then, later, decide we don’t trust the bank that leads to a bank run and it gets nasty when (as is often the case) the general lack of trust in the bank turns out to be well founded. Stopping bank runs from turning nasty is, actually, one of the things that we probably do need government for. The bank runs and related shenanigans in the bitcoin/crypto-currency world have been pretty catastrophic4.

Subscribe now

Mention of banks leads us to the concept of loans and interest. Again this isn’t anything complex. Alice cooks Katherine a cake today because she happens to have all the ingredients and a week from now Katherine cooks one for Alice. Effectively Alice loaned a cake to Katherine for a week. Assuming the cakes were the same then there was no interest on the deal. If Katherine’s cake was bigger than Alice’s then difference in size is interest on the loan (if it was smaller then the interest was negative and Alice makes a note to never bake a cake for Katherine ever again). Interest is a way of measuring the different value of money over time. Katherine really needed a cake today but didn’t have any flour and the shops were shut. Next Sunday she’ll have loads of flour so she can bake a bigger cake for Alice than the one Alice cooked for her.

We can extend the example further – say Alice cooks a big cake for Katherine and in return Katherine gives Alice a sticky bun every day for a week. If 6 sticky buns used the same ingredients as a cake then that 7th sticky bun is Alice’s interest and the profit on the deal. If Alice did this a lot then (apart from becoming overweight and sick of sticky buns) she’d get called a loan shark or worse because 1 cake in 6 is 16.7% and 16.7% interest for a week works out at something like 6000% on an annual basis. Of course if she got paid one extra bun a month it would still be steep but much more reasonable (~30% annual rate).

Now lets assume lots of people deposit excess cakes with Alice and agree that they’ll take back either sticky buns or cakes when they feel a touch peckish. If Alice now had 30 cakes which she loaned out to Katherine and 29 other people, she’d get just a single sticky bun every day but in 7 months she’d have a 5 extra cakes (or cake equivalents – remember 1 cake = 6 buns and 30 loans of a cake for 7 buns gives you 30 extra buns). Nice work if you can get it especially if the other cake owners who gave Alice their cake only ever want a sticky bun at a time and no more than one of them wants a bun each day. It goes a bit wrong however if one of the other cake owners (Louise) shows up with no warning on Saturday a month down the road and demands her entire cake back now because it’s her child’s birthday party. Unless Alice has a spare cake (or can make one quickly) she can’t pay Louise back as she promised which will make her very unhappy – and of course she’ll tell the mothers of the other 5 kids who were invited to her birthday party and they’ll start wondering what happened to their cakes and on Monday Alice will have demands for 5 more cakes which she won’t have either and that news will spread and so on. That is what we call a bank run and it probably results Alice getting a load of sticky buns in places she won’t enjoy.

Now Alice could avoid this situation by not loaning out all the cakes she got (perhaps she gets 40 cakes but only loans out 30) or by writing a contract that says that you can’t get your cake for 6 months or that you have to wait 6 days after you request a cake for it to be provided to you or some combination. For example Alice might say that if you want your cake back now you get nothing, but if you leave it with her for a year you get a cake and a sticky bun. In other words Alice is paying you interest for leaving the cake with her for a fixed period of time and she might make a rule that you can only withdraw one bun a day unless you give prior notice. With these sorts of rules and with 10 spare cakes sitting in her freezer Alice can be confident that Louise will be satisfied when she wants her cake back. Alice will get a decent surplus of sticky buns even assuming that some people decide that they’d rather have a cake and a sticky bun next year than just a cake today.

But it also assumes that all the 30 people Alice loans cakes to give her a bun a month. If Maria doesn’t pay after 3 months and Nina doesn’t pay at all then Alice is out a cake and a half. If that’s all it’s not too serious. The other 28 people will pay their full amounts so instead of ending up with a surplus of 30 sticky buns Alice ends up with 19 (7 lost from Nina + 4 from Maria). But if she gets it wrong and another three are like Nina then that 19 bun profit will turn into a loss of 2 buns.

Of course Alice could solve this by paying Bob a couple of sticky buns to go around to Maria and Nina and stand over them menacingly while they cook their buns (a 2 bun loss is far less bad than an 11 bun loss) and if Bob does it right to Nina when she hasn’t paid for 2 months then maybe Maria hears about how nasty it all was and manages to pay off her loan even though she really wanted those sticky buns herself. Either way Nina and, possibly, Maria now have a terrible credit rating – no one will lend them a cake again unless they show they really have turned over a new leaf and they probably have to pay 8 sticky buns back instead of 7 to cover the risk that they don’t pay back any sticky buns unless (or even if) Bob goes around and (threatens to) beat them up..

Right now we know what money is, how it works, how loans and interest work and things like that. This is all a basis for what we need in a possible new currency. Our new currency needs to be trusted, consistent in value, store-able and able to be used anonymously. Sounds easy….

Subscribe now

Comment here, or on Sarah’s new post (or both, and if both feel free to argue the opposite on each just for fun)

Back in the early 2000s there was the Axis of Evil, which was variously Iran, Iraq, Syria, Libya, North Korea, Cuba and maybe Venezuela. Different people may have added or removed different members of that group, but roughly speaking that was the group.

Now we’re in the 2020s and at the end of 2025 we could say that Syria, Libya and Iraq were no longer members of the Axis. Indeed they were (are) semi-failed states which aren’t great but are also unable to project any influence abroad.

In the Western Hemisphere Venezuela just got an attitude adjustment and Cuba appears to be in the process of getting one. Both of those being due to President Trump’s robust enforcement of sanctions and the like that previous regimes waffled over.

Iran is… in the desperately clinging onto power stage, and I’ll add more about that below, but its abilities to project force abroad are lessening day by day even as it tries to hold the world hostage. Its regional proxies - the Houthis, Hezbollah and Hamass are so far sitting it out, being taken out and already taken out respectively - and it has zero friends left in the region after it attacked almost all of them.

That leaves the sole survivor. The Democratic People’s Republic of Korea.

North Korea

North Korea is now looking pretty isolated. It has been isolated before but this time the Iran situation means that not only is it alone, but no one is paying it much attention. That’s not good because North Korea has no domestic oil/gas and little foreign currency to pay for imports, which have just become very expensive and in short supply. North Korea can also observe that President Trump’s USA is willing to tolerate global or regional disruptions if that results in a regime it disagrees with falling.

I’m not sure how little rocket man will react to current events, but I think we should be prepared for, at the very least, demonstrations of North Korean weapons and threats towards its southern neighbor. I wonder just how much fertilizer N Korea has stocked, how much it planned to buy or make using foreign oil/gas and what a lack of it will do for the harvest this year. It seems possible that by the end of this year North Korea may well be in a world of hurt. Ironically that hurt will have been caused primarily by its fellow Axis of Evil member, Iran.

It is also possible that no one is going to care what happens in N Korea because the world will have enough other problems. Even if the Strait of Hormuz blockade is cleared tomorrow the damage and delay means the world will be lacking several weeks worth of normal oil and gas consumption. If the war escalates to cause major damage to more Persian Gulf oil and gas resources then the world as a whole may be forced to enter a major diet of fossil fuel consumption for a prolonged period which means no one will have attention to spare for a N Korean famine and any related unrest. It will be interesting to see if either Russia or the PRC are interested in offering assistance. I suspect the Chicoms won’t, but Russia may find it worth trading some oil and fertilizer for troops to die in Ukraine and ammunition supplies to use there.

And then there’s the external threat. North Korea’s leadership is witnessing in real time just how poor Iranian operated versions of its air defenses are against US aircraft and missiles. OSINT analysts on X and elsewhere are posting pictures of secure bunkers with nice holes in the roof and noting that US bunker buster bombs appear to be able to penetrate pretty thick mountains. Possibly more concerning is that the Persian Gulf states are successfully using South Korean air defense systems to protect locations. That implies that when South Koreans use them themselves they will also be able to shoot down a lot of North Korean attacks. South Korea has a lot of these systems, which implies that the “Seoul as hostage puppy” strategy may no longer be quite as viable as previously thought.

Now it is true that, unlike, Iran and Hormuz, the Nork plan for Seoul includes lots and lots of artillery, but that’s so well known that no one seriously doubts that the South has counter battery plans to take out as much Nork artillery as possible as fast as possible when an attack occurs. So yes, each gun gets off one round1, which will definitely cause significant damage to Seoul, but if it is then destroyed before it can fire the second or third the damage is survivable and the expected number of fatalities are in the low thousands. This substack and associated academic paper dig in to the likely damage

Blue Blaze

How Dangerous is North Korea's Artillery Threat to Seoul?

For decades, North Korean leaders have vowed to turn Seoul into a “sea of fire,” threatening destruction with the thousands of artillery pieces and rocket launchers it has positioned just across the Demilitarized Zone—only 30 miles from the South Korean capital. Seoul’s vulnerability to North Korean bomba…

Read more

10 months ago · 2 likes · 2 comments · Nicholas Anderson, Daryl Press, and Jennifer Lind

My belief is that, based on what we’ve seen in the Middle East this year, that analysis may actually overstate the likely damage to Seoul and South Korea. In summary, rather than killing millions and making Seoul a ruin of uninhabitable skyscrapers as the Norks and their sympathizers have loudly claimed is to be expected if the Korean War resumes, damage is more likely to be not dissimilar to a couple of 9/11 attacks in total. Now that’s not good, 2000-4000 dead and several thousand injured is a serious attack, but it’s not close to nation ending. Neither is repairable damage to a few hundred or thousand buildings. Yes it’s trillions of Won (billions of US$) to repair but the South Korean economy is large enough that this is a relatively minor blip.

More to the point the damage isn’t going to be one way. While a surprise Nork attack will clearly do a fair amount of damage to Seoul and South Korea in general, it will pale in comparison to the response, which North Korea will be unable to stop. Not only would the total destruction be far more, based on what we are seeing in Iran, the damage would be disproportionately heavy on actual regime critical infrastructure rather that spread over the population in general. Little Rocket Man might manage to avoid the Ayatollah “should have been a Zoom meeting” decapitation strike, but it is still likely that most of the top echelons on the Nork leadership would be killed along with the destruction of the command and control infrastructure of the regime (and much of its military). In other words the blackmail/hostage strategy that has allowed North Korea to survive is no longer as viable as it once was.

In other related bad news for Little Rocket Man, his nuclear missiles will probably be shot down if they launch successfully. Japan, the US and S Korea have all tracked Nork missile launches from very shortly after launch even though none have yet launched counter missile strikes. Based on Iranian missile performance against Israel’s Iron Dome and the known existence of THAAD batteries in the East Asia region, it seems likely that any missile launch that looked like a nuclear threat would be destroyed well before it reached its target.

North Korea’s other survival strategy is to pull in its larger neighbors, Russia and West Taiwan, as it did in the original Korean war. In the case of Russia that’s not going to work because Russia doesn’t have any resources it can put towards assisting North Korea. Russia has burned through its Soviet-era stocks of weaponry and while oil and gas exports are keeping the country afloat, its military has been shown to be a hollow shell by the war in Ukraine. Indeed it has been importing (and bitching about the quality of) Nork artillery shells because it can’t make enough of its own. West Taiwan is different. It does have resources it can divert to help North Korea and, of course, it can choose to take advantage of a response to a Nork attack to try taking Taiwan, which will split the attention of the US and Japan. On the other hand even assisting the Norks, let alone attacking Taiwan, is a potential “start WW3” gamble and as I’ve written before the Chicom leadership is risk averse.

In summary, the campaign against Iran has just showed North Korea that the cards it holds are not as solid a winning hand as it may have thought. Combine that with supply issues and North Korea may be in for a tough year. Not quite as bad as Cuba, but if the Iran war wraps up in fairly short order as I expect it to then the DPRK rulers will be well advised to have some behind the scenes chats with President Trump lest he decide to go for removing the final Axis of Evil regime this year too. But it is quite possible that Little Rocket Man will decide he needs to do some kind of “Senpai watch me” demonstration first, and not unlikely that that demonstration provokes a response he didn’t anticipate.

Update: Interestingly Little Rocket Man is hosting Belarusian President Alexander Lukashenko in Pyongyang yesterday. There seems to be no published agenda that I can see but I can confidently predict two of the topics under discussion - Ukraine and Iran

Subscribe now

Back to Iran

The news media seems to think the war over Iran is a quagmire, disaster etc. that will cause another 1970s style oil shock etc.

As I wrote in my previous post2 that is unlikely to be a global phenomenon, however some countries that failed to prepare for the obvious danger of an oil shortage (Australia3, most of Europe to name a few) are going to be in some kind of trouble. Moreover there are already shortages seen in various chemical production chains that relied heavily on Arabian oil and gas as a start point (nitrogen and phosphate fertilizers, ethylene and more) as well as looming issues for industrial users of helium which the West Taiwanese chip industry will definitely not appreciate. There will no doubt be other shortages that will hit at odd times and locations because international supply chains turn out to be highly vulnerable to a 10-20% decrease in global oil and gas availability. Some of these, the helium one for example, are now baked in because there is no way to recover the lost Qatari helium, others are still in process as plants still have some downstream supplies stockpiled but may not have enough to tide them over until supplies can resume.

However just as the Covidiocy caused upset but failed to kill off large chunks of humanity as promised threatened, any oil shock is likely to have fewer dire effects than the gloomsters predict. Though that’s not saying that countries without decent stockpiles of oil/gas are not going to suffer economically.

Either way, we’re about three quarters of the way through the original 1 month/5 weeks timeline and the Mullahs have not yet failed. Indeed there are reports that the IRGC are charging tankers $2m for safe passage though the Strait of Hormuz, which means that suppression of IRGC assets in the area has not been as extensive as one would have hoped for. [Aside, I am curious where the payments are going, and I would expect them to be tracked and probably seized by any Western banks they may pass through ]

Similarly, the numbers of missiles and drones being launched has not dropped further - in fact it went up slightly in the last couple of days.

However I am fairly sure that Iran’s missile and drone factories are now inoperable and Israel just sank the Caspian sea fleet so drone (and other munition) supplies from Russia via the Caspian are probably cut off. I think that means that although Iran may still have stockpiles that have not yet been neutralized, it is not going to get any resupply.

In urban Iran things are looking a little more dicey for the regime. There are reports (e.g. this and this) of the local population starting to shoot back at the Basij/IRGC using weapons the Basij abandoned at checkpoints when they were droned. If this becomes widespread then the regime is doomed. The top are being systematically eliminated by strikes and the bottom is threatened by the population. Because yes they can shoot job lots but then they have to go somewhere to sleep/eat etc. and Israeli drones are hitting large concentrations of them. Individual Basij are extremely vulnerable to sniper/mob action if they are not in a group, if they are in a group they are vulnerable to being droned. It’s a tough choice and really couldn’t happen to more deserving scumbags. We haven’t seen the overthrow of as murderous a regime as the mullahs for a long time, if ever, so I can’t predict how things will pan out but I will not be surprised if we see defections at the bottom and then mass bloodshed as the population and defectors fight the true believers. I don’t know how long it will take to get there, but I doubt the regime has another month left assuming the pressure is kept up on it. Various commenters have observed that the Mullahs can spin survival as victory by “standing up to the US” but that only works if they can still project force to disrupt international shipping and so on. I do not expect that to be considered acceptable by Israel or the US so I expect, despite the mutterings of ceasefire, that the war will continue until they cannot project force. On that note and to ratchet up the pressure, the MEFs are due in the region any day now. I think it is unlikely there will be long term boots on Iranian soil, but raids in force and temporary occupation of Qeshm Island and other places in and around Bandar Abbas are quite possible. Then, as I suggested in my previous article, the toll booth will be operated by America not the IRGC

Subscribe now

PS I’ve been reading this substack which I find stimulating in its analysis.

Velina Tchakarova

THE GLOBAL SYSTEM RUPTURE

Last week I called this the Uncharted Water scenario. I was right about the systemic risk exposure. I was wrong about the scale and scope. What we are watching as of Day 22 is not a scenario. It is a rupture in progress…

Read more

5 months ago · 43 likes · 8 comments · Velina Tchakarova

We agree that Iran must hold Hormuz to survive. In other places we disagree.

For example, while it is true that, unlike its GCC neighbors, Iran does not get much of its water from desalination, this is not a strength for Iran as it is in the midst of a prolonged drought with water supplies for major cities like Tehran at extremely low levels. Because Iran has limited desalination capability and no water pipelines over the mountains, the water supply to many of its major cities is extremely vulnerable to disruption and simply running out. Thus, while it is quite true that an Iranian strike on neighboring desalination plans would cause mass death from dehydration, the exact same effect could be achieved in Iran by destroying the dams and pipelines that supply water to its cities.

I also disagree that the PRC is benefiting from the war. Russia is, in the short term, because of the spike in oil and gas prices, the PRC is not. In fact it is subject to the same oil shortage that the rest of Asia faces despite still getting some Iranian crude.

2

L'Ombre de l'Olivier

Trumping Everyone on Iran

The general destruction of the mullahcracy of Iran is proceeding well militarily, with almost every major prewar leader dead and/or seen only on cardboard. The US has said they have moved on to destroying the factories that make the missiles and drones having pretty much defanged the current inventory of missiles…

Read more

5 months ago · 60 likes · 35 comments · Francis Turner

3

Mentioned previously, but more here

Craig Peacock | No Easy Way Out

A RESOURCE SUPERPOWER WITHOUT RESERVES

Australia and Japan. The difference in oil reserves policy is startling…

Read more

5 months ago · 2 likes · 3 comments · Craig Peacock

and here

Dystopian Down Under

Energy lockdowns Down Under? Call for 'Covid-style' rules as fuel crisis escalates

The spectre of lockdowns 2.0 looms as an Australian Premier has called for “Covid-style” rules to manage the escalating fuel crisis sparked by the war in Iran…

Read more

5 months ago · 67 likes · 19 comments · Rebekah Barnett

The general destruction of the mullahcracy of Iran is proceeding well militarily1, with almost every major prewar leader dead and/or seen only on cardboard. The US has said they have moved on to destroying the factories that make the missiles and drones having pretty much defanged the current inventory of missiles.

There’s probably a degree of over confidence in those statements in that Iran is still firing a few missiles but generally speaking it seems to be correct - the numbers have dropped enormously. While that doesn’t mean the mullahs are all dead and/or impotent, the thousands of missiles they had before February 28th have declined by at least an order of magnitude, with many of the survivors trapped in underground bunkers that the US/Israel have drones monitoring for traffic in and out.

In fact, the US and Israel have got full control of Iranian airspace and can hit anything they want with pretty much any weapon they want, as has been noted by the way they have drones taking out individual Basij checkpoints or buildings. Sure some US/Israeli drones get shot down, but that’s why you fly unmanned drones not manned planes for such missions. The largest problem seems to be that Iran is so big there are parts (e.g. the NE corner around Mashhad) which have few hits because of range limitations by the shorter ranged US/Israeli aircraft/drones.

What that means is that, in general, the missile threat from Iran is over. At least in regards to defended locations like airports, military bases and the like. Which in turn means all this hyped up worry about the gulf states or Israel being about to run out of expensive Patriot/THAAD etc. missiles was either deliberate disinformation by Iran/allies or idiots spouting off (or both).

However missiles are not the only way the regime can attack. There are also drones. The Iranian drones sold to/built under license by Russia have been an immense problem for Ukraine and those same drones are problem for Iran’s neighbors too (see chart above) because the number of attacks using those have taken longer to decrease.

That’s only part of the problem. Many of the successful hits on buildings and bases seem to have been from drone attacks because, for reasons, drones are hard to defend against if you haven’t planned on defending against them. Fabian Hoffman explains in some detail why

Missile Matters — with Fabian Hoffmann

Why Cheap Drones Destroy Expensive Things - And How to Deal with Them.

Picture: U.S. Central Command…

Read more

5 months ago · 24 likes · 3 comments · Fabian Hoffmann

Now, obviously, there are defences against Iranian drones and once you deploy the right detectors they are fairly easy to spot and then remove. And in fact it seems like one way or another the strikes are generally being handled, but there are still a few leakers every day. However it seems to me that a lot of the footage of drone attacks seem to be of drones being hit by close in fire near their end targets rather than drones actually making a hit. Because the take out is so close it can be unclear whether they struck the intended target or not.

But as is quite clear from Ukraine, drone attacks against fixed infrastructure are eminently defendable once you have all the right bits. It sounds like Ukraine is willingly accepting large amounts of Persian Gulf state cash in exchange for hardware and trainers. So drones against desalination plants, oil refineries etc. are probably not going to succeed often enough that they cause long term problems. In fact there is one very obvious difference between Iran’s drone warfare and Russia’s (or Ukraine’s), which is that Iran cannot put together a swarm of drones that can overwhelm defences for a specific target. Instead they send, AFAICT, either singletons or just a handful at a time to a particular target and that, in turn, makes them easy to shoot down. So the drone threat is also over-hyped, in general.

The Dire Strait

However drone, missile and other attacks against shipping are a whole other story. In fact the Strait of Hormuz has been effectively closed to shipping for the whole of March. Perun has a good hour long video on the topic

The key here is that there’s an asymmetric knowledge/risk problem here where it is easier to close the thing than to open it. Ship owners and crews are well aware that their ships are large, relatively slow moving targets passing within visual distance of Iran and thus risk an attack by a missile, drone, boat… that they can’t dodge or defend against.

It is true that there are ships passing through, but relatively few compared to the numbers normally observed. In order for shipping to resume at scale, the owners and crews need to be sure that the chances of a ship being hit is down in the 0.00something% range.

[Brief probability digression. Transits of the Strait of Hormuz are normally in the 100 vessel a day range. That means that if there is a 1% chance of a vessel being hit, then on average one vessel will be hit a day. 0.01% puts it at one vessel being hit every three months. Owner, insurers, and crews can probably handle a sub 0.01% chance but that is about the upper limit]

That is a problem, because significant fractions of the global supply of oil, LNG and fertilizer pass through the Strait. People like Phillips P O’Brien think that this means that Iran is in one way winning its war while losing its war with the US.

It’s not a bad summery of where we are and makes some good points. However, and this is very important, it misses one key factor.

That factor is that closure of the Strait of Hormuz is not, directly, a problem for either Israel or the USA. The US (with the exception of the idiots in California) has no need to import oil, gas or fertilizer and Israel’s imports of all three are sufficiently small that the US can trivially supply them in the event that no other supplier is available.

Subscribe now

It is a major problem for several other nations though. First, many of the gulf nations depend on it not just for exports of oil etc. but for imports of things like… food, as well as clothing, Chinesium consumer electronics and the like. Second many European and Asian nations really need that oil, gas and fertilizer.

Some nations (e.g. Japan) have significant stockpiles and so probably can cope with several months of shortage, but even Japan will eventually run out. Many other nations (e.g. Pakistan, Bangladesh, India and probably West Taiwan) have considerably less stockpiled and so feel rather more urgency. Europe should do too for that matter.

Numerous talking heads have ridiculed the US effort claiming that the US made no plans to stop the closure of the Strait and suggesting that President Trump, Defense Secretary Hegseth and so on were taken by surprise that the Iranians would close the Strait. Those talking heads are almost certainly wrong. What they miss is that under Trump 2.0 the US is not interested in acting as the free world’s volunteer coast guard any more. Once you realize that, you see that Iran is fighting to provide President Trump with a lever he can wield against the rest of the world. The mullahs, BTW, have no choice here. They either surrender to Trump, accept regime crippling to the level of regime change, or they try to close the strait and make the rest of the world force the US/Israel to back down. Their problem is, I think, that Trump has planned to take advantage of precisely this disruption and he won’t be backing down.

Consider. All the way back in Trump 1.0, the President made it clear he was unhappy with the way he saw other countries free loading off the back of the US. He and his 2.0 cabinet - e.g. JD Vance and Marco Rubio - have made many speeches stating quite clearly that under Trump (and presumably successor MAGA-style presidents) the US will only respect allies that actually meaningfully contribute. The Strait of Hormuz closure is, as Hans Mahncke points out, where the US tells its freeloading allies and non-allies that the freeloading stops

The Strait of Hormuz issue is the perfect “allies” test. The United States has very little direct interest in the strait, as almost none of its ships or energy flows pass through there. So before the United States starts doing all the heavy lifting, it is the ideal moment to ask those who actually benefit from the strait being open what they are willing to contribute.

The fact that they all said nothing is proof of what Trump has been saying all along, that supposed “allies” are really just freeloaders. So this was never about whether the Europeans could actually help with a handful of dilapidated frigates. It was a test, and they all failed.

What Hans doesn’t mention, that Perun covers in his video, is that the gulf states have quite a lot of naval vessels that would be suitable for escort duty if they are actually as capable as they are supposed to be. Forcing the gulf states to be a bit more aggressive is, probably, one of the goals of this exercise. Either the gulf states step up and use their expensive naval toys or they quietly admit that they blew shipyards of money on kit that they can’t operate properly in actual combat situations. My bet is that they’ll do option 2, though I could be wrong.

They can then reflect on how that differs from, to pick a neighboring country completely not at random, Israel which can not just operate advanced technical weapons but develop them too. I’m not entirely sure how this works, but I suspect the Abraham Accords are going to get more signatories and some kind of mutual defence treaty that allows Israel to protect gulf state shipping and infrastructure, by, for example, having its aircraft and drones bang next door to the US ones in the various states. Certain temporary residents of certain gulf states (coff Hamass coff Qatar coff) may have some fatal accidents as that comes to pass.

Then there’s the oil, gas etc. importing countries.

Europe, Who Cares?

In Europe, France may get a partial credit by at least having a naval presence in more or less the right place, though statements like this suggest not. It is possible that Italy and Greece and a couple of other countries will too, but Germany and the Yookay, to pick two obvious examples, have just shown to the entire world that they have no real navy.

However in the grand scheme of things, despite their self belief, Europe is a side show. European nations are in most respects no more important or relevant than, say South American ones and over the last quarter century they have intentionally embraced idiotic deindustrialization policies that mean they are unable to respond to entirely predictable world events. That’s not so true of Poland or Eastern Europe in general but the fact that four years in to the Russia Ukraine war, European nations have failed to seriously rearm or supply armaments to Ukraine. And then there is the whole anti-Semitism thing and the way that they have allowed themselves to be colonized by Islam. What the current Iran War shows is that Europe is unable project force in a meaningful manner and also, apparently, unable to do anything other than whine when events occur. As wretchard Xeets:

In retrospect the Global World was excessively dependent on chokepoints controlled by terror groups and theocracies for decades. These narrow points, like mountain passes, could be crossed only by their leave. They were passable only at their pleasure.

Rather than mitigate supply chain risk, the Global World for ideological and economic reasons, increased their dependence without corresponding insurance, either in the form of stockpiles, alternative suppliers or military forces. They relied instead on the charlatans who offered "international law" as a substitute for physical security.

Time and again they made themselves even more dependent on bad actors: on artisanal mining of rare earths in Africa, on Russian pipelines, on undefended Taiwanese chip foundries. Until, in a moment of reckoning Iran partly closed the Bab el-Mandam and more consequentially menaced the Strait of Hormuz and there was no answer. The Global World discovered it had not the fleet to secure either passage or as in the case of the Royal Navy, found it had not a fleet at all.

Suddenly Europe rued closing down its nuclear plants. Australia regretted closing down its fuel refineries. Perhaps before long the Global World will worried about the security of Taiwan. It is easier to be wise after the fact

There’s more to the Xeet than that, and it is all good, as is the Xeet it refers to about the empty suit that is 2 Tier.

Asia and Oceania

Of course, as the that Xeet above notes, it hasn’t just been Europe that has blindly allowed itself to be put in a position of weakness regarding entirely predictable world events.

[On weakness front, the mullahs seem to have thought they could threaten, bully etc. the US as they had done for the last almost half century. It turns out that President Trump’s reaction to being threatened like that is “kill them”, unlike previous presidents. ]

Starting west and working east, Iran’s neighbor Pakistan’s government is showing signs of willingness to stand up to Iran and escort ships, but very much for its own shipping. That’s perhaps because it’s busy laying smack on Afghanistan but it may also be because it dreads being seen cooperating with its other neighbor, India and because it is trying to play both sides.

India seems to be taking it seriously. It has both done some diplomacy with Iran to get promises of safe passage and is sending naval assets to provide escorts. Plus India seems to have had a word with President Trump and got permission to take a bunch of shadow fleet oil.

Sri Lanka and Bangladesh are unable to project any force and are entirely dependent on gulf oil and fertilizer. The words “completely fucked” come to mind. In the case of Sri Lanka there’s the additional issue that they got in bed with the Chicoms, which may not have been a good choice.

Indochina is in similar state for the most part, though Malaysia and Indonesia are less dependent on the gulf and can probably export enough to obtain what they lack. They may also be able to support some of their neighbors. But a prolonged shortage is going to hurt all the countries in the region that don’t have oil/gas assets as they don’t have much in the way of suitable naval assets.

Taking a jump south there’s Australia. Which is being very European with touches of California and is now wondering whether closing all those refineries was quite so smart after all2.

Jumping back north of the equator we have Taiwan, Japan, the Koreas and the Chicoms. Taiwan is, for obvious reasons, not projecting any force a third of the way around the globe, but does, also for obvious reasons, have significant oil stockpiles. I assume the amount is similar to Japan’s 7-8 months consumption but I’m not sure. Korea probably has less, but still some. And as I discussed in a previous post, the Chicoms almost certainly do not have as much oil as they want.

Both South Korea and Japan could escort ships and, AIUI, have a few assets in the region that were used to escort ships threatened by the Houthis. Both countries have additional vessels that could be used but are not in the region, however both are currently unwilling to do so. I don’t have any insight into Korean politics, but my read on PM Takaichi is that under certain conditions she would be willing to deploy JMSDF ships in an escort role. One of the conditions though is that she be as sure as can be that the Chicoms aren’t going to do something stupid regarding Taiwan, another would be a UN Security Council resolution. She does however face significant domestic resistance to getting involved even if those base conditions are met3 where the dominant media narrative seems to be that President Trump just decided to attack Iran for no valid reason or because of the Jooz, and I’ve seen no coverage (I may have missed some, but it certainly hasn’t been major) of the US and Israeli claims that Iran was threatening to make nukes again.

Which leads us to China. China imported a lot of oil and gas through the Strait, some from Iran but plenty from other places too. China also has a large navy and thus should be able to provide escorts but that puts them in a very delicate position because they have been a major supplier of weapons to Iran, as well as consumer of most of Iran’s oil output. So if the PLAN show up to escort ships they are basically escorting them against the threat they helped create which is very embarrassing, on the other hand if they don’t they look pathetically weak and dependent on the US for oil and gas - especially given that their other play, Venezuela, just got taken out at the start of the year.

If you think placing Winnie the Flu in this dilemma was an accident then I have some swampland near Mar A Lago to sell you. It turns out President Trump can not just walk and chew gum, he can also talk and do two or three other things at the same time. In this case he’s making China look weak and unreliable as a partner because either it supports its erstwhile ally, Iran, and doesn’t help break the blockade or it dumps that ally and helps itself and its other allies/dependents like Sri Lanka. In neither case are the Chicoms getting what they want and they are seen to be dancing to the tune of the MAGA USA.

It’s a lose lose situation for Winnie the Flu and a teachable moment for the rest of the world. So all in all I think, as indeed he recently posted on Truth Social, that President Trump may be quite happy to let other people flail around and fail to solve the safe shipping problem for a while

Toll Booth Time?

And it may come to pass that in a few weeks when the MEF shows up, if there are still no patrols and escorts that President Trump takes some advice from the Instapundit4

What are the odds that Trump puts a tariff-like “security fee” on all oil transiting the Strait of Hormuz? . . .

People have speculated that the Marines are intended to seize Kharg Island and thus grab the Iranian regime by the financial short and curlies, but I wonder if there isn’t a better choice of island. Specifically Qeshm Island5 in the Strait of Hormuz. Taking that island, which has at least one port and an airport, would put the US in the position to control access to something like half the world’s oil and gas supplies.

Of course the US would have to hold it, and it’s in swimming distance of Bandar Abbas where the Iranian navy used to be, but given the general degradation of the IRGC I expect that in a couple of weeks there won’t be much that can shoot at it and meanwhile it can act as a base to stop attacks on ships passing through the Strait - for a price…

Subscribe now

2

The Concerned Observer

10–18 Days Until Australia Runs Dry

Saturday. 14 March 2026…

Read more

5 months ago · 185 likes · 30 comments · The Concerned Observer

3

Observing Japan

Takaichi's test in Washington

Thank you for reading Observing Japan. This post is available to all readers…

Read more

5 months ago · 24 likes · Tobias Harris

4

More here -

Glenn’s Substack

Don't Drone Me, Bro!

The attacks on the Islamic Republic of Iran are going almost comically well. Where before Iranian leaders used to mock the “Great Satan” as a pitiful helpless giant, now most of them are dead. The ones left alive are, well, comically begging “don’t drone me, bro!” like Iranian President Mazoud Pezeshkian, hilariously pictured below…

Read more

5 months ago · 74 likes · 9 comments · Glenn Harlan Reynolds

Treason and Fossil Fuels

Ever since the start of the Russian invasion of Ukraine in 2022 it has been clear that giving your enemies or geopolitical rivals control over your energy supply is … how shall we put it? suboptimal.

See Europe in general and Germany in particular with respect to oil and gas from Russia as an example.

Now, thanks to Operation Epic Furry Fury, that same green inspired insanity is going to hit California right in the nuts gas tank.

As has been noted by many, California has lost something like 50% of the refineries it had once upon a time. Now that Valero’s Benicia refinery has shut, it is down to 11 from 23 refineries in 2000. In fact thanks to Valero and Phillips shutting refineries down in the last few months California’s refining capacity has dropped 20% compared to a year ago.

Since California’s demand for gasoline, diesel etc. has not declined by 20% California needs to get its refined oil product from somewhere else. That has led to idiocies such as the Bahamas (a noted oil producer NOT) providing almost 40% of California’s requirements recently. Japan, India and South Korea have also contributed:

Jet fuel imports surged to a five-year high of 90,000 b/d in November, largely on the back of sharp increases from Japan. Gasoline imports reached record highs as well, averaging almost 50,000 b/d in 2025, sourced mainly from India, South Korea and more recently Japan.

All of this, while stupid, is not deadly serious as long as there are no disruptions to global oil supplies.

And, um, well, sadly, as we are all well aware, there is currently a major disruption in global oil supplies.

China has announced it is forbidding all oil exports. I have not seen any announcement from Japan but given that Japan imports over 90% of its oil from the gulf, I doubt Japan is going to be selling any more jet fuel to California either even though it has something like eight months supply stockpiled. Much the same goes for S Korea and probably India - though India may temporarily do OK because it is temporarily being allowed to purchase the shadow fleet oil hanging around in various tankers. Either way, when push come to shove, countries want their refineries to produce oil for domestic consumption first and exports across the Pacific to California are heavily de-prioritized.

The USA as a whole is a net oil exporter, that’s the source of the Bahamian oil California imports, but California is, for reasons, almost entirely removed from the rest of the US oil market. This is partly by design in that, AIUI, plans to build an oil pipeline from the gulf/Texas to California have been blocked hard by Californian regulators and their environmental NGO allies. You might ask why the US Gulf oil can’t be shipped to California. That’s where the Jones Act gets to do its thing. There are hardly any Jones Act compliant tankers and most of them are busy delivering Alaskan oil to Hawaii and the US West Coast already. There simply isn’t any spare capacity in the Jones Act compliant tanker fleet which is why oil to California from the US Gulf producers goes to refineries in the Bahamas and is then shipped on from there in tankers passing through the Panama Canal to California.

In general one would be tempted to tell California to suck it up buttercup and think about electing some saner politicians ASAP. However, despite all the billionaires and companies fleeing the state, California is in fact critical to the US economy thanks to the extensive ports in Los Angeles, Long Beach and so on. Not forgetting the agriculture of the central valley that feeds more than just California. Every import and export that passes through those ports is transport to them, on the landward side, diesel powered trains and trucks and likewise every tomato, garlic, raisin, rice grain and so on from the California’s fields goes to its consumers on diesel powered trucks and trains. So if the price or availability of diesel worsens that increases the cost of those foodstuffs, imports and exports. As it is diesel in California is some 50% more expensive (~$6/gal vs $4) than in the central US1 and, although California recently passed some law mandating that refiners hold some stock of refined product, it doesn’t look the amount actually stockpiled is anywhere near the several months worth that Japan (and probably China - see below) has.

California is also critical for the US military in the Pacific thanks to the collection of US Navy and USMC assets in Southern California as well as numerous other locations elsewhere such as Travis AFB. Travis AFB, by the way, used to be supplied jet fuel via a dedicated pipeline from the now closed Benicia refinery.

It is this importance that makes me wonder whether passing laws that make the oil industry impossible in California rise to the level of actual treason against the US as a whole. In the event that the West Taiwanese decide to get really sporty, San Diego is going to need to be refueling lots of naval vessels and air force bases like Travis are going to be needed to send out patrols as well as supplies to bases in Guam and so on. Artificially and gratuitously restricting supplies by government fiat surely looks to me like giving “aid and comfort” to enemies of the US.

True we are not yet (and hopefully won’t ever be) at war against West Taiwan, but if that happens and if it turns out that lack of fuel in San Diego limits the capability of the USN to deploy then I would think treason charges could be laid against all the politicians and bureaucrats who have overseen the mess that is the current California oil market. Sadly because the US constitution prohibits cruel and unusual punishments, these wastes of oxygen will not be hung, drawn and quartered or any of the other punishments that Britain has traditionally applied to traitors. But one can dream.

Anyway enough about treasonous greenies.

Subscribe now

How Much Oil Does China Actually Have?

The announcement that the Chicoms were banning oil exports coming so soon after war broke out in the Middle East led me to wonder just how much oil they actually have.

According to this article and links from it

China’s oil stockpiles and crude in storage in Asia and on the water leave it well positioned to weather a multi-month disruption. As of early January, China had 1.206 billion barrels of oil in storage onshore, which would cover 104 days of net crude oil imports at the 2025 level.

There also are likely ample supplies of Venezuelan and Iranian crudes on the water. In early January, more than 29 million barrels of Venezuelan crude and 170 million barrels of Iranian crude and condensate were on tankers or floating storage. There are also Iranian barrels in bonded storage in Chinese ports.

That sounds like a lot. 3 1/2 months or so. But I wonder if it’s all actually there. Failing to actually fill an oil reserve and instead selling the oil is an obvious corruption path for middle managers in the right places. It might also help China’s small domestic oil production industry. That industry produced some 3,850 million barrels of oil (216.05 million tons) last year, slightly above the year before and which if I do the sums from this article was a record level.

In 2024, China produced 213 million tons (4.27 million bpd) of crude oil, virtually equaling the historical peak of 215 million tons (4.32 million bpd) in 2015.^[xxvii] After setting that production record, China’s oil output fell to 189 million tons (3.8 million bpd) in 2018.[xxviii] This decline appears to have alarmed China’s leader Xi Jinping. In July of that year he issued an instruction to “vigorously increase exploration and development efforts to ensure national energy security.”^[xxix] China’s national oil companies subsequently developed their first-ever seven-year action plans to increase domestic reserves and production for the period 2019-2025.^[xxx] To this end, the companies invested heavily in exploration and production.^[xxxi] China’s oil output grew by almost 500,000 bpd.^[xxxii]

I wonder whether the domestic production really increased over the last few years, or whether it turned out to be easier to pretend it did and “borrow” oil from the long term stockpiles. Total domestic production is a fraction of a percent of consumption so it would be easy to goose that number and meet the goals. But once you’ve borrowed oil for one reason, it becomes easy to borrow it for other more directly individually profitable reasons.

Lack of belief in those official statistics may well explain why the West Taiwanese have halted all exports of refined oil products – they just had some people who are not an actual petroleum company managers do an audit and noticed that instead of 1.2 trillion barrels they were supposed to have that actually had significantly less. Perhaps as little as half the amount.

This is purely speculation, but a country with 3.5 months of oil stockpiled ought to be a little more confident even if it has just seen a hit on cheap imports. I mean, yes you do want to think about reducing consumption and make plans but an announcement less than a week after war starts seems extremely conservative. On the other hand one with rather less reserves than that, say less than a couple of months, might be rather more concerned that the current week or so of interruption could extend to where oil stocks look very low.

On the gripping hand, to return to California, I very much doubt that California has even a month’s supply stockpiled. If disruptions continue then California may be looking at rationing in a week or three.

Subscribe now

I don’t have any particular insights or knowledge about the Iran war, but I do have some longer term thoughts.

It turns out that Western Democracies do not have to cower and pay Danegeld to terrorists and barbarians. The combination of Israeli PM Netanyahu and US Preisident Trump has shown that there is an alternative. That alternative is killing the leaders of the people who say they want to kill you unless you do as they say. Dead people, amazingly, have very little ability to wage war and dead leaders are unable to make rabble-rousing speeches. Moreover, while one dead leader may be a martyr, dozens of them dying looks more like loserdom and (almost) no one wants to identify with dead losers. It seems that Khameini was unable to grasp the rules had changed and paid the ultimate price. It couldn’t happen to a more deserving person.

As I wrote back at the end of January, just over a month ago

I predict that Iran is going to fall. It’s going to be really messy but it isn’t going to affect America much at all.

That post also talks about Iran’s general lack of air defenses and the priorities of the attackers and how the CIA and Mossad have countless agents inside Iran to provide targeting data. All of that seems to have been right. The Mullahs and their regime can do nothing to stop the US and Israel from hitting anything they want to. That doesn’t mean they can’t inflict some casualties in return, but those casualties are likely to be minimal. The main concern for people outside Iran is to what extent Iran is able to continue attacking its Arab neighbors and shipping passing through the Straits of Hormuz. It seems likely that disruption will be of relatively short duration. The Straits and the Gulf States are basically the hostages the Iranians had against widespread attack. Unfortunately for Iran it looks like much of their offensive abilities have been neutralized and most of what they do manage to launch gets shot down. So yes, it’s not great to be a Gulf State civilian, but so far it seems to be more inconvenience than actual danger

That leads to the Mullahs’ other hope for survival, globalizing the war through “spontaneous” demonstrations and protests as well as telling all their proxies that they have to engage NOW.

The Saulstack

Iran's Only Hope: Play Hitler in the Bunker

Why is Iran doing this? Why would the IRGC risk dragging more states into the fight against them, alongside Israel and the USA? And why waste their limited stockpile of missiles—of which they have already burned through around 20% in just 36 hours, not even counting those being hunted on their launchers and in storage—on this tantrum…

Read more

5 months ago · 14 likes · 2 comments · Saul Sadka

They’re trying to shift the battlefield from the actual battle to the battle of ideas online. (It doesn’t help that many of their influential boosters are morons and/or frauds, from Candace to Scott Ritter.) This is why they tried—and are still trying—to flood the zone with obvious lies. It’s why they’re desperate to kill US servicemen. Their only hope is to frag US public support faster than their missile stockpile runs out. Read their spokesmen, like Marandi, and it’s all Epstein-conspiracy garbage—because it’s their only chance.

The problem the Mullahs have is that this isn’t working. The indiscriminate drone and missile fire on the gulf states seems to have caused them all to unify into an anti-Iran coalition for the first time in decades (if not ever). The Hez(no)bollah rocket attacks on Israel seem to have basically caused the Lebanese government to say “it’s all them, don’t get us” and for civilians to pack up and go and refuse to be human shields. And it looks like Hez(no)ballah is losing leaders again, so my guess is they are going to stop attacking pretty soon. The strike against the RAF base in Cyprus (possibly by Hezbollah?) seems to have simply provided the excuse to 2 Tier to U turn again and tell the nice Mr Trump that it’s fine to use British bases to attack Iran, sorry for the misunderstanding, even though this naturally upsets his left wing MPs.

Their stooges in Pakistan attempted to storm the US Consulate there and were repulsed by the Marine guards. The Pakistan authorities do not seem to be siding with the protestors, much to the sadness of the MSM. In fact MSM in general seem to be doing as the Saulstack post above suggests and doing their best to report much protesting against the attacks while skipping over the celebrations from exiled Iranians and others celebrating the deaths of the Mullahs and the likely overthrow of their regime.

However I don’t think this is going to work. Everyone knows Iran has been sponsoring terrorism all over the globe - remember the bombing in Buenos Aires in 1994? - and it has been hanging homosexuals and mistreating women all the time too. Sure there are some idiots who try to argue that the poor widdle Mullahs are oppressed brown people who deserve a pass, but none of them are in a position of power in the US or Israel and the leaders of those two countries have repeatedly shown that they don’t care about the clamoring of the UN and other international bodies. Fundamentally Iran has been pissing off its neighbors and the world in general for decades, with the threats of destabilizing the global economy through blocking the Straits of Hormuz and exporting terrorism as reasons why it, in turn, cannot be safely hit back at.

It turns out that, as Israel showed last year, Iran’s Russian and Chinese air defenses aren’t up to scratch against a real military. They definitely aren’t up to scratch against two. That means that the missile and attack boat threat against shipping is substantially reduced - my guess is that by next Friday, if not sooner, the surviving Mullahs and their IRGC stooges will be unable to project force much beyond the lethal range of, say, an AK-47 and so ships will be free to export Kuwaiti, Iraqi, Qatari etc. oil and gas without worrying about being struck by missiles, torpedoes, mines etc.

The last couple of years of destruction of HamAss, Hez(no)bollah and Syria have cut the regimes ability to threaten terror significantly and last summer’s raids put a major crimp in their plans to become a nuclear power and terrorize that way. Finally the Venezuela action made it extremely clear that Latin American countries that buddy up to the IRGC will face abrupt changes of leadership so the global network or terror and oil smuggling suddenly shrunk to a merely regional one. All in all, it turns out those threats that they used to have are no longer very threatening.

Which means that the only other protection Iran had was its alliances with Russia and China

Subscribe now

Winnie The Flu Hardest Hit?

And that’s a problem. Russia doesn’t have any spare capacity, it’s committed to invading Ukraine and is completely bogged down there. Russia can send thoughts and prayers, but apart from possibly drones, it cannot send any hardware and, as we have now seen repeatedly, Russian Air Defense systems are not good against Israeli or US air attacks. Allegedly Russia and/or Belarus sent a few thousand enforcer troops to help the Basij and IRGC shoot Iranian protestors, but those enforcers are not able to defend against bombs and air attacks and, in fact, may well be some of the casualties from those attacks, along with their new buddies in the IRGC and Basij.

That means that the Chicoms are the only country that could realistically come to Iran’s assistance. But short of the West Taiwanese deciding to open a second front against the US, there are distinct limits to the assistance they can provide. The USN is now effectively blockading Iran so freighters full of weapons are unlikely to make it to port. And if they do make it to port, they probably won’t make it off the dock and will provide an exciting fireworks display instead when the ship and/or cargo gets a visit from a friendly American missile. There are in fact reports of mysterious massive explosions in various remote parts of Iran that may be the US removing previously supplied Chicom weapons. Flights into Iran are contra-indicated and transfers via Russia are going to look like Russian transfers and get hit the same way.

The problem for West Taiwan is that if it gets a little too sporty, it faces the prospect of a de facto oil embargo and as of right now it almost certainly doesn’t have the oil stockpiles it needs to survive even a short term one. Sure it can survive for a couple of weeks to a month, but despite all the solar panels and EVs, if there is a significant disruption then the economy literally runs out of fuel because it simply doesn’t have a couple of months supply of diesel. And those supplies get even shorter if it tries for large scale naval operations because most of the PLAN uses diesel or similar.

I have no doubt that given a year or two to prepare, the Chicoms could stockpile enough oil and coal that they could last longer, but although they were stockpiling fuel a bit in 2025, they just lost Venezuelan and Iranian oil supplies. Sure they seem to have partly replaced those with Saudi and additional Russian supplies, but replacement does not mean additional, and, more importantly, as the first link above says, they haven’t yet finished building the infrastructure to store more. Moreover, while they have Russia over a barrel, Saudi Arabia is going to be expecting market price for its oil. West Taiwan certainly can afford to pay market price, but having to do so for ~10% of its oil that it previously got for very cheap adds to the difficulty of building up stockpiles for the future. Summing up that means they remain vulnerable, and, as I said back a couple of years ago1 in January, Winnie the Flu is not a risk taker.

Not only that but Winnie the Flu is still removing top military leaders. I feel that Winnie the Flu and his fellow Chicoms are struggling more than most to keep up with the frenetic pace of action in Trump 2.0. In any “normal” US presidency, there would be significant pauses between operations that would allow other countries to figure out what was coming next and prepare. Trump 2.0 isn’t doing that and that means that plans to spend, say, six months stocking up on oil no longer work.

The Instapundit sees this pressure on the Chicoms as a beneficial side-effect:

Trump wants to weaken China without going to war with China. He has now cut off two major suppliers of oil to the PRC, which produces hardly any oil of its own. (It’s worse than that, because China wasn’t paying for that oil with dollars, and now it will need dollars to buy oil elsewhere.) That applies a squeeze to an already squeezed CCP, and will make Xi’s position, domestically and internationally, weaker. Also the military excellence recently displayed has to inspire second, third, and fourth thoughts about invading Taiwan.

Glenn’s Substack

Cutting Off the Head of the Snake

November 4, 1979 — almost 47 years ago — Iran seized the American embassy in Tehran and held its staff hostage. Ever since then, American presidents have struggled with what to do…

Read more

5 months ago · 320 likes · 40 comments · Glenn Harlan Reynolds

I think there may be more to it than that. He is in fact taking out West Taiwanese allies one at a time and strongly incentivizing everyone to reduce business with the Chicoms will doing more business with the US. That squeezes them even more.

This also, BTW, impacts Cuba, which is steadily going titsup.com while everyone’s attention is on the Middle East. It turns out that the US Navy can do two or three things at once and enforcing an oil blockade on Havana is dead easy. I assume the top Castro-ites are have serious conversations about how they can safely get out of the country before they get am Easter-tide Romanian Christmas.

One other thing that Trump 2.0 has made clear is that when he gives you a deadline and an agreement to sign on to, if you try to weasel out you get whacked. That applied to Maduro and the Mullahs. He has not, yet, given the Castro-ites a public deadline, but I don’t think anyone doubts that he has every intention of ensuring that Cuba undergoes regime change in the near future, and that certainly includes the heirs of Fidel themselves.

Going back to Russia. I strongly suspect that President Trump is trying to split Russia off from West Taiwan. That may be a challenge, but unlike my friend Nicki Kenyon, I don’t think chaos in Iran is going to benefit Putin

Into the Void

Iran War 2026

I have spent the last three days with my parents. My mom has dementia, and she has taken a severe turn for the worse. She can no longer speak, she doesn’t understand anything but very basic phrases, and her blood pressure was so high on Friday, she lost consciousness…

Read more

5 months ago · 7 likes · 4 comments · Nicki Kenyon

I do not expect Russia to be able to install a friendly puppet there because I strongly suspect that the goal of the current war is to totally eliminate the Mullahs, the IRGC and the Basij as well as all the corrupt oligarch class that supported them. Russia will have zero influence on the various groups that emerge from the wreckage because they will all be groups that the mullahs oppressed. It is possible that Russia will benefit from higher oil prices, but not as much as Nicki thinks because Russia’s customers have shrunk and the “shadow fleet” is under considerable pressure. The only real major customer for Russia now is West Taiwan and for all the obvious reasons, the Chicoms are going to demand significant discounts to market price for Russian oil.

However, Trump does like making deals. I expect that someone on the Trump team is going to give Putin a deal of the “you can’t refuse it” kind any day now. The deal is going to look something like this.

1. Stop invading Ukraine and give some territory back (how much TBD but significant)

2. Stop supplying China with cut price oil or other cut price resources

3. Don’t assist the mullahs, the Cubans, the Venezuelans etc.

4. In exchange we will let you sell gas to Europe and sell your oil to anyone at market price

5. By the way, here are some pictures of buildings in Moscow, and some technical characteristics of bunker busting bombs

I do not know whether Putin will take this deal. If he withdraws from any part of Ukraine he will face significant push back in Russia. But after recent reverses on the ground in Ukraine - Ukraine has retaken control of 100-400 km22 in the last month - and continuing destruction of Russian industry, and manpower, he may be willing to accept a deal.

Now of course it takes three to tango here, and Ukraine may not be willing to join in without significant Russian withdrawal and some kind of solid guarantee that Russia won’t simply try again in a few years, but Ukraine is also in poor shape after 4 years of war. This last winter has in fact been hard on both sides and so I suspect there is a point that both sides can accept and call a win. Note that I disagree with chunks of this foreign affairs piece in that I think it understates Russian casualties and overstates remaining Russian power but I tend to agree that Ukraine is not at all close to winning, where by winning I mean regaining most of the territory it lost to Russia.

However, see Iran, if such a deal is offered to Putin, it will have a time limit. And some serious teeth in the event that the deal is not taken by the time the limit expires. Teeth being far greater support for Ukraine and a willingness to enforce NATO blockades of Russian oil exports from western ends of Russia, among other things. The mullahs appear to have tried to run the clock on negotiations even though President Trump told them that they had until the end of February. Putin would be wise to bear that in mind.

My bet is that Putin will do just like the mullahs and misread Trump. He will look at how he has strung Trump along for the last year and assume he can continue to do so. And Trump will supply Ukraine with target data and GPS etc. tricks such that Ukraine’s missiles can hit anywhere and everywhere in Russia. Along with guidance on how to spoof Russian radar and quite possibly some explicit tips and hints about how to penetrate reinforced concrete bunkers.

Thank Sinwar

It seems to me that a major reason why President Trump is willing to take out dictators is that he has seen how well this worked for Israel. Israel almost certainly would not have thought of it prior to October 7 and October 7 was Sinwar’s bright idea. After that, the Israelis discovered that eliminating the top leaders of their enemies was remarkably effective. Indeed, given the Chicom and Iranian funded anti-Semitic protests in the West, Israel was forced to take out leaders. It did so and found when it did it that the resulting panic and chaos was extremely beneficial. Taking out successive waves of top leaders drastically reduced the ability of the enemy to coordinate a response or to conceal it, and that meant it was far easier to take out the fanatical lower layers still trying to do things.

The US in Venezuela is trying a slightly different tactic - take out one leader, tell the replacement that they will gradually reform or they get similar treatment - and it is possible that once Iran’s missile and drone threat appears to have been thoroughly neutralized the US will try the same thing to the handful of remaining members of the current ruling regime. But, based on other comments, I suspect the preferred outcome is for the Iranian people to rise up and kill the regime survivors3.

However, if it weren’t for Sinwar, I suspect none of this would have happened, so in a twisted way we can thank him for causing the downfall of his own terrorist organization, his allies and his sponsors. Well Done! Rot in hell!

Subscribe now

Update/PS: I’m not the only one thinking that the Chicoms are part of the losers here - Steven Mosher thinks so too.

The Gorton and Denton by-election has happened, and as I predicted, Labour came third.

Though more people voted Labour than I expected (25% actual vs 10-15% prediction) and sadly not enough people were convinced to vote Reform so the Greens won. But, while turnout was lower than one might have hoped, there’s a real humdinger of an allegation that makes the Green victory very iffy.

Today we have seen concerningly high levels of family voting in Gorton and Denton. Based on our assessment of today’s observations, we have seen the highest levels of family voting at any election in our 10 year history of observing elections in the UK.

Family voting is not a term I’ve heard of before, but it is the situation where two voters either confer, collude or direct each other on voting. And obviously cases where one voter oversees the votes of more than one other person as well.

Democracy Volunteers, the organization making the allegation, is a reputable decade plus old organization and not a partisan one.

Democracy Volunteers is run by Dr John Ault, a former Liberal Democrat politician who has observed elections in countries including Britain, Sweden, Norway and Finland.

They give more detail on their webpage

2023 saw the enactment of the Ballot Secrecy Act, which made the practice of family voting more clearly a breach of the secret ballot, making it more enforceable by staff in polling stations. Signage is now available to discourage the practice. Signage was only seen in 45% of the polling stations observed.

The observer team saw family voting in 15 of the 22 polling stations observed, some 32 cases in total, nine cases in one polling station alone. The team observed a sample of 545 voters casting their vote – meaning 12% of those voters observed either caused or were affected by family voting.

Commenting John Ault, Director of Democracy Volunteers said;

‘Today we have seen concerningly high levels of family voting in Gorton and Denton. Based on our assessment of today’s observations, we have seen the highest levels of family voting at any election in our 10 year history of observing elections in the UK.’

‘We rarely issue a report on the night of an election, but the data we have collected today on family voting, when compared to other recent by-elections, is extremely high.’

‘In the other recent Westminster parliamentary by-election in Runcorn and Helsby we saw family voting in 12% of polling stations, affecting 1% of voters. In Gorton and Denton, we observed family voting in 68% of polling stations, affecting 12% of those voters observed.’

[…]

The team also saw a number of voters taking photographs of their ballot papers and one voter being authorised to vote despite them already having been marked as voted earlier in the day.

What they do not say, unfortunately, is which polling stations they observed this in. We can guess. In fact the Torygraph reports that Reform has explicitly made the obvious accusation:

Nigel Farage, the Reform leader, said allegations of family voting raised “serious questions about the integrity of the democratic process in predominantly Muslim areas”.

Subscribe now

If the numbers reported are extrapolated to the total number of votes cast then it is not impossible that the Green’s 4402 vote victory came from family voting as 12% of the total vote is 4417 votes. I think that may be overstating things, because 32 out of 545 is 5.9%. It’s only 12% if you double the number of voters and imply that basically every, say, husband ensured that his wife voted the same way. That of course ignores cases where voter H ensures that voters W, S and D all vote the same way and so on.

But whether or not the Greens would have scraped a win without this kind of assistance, and my guess is that they probably would have, this credible accusation is going to cause a lot of reaction and subsequent accusations of “Islamophobia” etc.

I anticipate a lot of pressure being placed on DV to either retract their statement or not provide details that could help fuel “sectarian division”. This will of course not work because we’ve seen all this before. Specifically those of us who are paying attention have seen numerous cases where people of “South Asian” origin have been involved in voting irregularities - in one case it was Asian A vs Asian B - and thus the likely accusations of Islamophobia, racism etc. are easy to rebut with evidence.

In the short term I anticipate this will help 2 Tier survive a bit longer because controversy about voting will take the spotlight off the collapse of the Labour vote. I doubt this will last. If (when) Labour lose massively in the May local elections, then I expect resignation pressure to be intense, but, assuming no further government cockups in the next couple of months1, he will probably stagger on until those elections.

Depending on the details of the full report from DV, we may well see Reform attempt to have the election anulled and have another one. I am sure they are talking to lawyers as I write this to see what legal options they have. I have no idea if this will succeed or not, but it will certainly draw more attention to the issue.

A more important consequence is that there will be close examination of the voting in those local elections. I anticipate a lot of requests for widespread election observers and/or other monitoring of the voting in May with requests to enforce the ban on family voting and, if those requests are seen to fail, attempts to have violators arrested and then probably disruption as one side or another takes offense. Which side causes the disturbance will of course depend heavily on whether the police attempt to arrest violators or not. My prediction is that the police will mostly not try arresting violators and that they will instead try arresting the people complaining about the family voting.

Bottom line: those May local elections may provide rather more fireworks than expected

Subscribe now

Update: Rod Dreher has some discussion of the result here

In particular he links to this oped by Daniel Hannan in the Daily Mail which discusses the Muslim pandering of the Greens and others

There’s been a considerable amount of hype recently about the capabilities of recent LLM AI. There was the Clawdbot Moltbot OpenClaw hype which mostly demonstrated the credulousness of tech journalists and the lack of security thinking of AI researchers and developers. There’s Matt Shumer’s famous screed - which Ed Zitron poked a few holes in (archive) as does Gary Marcus

So let’s get this out the door now. There’s a lot of hype and exaggeration. LLM AI is not going to take everyone’s job next week or even next decade. In fact as Gary Marcus points out, IBM is citing AI as a reason to increase US hiring 1

But

It is going to take away a lot of routine administrivia jobs and low level technical stuff fairly soon, indeed it already is. Shumer’s claims that Claude AI is programming itself and that lawyers love it more than 100 interns are probably exaggerated, but ES Raymond, Perry Metzger and various other real programmer sorts are using AI to do more things, faster and better than they could before. Likewise there was a big story about accountancy firm KPMG getting a discount from their own auditors because of AI efficiency which has led numerous people to wonder if we’re seeing the end of the billable hour. And that is a legitimate issue for lawyers, accountants and other similar professionals.

Also yes, the last year or so has been a learning process (a real one, not a “learing” one) as we figure out how to work around the AI hallucination problem, the too big for context problem and so on. It looks like we are finding work arounds to those problems, but the openclaw shenanigans have shown that trusting AI to do something without user intervention is still fraught with danger. I predict that there will be considerable opportunity to write AI verification and validator tools and possibly employment opportunities as the person who confirms that the AI output does what it says it does and isn’t a lie. For what, I trust, are obvious reasons, we don’t ask the AI to do final validation of its own output - or even the output of a different AI

However auditing, routine legal work (e.g. contracts), basic office functions like much of HR are vulnerable to being AIed away. A lot or marketing and sales too. Create this powerpoint, generate that quote/invoice and so on. Your AI friend will do that for you just fine, but be sure to double check that it didn’t get prompted to add an extra discount to the invoice or some extra fingers to the photo in the presentation.

However there’s a catch here. These phony baloney jobs will only be AIed away if the AI option is cheaper than a a human being paid, say, $20/hour as I discussed in my last AI post

How much does it cost?

If you want to write, say, a C compiler then there are reported costs2:

To stress test it, I tasked 16 agents with writing a Rust-based C compiler, from scratch, capable of compiling the Linux kernel. Over nearly 2,000 Claude Code sessions and $20,000 in API costs, the agent team produced a 100,000-line compiler that can build Linux 6.9 on x86, ARM, and RISC-V.

However that $20,000 is almost an enormous understatement of the actual costs involved. For a start there’s the fact that Anthropic spends at least $2 for every $1 of income it gets. I say, at least, because the sums are murky. Ed Zitron’s latest has this:

Even after a year straight of manufacturing consent for Claude Code as the be-all-end-all of software development resulted in putrid results for Anthropic — $4.5 billion of revenue and $5.2 billion of losses before interest, taxes, depreciation and amortization according to The Information — with (per WIRED) Claude Code only accounting for around $1.1 billion in annualized revenue in December, or around $92 million in monthly revenue.

The AI companies are generally speaking giving their product away. Even when they charge for it, the fee they charge is far lower than the cost of providing the service. That is the direct inference cost of that particular user’s usage is more than that user pays and ignores all the compute being given away for free and the costs of training the models used, the data center construction etc.

Let’s talk training for a second. Again quoting from the excellent Ed Z:

[A]arguably the most dishonest part is this word “training.” When you read “training,” you’re meant to think “oh, it’s training for something, this is an R&D cost,” when “training LLMs” is as consistent a cost as inference (the creation of the output) or any other kind of maintenance.

While most people know about pretraining — the shoving of large amounts of data into a model (this is a simplification I realize) — in reality a lot of the current spate of models use post-training, which covers everything from small tweaks to model behavior to full-blown reinforcement learning where experts reward or punish particular responses to prompts.

To be clear, all of this is well-known and documented, but the nomenclature of “training” suggests that it might stop one day, versus the truth: training costs are increasing dramatically, and “training” covers anything from training new models to bug fixes on existing ones. And, more fundamentally, it’s an ongoing cost — something that’s an essential and unavoidable cost of doing business.

As far as I can tell this is 100% correct. OpenAI and Anthropic put AI training in the CapEx budget to flatter their OpEx numbers to make it seem like they are closer to breaking even on a income to COGS basis. They aren’t. Even with this tweakery they are still losing money on every user, but they reduce the apparent loss. It’s kind of like a pottery putting its kilns and the price of fuel to run them as CapEx and just counting the clay and glaze as OpEx.

It is phenomenally difficult to get a straight answer about how much Microsoft, Google, Amazon, Oracle3 etc. are spending on AI operations costs, and how much they get in income from them. Of course OpenAI and Anthropic use these other company’s data centers and pay for them, but this is extremely incestuous as they often have “investments” from the data center companies that they then pay back to them in usage fees. It is very reminiscent of the 2000s and Nortel/Lucent, if not Enron. Money is being spent on power, servers, buildings, wages and so on and the bills are being paid by money from somewhere. But where exactly is mysterious, and that seems to be a deliberate choice by all involved.

Which leads us to the data center costs.

Subscribe now

$1T CapEx is planned

As many people have noticed the AI companies are spending telephone number amounts of money to build more data centers and fill them with AI crunching computers. The even cost more than a Somali Day Care Learing Center

In fact they cost so much that companies are doing interesting financial engineering to come up with the money. Alphabet (Google) is planning to sell 100 year bonds, which seems to me to be a stretch. Oracle is seeing problems raising the debt to pay for its Abilene Stargate complex. Meta is doing something weird to finance its data center in Louisiana. And so on.

On that note SpaceX’s purchase of xAI could also be a bailout as Gary Marcus notes:

I think the merger is really a kind of bailout, to give a lot of cash to a company that is otherwise in distress. Fact is, xAI ain’t doing all that great. It’s burning money fast, with no obvious business model or market niche, and has little to show for it. They have also faced a lot of backlash for being reckless and irresponsible; hardly a good brand name. Grok (their main product) doesn’t have the users that ChatGPT has, and it doesn’t have the prestige that Google’s Gemini seems to be racking up. It also doesn’t have the clear corporate focus that Anthropic has. Nor does it have any obvious secret sauce.

And he includes this helpful graphic

Now I wouldn’t bet against Elon, and he mentions other reasons for the merger too. But nothing says that Elon might not want to corner the market in space/lunar data centers AND bail out his AI company, which is kind of where I’m leaning.

Eventually the roughly $1T in current and planned AI data center investment has to be paid back in addition to the ongoing cost of running the gigawatts of data centers that investment has built4. That’s a challenge at $20/month or even $200/month which seems to be about the top end of what the market will currently bear and the use case of replacing low end clerical work.

If pricing needs to go up to $2000/month then now we’re approaching the cost employing a human - an offshored Sanjay providing the needful, to be sure, but a human never the less. The AI tools have to be better and more reliable than they currently appear to be for that to be a broadly applicable worthwhile tradeoff.

That’s not to say they aren’t worth it at that kind of cost in certain situations. The coding/development assistants are absolutely going to pay for themselves for the right kinds of task at $2000/month or even, like the Anthropic C compiler at $20,000 for two weeks of work. Depending on the task they might pay for themselves at even higher amounts such as $50,000 for a week. I can, for example, think of tasks that $dayjob and/or $dayjob’s partners want to do that we/they don’t have the manpower to do but which would pay off a $50,000 spend in a year.

The question is whether there are enough $50,000 tasks to pay for the investment. Annually, as I said in a previous post5 almost year ago, that means a minimum of $100B. When you add actual OpEx and a need to make some kind of profit you need to get an additional $50-100B/year. That’s a LOT of $50,000 tasks required and I suspect most organizations only have a handful. Because they too can do similar sums, that is why many people say AI is a bubble that must pop and much of that $1T will never be paid back.

It is also worth pointing out that a lot of the CapEx spending is on things that depreciate fairly quickly - the Nvidia chips and now custom chips from others. Not only do these chips fail, they look to be no longer useful in 3-5 years. So while the data center building and some of the internal bits may last a decade or more, the servers will need to be replaced multiple times before the building is no longer suitable.

So when I said that “it’s kind of like a pottery putting its kilns and the price of fuel to run them as CapEx and just counting the clay and glaze as OpEx”, it’s actually worse than that because parts of the kilns themselves need to be rebuilt every few years too. A significant fraction of the $1T has to be replaced in 3-5 years which means AI needs to generate at least $200B/year to pay for itself, and quite likely closer to $300B

The Optimistic End State

The optimistic view is that AI will enable entrepreneurs to come up with new things to do that will profitably use the AI data centers in addition to the replacement of grunt work. I think this is almost certainly true. Eventually. For example, Elon Musk’s robots will need masses of AI to tell them what to do in detail as they do enormous amounts of grunt work for humanity. And I’m sure there are plenty of other people who have ideas. Lenny Rachitsky Xeeted a summary of a discussion he had with a lead OpenAI engineer which includes the following:

8. The one-person billion-dollar startup is coming, but with unexpected second-order effects. As AI makes individuals more productive, we’ll see not just billion-dollar solo founders but an explosion of small businesses: hundreds of $100M startups and tens of thousands of $10M startups. This will transform the startup ecosystem and venture capital landscape.

9. Business process automation is an underrated AI opportunity. While Silicon Valley focuses on knowledge work, most of the economy runs on repeatable business processes with standard operating procedures. There’s massive potential to apply AI to these workflows, which are often overlooked by the tech community.

I agree with both of these points and some others too6.

However, in order for the current AI investment frenzy to not generate lots of large bankruptcies, the one-person $10M+ startups replacing repeatable business processes need to be up and running profitably a couple of years from now. Because if they aren’t then the bondholders for the data center debt aren’t going to get paid and that is going to cause financial damage that looks quite a lot like 2008. It may not be quite as bad - sub-prime mortgage debt was around $1.5T just before the wheels came off and we’ve had a fair amount of growth and inflation since. But a debt shock that is a third to half the 2008 shock is not a good thing. Especially since

1. Countries don’t have the resources to bail out the overstretched this time around thanks to a) not paying off the debt from 2008 and b) getting further in debt in 2020/1 for Wuflu

2. The Chinese property market is at least one other massive problem and may also soak up resources that would be needed to bail out AI (or vice versa - see how the US sub-prime mortgage crisis ended up causing the southern European Euro crisis)

3. One of the obvious uses for AI is going to be crime - indeed we’re already seeing AI assists in crime. Specifically fraud and cybercrime. If there is a widespread rise in such crime then laws may get passed that end up stopping much otherwise beneficial AI usage in an attempt to stop the AI powered crime wave7. For example an obvious fraud method is to have an AI impersonate a real person. A ban on AI impersonating humans (as originally requested a couple of years ago) could end up hurting any number of positive usages for human impersonation.

Eventually, AI is going to transform society, positively. For example nothing says that AI’s use in crime can be only on the offense. An obvious use for AI right now is to scan the HHS Medicaid data for fraud, and there is absolutely no doubt that AI will be a wonderful tool to check up and validate applications for government grants and payments in general in the future.

But in the next two or three years it may just impoverish society first.

Subscribe now