Pundit ClaimsAI claims by Gary Marcus and Ed Zitron, with the evidence

Ed Zitron Where's Your Ed At Published

AI's Brokenomics

We recorded 41 claims from this post. 21 were checked against evidence. 20 can’t be checked. 14 more are held. Read the post at its source.

Claims from this post (41)

In the order they appear in the post. Each quotation is the speaker’s words in this post. A label such as “Supported” or “Can’t be checked” is the claim’s current assessment; its page has the evidence and the reasons.

  1. Assessment: MixedSaid

    “Anthropic has not created recursive self-improvement, nor has it done much more than create a very large language model that gets higher benchmarks in tests built for large language models, wrapped in a veneer of mysticism and panic-hype built to scare organizations in paying them to use it.”

    What we tested: Anthropic has not created recursive self-improvement, nor has it done much more than create a very large language model that scores higher on benchmarks built for large language models, wrapped in mysticism and panic-hype meant to scare organizations into paying to use it.

    What held: Anthropic explicitly acknowledges that it has not reached the autonomous successor-development capability it defines as recursive self-improvement. This supports that component of Zitron's statement. (Anthropic, )

    What did not: Mozilla identifies the evaluation as its application of Mythos Preview to Firefox. Vulnerabilities fixed in a shipped browser constitute practical results beyond benchmark scores. (Mozilla, )

  2. Assessment: Can’t be checkedSaid

    “Fable is, of course, just another large language model that’s an indeterminate amount of “better” than the last one.”

    What we tested: Anthropic's Fable is just another large language model only an indeterminate amount better than its predecessor.

  3. Assessment: Can’t be checkedSaid

    “The outright mythology of Mythos existed to scare people and help Anthropic raise at a $965 billion valuation”

    What we tested: Anthropic's Mythos marketing existed to scare people and help Anthropic raise money at a $965 billion valuation.

  4. Assessment: Can’t be checkedSaid

    “Some are attempting to read this story as bullish for Anthropic — that the government will work with it to bring the models back online, creating a proxy marketing campaign for its models — and while I think that’s possible, if not likely, I think there’re many other possibilities.”

    What we tested: It is possible, if not likely, that the US government will work with Anthropic to bring its models back online, in effect creating a proxy marketing campaign for them, though there are many other possibilities.

  5. Assessment: Can’t be checkedSaid

    “This, combined with Microsoft AI CEO Mustafa Suleyman saying Anthropic’s models were too expensive and Andy Jassy likely being part of the reason that Anthropic got banned makes me think that hyperscalers might be trying to cast doubt on the inevitability of AI labs.”

    What we tested: Hyperscalers such as Microsoft and Amazon may be trying to cast doubt on the inevitability of AI labs.

  6. Assessment: Can’t be checkedSaid

    “It’s hard to see how everything goes back to normal from here. Even if Anthropic gets its models greenlit for availability, it’s clear the government has some animus against it after Q1’s battle with the Department of Defense”

    What we tested: It is unlikely Anthropic's situation goes back to normal, since the government has animus against it even if its models are re-approved.

  7. Assessment: Can’t be checkedSaid

    “The other problem is that this is terrible, terrible timing for an AI industry in the throes of a cost crisis. Anthropic and OpenAI’s IPOs depend on myth, hype, and certainty that their growth will never slow.”

    What we tested: The AI industry is in a cost crisis, and Anthropic's and OpenAI's IPOs depend on hype and certainty of never-slowing growth.

  8. Assessment: Can’t be checkedSaid

    “Amazon and Microsoft are far more scared of the White House than they are of killing their golden goose, and may honestly be relieved to find a reason to bring this era to an end.”

    What we tested: Amazon and Microsoft fear the White House more than losing AI labs and may be relieved to end this era.

  9. Assessment: UnresolvedSaid

    “Their business models don’t fucking work.”

    What we tested: OpenAI's and Anthropic's business models do not work.

  10. Assessment: UnresolvedSaid

    “I’ve been saying for years that the underlying economics of AI don’t make sense — that AI labs were intentionally obfuscating the costs of subscriptions and heavily subsidizing users’ compute, and that the moment that that changed, everything would begin to fall apart, and god damn has it finally begun.”

    What we tested: AI labs intentionally obscured the costs of subscriptions and heavily subsidized users' compute, and now that this has changed, everything has begun to fall apart.

  11. Assessment: UnresolvedSaid

    “that the majority of users will refuse to pay the actual cost of AI.”

    What we tested: The majority of AI users will refuse to pay the actual cost of AI.

  12. Assessment: ContradictedSaid

    “In a single podcast, Andrew Macdonald gave the entire tech industry permission to say the truth: that nobody was actually able to show any ROI despite its massive costs.”

    What we tested: Nobody adopting AI has been able to show any return on investment despite its massive costs.

    What happened: The survey documents positive quantified returns relative to investment, rather than only usage or anticipated benefits. Fieldwork ended September 17, 2025; the results were publicly announced before Zitron's post. (Snowflake, reporting research conducted by Omdia, )

  13. Assessment: UnresolvedSaid

    “This was always going to be a problem. By starting everybody off with subsidized subscriptions, AI labs shielded users from the costs, training them by proxy to use AI models without any concern for efficiency.”

    What we tested: Subsidized subscriptions shielded users from costs and trained them to use AI without regard to efficiency, causing problems once real billing started.

  14. Assessment: UnresolvedSaid

    “As I’ve argued previously, the vast majority of Anthropic’s recent surge in revenue comes from experimental revenue from paypigs that it doesn’t deign worthy of clear visibility into their organizational token spend.”

    What we tested: The vast majority of Anthropic's recent revenue surge comes from experimental spending by customers whom Anthropic does not give clear visibility into their organizational token spend.

  15. Assessment: UnresolvedSaid

    “I cannot express how deadly these price cuts would be to the AI industry, and how dangerous this conversation has become.”

    What we tested: If OpenAI and Anthropic make the planned token price cuts, they would be deadly to the AI industry.

  16. Assessment: Can’t be checkedSaid

    “A 50% price cut doesn’t actually answer the question of “why am I paying so much for this,” and unless the price cuts are to DeepSeek levels (which would also be fatal), it’s hard to see how organizations are going to be won over.”

    What we tested: A 50% price cut does not answer customers' question of why they are paying so much, and unless the price cuts go as low as DeepSeek levels, which would also be fatal, it is hard to see how organizations will be won over.

  17. Assessment: Can’t be checkedSaid

    “This means that there’s every chance this both cuts revenues and ends up with organizations using fewer tokens.”

    What we tested: If the token price cuts happen, there is every chance they will both cut the AI labs' revenues and lead organizations to use fewer tokens.

  18. Assessment: UnresolvedSaid

    “Newsflash! These price cuts are not happening because Anthropic or OpenAI made their products more efficient! They’re making these price cuts because their customers don’t want to pay their current prices!”

    What we tested: OpenAI's and Anthropic's price cuts are caused by customer unwillingness to pay current prices, not efficiency gains.

  19. Assessment: Can’t be checkedSaid

    “In fact, their costs appear to be increasing, which is why they’ve raised (assuming the rounds completely close) over $230 billion in the last six months.”

    What we tested: OpenAI's and Anthropic's costs appear to be increasing, which is why they have raised over $230 billion in the last six months, assuming the rounds completely close.

  20. Assessment: SupportedSaid

    “That’s right. Anyone with a $200-a-month Anthropic subscription can burn $8000 in tokens, and with a $200-a-month ChatGPT subscription, you can burn $14,000 in tokens.”

    What we tested: A $200/month Anthropic subscription allows $8,000 of token usage and a $200/month ChatGPT subscription allows $14,000.

    What happened: The article reports the June SemiAnalysis test of both companies. This passage supplies the Anthropic component and identifies an approximate API-equivalent ceiling. (Tom's Guide, )

  21. Assessment: Can’t be checkedSaid

    “This obscene and wasteful subsidy is what you do when you have little to no confidence in the actual value of your product!”

    What we tested: OpenAI's and Anthropic's obscene and wasteful subscription subsidies are what a company does when it has little to no confidence in the actual value of its product.

  22. Assessment: SupportedSaid

    “That’s right folks. With the current subsidies, all it takes for a user to have a gross margin of at best negative 25% is for them to use as little as 25% of their rate limit.”

    What we tested: Even assuming a 75% token gross margin, a subscriber using 25% of their rate limit yields at best a -25% gross margin.

    What happened: The reproduced inputs yield -25% for the two lower Anthropic tiers and lower margins for every other listed tier. This supports the arithmetic under the specified assumption, not actual profitability. (Vlad's Playbook, )

  23. Assessment: Can’t be checkedSaid

    “This isn’t a real business model because generative AI companies are not real businesses.”

    What we tested: Generative AI companies are not real businesses.

  24. Assessment: UnresolvedSaid

    “It requires endless pre-training, post-training and script-based MacGuffins to do tasks with mathematically-guaranteed hallucinations that burn more tokens, raising costs on customers who are already in mutiny less than a quarter into being forced to pay a cost that is already unprofitable.”

    What we tested: Generative AI requires endless pre-training, post-training and script-based workarounds to do tasks with mathematically guaranteed hallucinations that burn more tokens, raising costs on customers who are already in mutiny less than a quarter into paying an already-unprofitable price.

  25. Assessment: UnresolvedSaid

    “It isn’t becoming “better” in a way that’s measurable using anything other than benchmarks invented specifically for generative AI — an industry-wide coddling of a mediocre technology that only makes money through massive subsidies, FOMO and executive ignorance.”

    What we tested: Generative AI is not measurably improving except on AI-specific benchmarks and only makes money through subsidies, FOMO and executive ignorance.

  26. Assessment: ContradictedSaid

    “Generative AI does not have a business model. It is not a tool with value remotely commensurate with its costs. It isn’t getting cheaper for the providers or the customers.”

    What we tested: Generative AI does not have a business model, is not a tool with value remotely commensurate with its costs, and is not getting cheaper for providers or customers.

    What happened: Anthropic announced a lower customer price for its fast-mode service before the post. This contradicts the categorical customer-price component, but establishes neither lower aggregate bills nor provider profitability. (Anthropic, )

  27. Assessment: UnresolvedSaid

    “If stopping training was all that it’d take to turn these businesses profitable, they’d have done it already, because inference would be a money-printer rather than a cursed object eating away at Altman and Amodei’s souls.”

    What we tested: Stopping model training would not make OpenAI and Anthropic profitable; inference is not a money-printer.

  28. Assessment: UnresolvedSaid

    “if they stop training, the models will eventually drift into obscurity.”

    What we tested: If AI labs stop training, their models will eventually drift into obscurity.

  29. Assessment: Can’t be checkedSaid

    “I’ve said it once and I’ll say it again: I believe a large majority of AI token spend — and specifically Anthropic’s revenue growth — has come from Business Idiots disconnected from any real work that have become convinced that “lots of AI” would do something other than rack up massive bills.”

    What we tested: He believes a large majority of AI token spend, and specifically Anthropic's revenue growth, has come from 'Business Idiots' disconnected from any real work who became convinced that 'lots of AI' would do something other than rack up massive bills.

  30. Assessment: UnresolvedSaid

    “The problem with this argument is that we’re yet to prove if running these models is profitable (or even sustainable) for any provider, nor do we have tangible proof that they can compete at scale with Anthropic or OpenAI’s more-complex LLMs.”

    What we tested: It is not yet proven whether running cheaper open-source models is profitable, or even sustainable, for any provider, and there is no tangible proof that they can compete at scale with Anthropic's or OpenAI's more-complex LLMs.

  31. Assessment: Can’t be checkedSaid

    “I am still skeptical that this is a real shift away, if only because using open source models requires you to either work with an inference provider or run your own GPUs.”

    What we tested: It is doubtful there is a real enterprise shift away from frontier models to open-source models.

  32. Assessment: Can’t be checkedSaid

    “The fact that Anthropic and OpenAI have any business left after stories like these is proof that the vast majority of companies paying for these services are doing so because they feel pressured to by their peers, investors or the media.”

    What we tested: That Anthropic and OpenAI have any business left after stories like these is proof that the vast majority of companies paying for their services do so because they feel pressured by peers, investors or the media.

  33. Assessment: Can’t be checkedSaid

    “Hell, let’s take it one step further: neither OpenAI nor Anthropic is a real business.”

    What we tested: Neither OpenAI nor Anthropic is a real business.

  34. Assessment: Can’t be checkedSaid

    “it’s becoming clear that token-based billing might genuinely not work as a viable business line.”

    What we tested: Token-based billing may not work as a viable business line for AI labs.

  35. Assessment: UnresolvedSaid

    “Anthropic negotiated discounted compute for the first two months of its SpaceX deal as a means of pretending to be profitable for a single quarter, but any price cut — or even customer churn! — will immediately put its finances in a kind of red usually reserved for the deeply embarrassed or steroid-enhanced.”

    What we tested: Anthropic negotiated discounted compute for the first two months of its SpaceX deal as a way of pretending to be profitable for a single quarter, and any price cut or even customer churn will immediately put its finances deep in the red.

  36. Assessment: Can’t be checkedSaid

    “They do not have a plan.”

    What we tested: OpenAI and Anthropic have no plan to become profitable.

  37. Assessment: UnresolvedSaid

    “There is no reversing this trend, because if there were, OpenAI and Anthropic would’ve reversed it in any way, shape or form, rather than raising more money than anyone has ever raised before for what appears to be no reason other than to burn it.”

    What we tested: OpenAI and Anthropic cannot reverse the trend of high costs and price-sensitive customers.

  38. Assessment: UnresolvedSaid

    “The completely imaginary assumption about the AI industry is that it’ll magically get cheaper. That is not something that’s happening. More data centers will not make OpenAI or Anthropic profitable. More data centers will not make customers more willing to pay the actual cost of AI. More venture capital funding will not make Anthropic or OpenAI real businesses.”

    What we tested: The AI industry is not getting cheaper, more data centers will not make OpenAI or Anthropic profitable or make customers more willing to pay AI's actual cost, and more venture capital funding will not make Anthropic or OpenAI real businesses.

  39. Assessment: Can’t be checkedSaid

    “I agree with anyone saying there should be a pause in the development of generative AI, but I do so based on the belief that this is a doomed grift and science experiment masquerading as an industry”

    What we tested: There should be a pause in the development of generative AI, because it is a doomed grift masquerading as an industry.

  40. Assessment: Can’t be checkedSaid

    “Anthropic and OpenAI represent what I believe may be the last hypergrowth startups, and their collapse (however it may happen) will represent the end of the dream of founding a little company that turns into the next Google or Meta.”

    What we tested: He believes Anthropic and OpenAI may be the last hypergrowth startups, and their collapse, however it happens, will represent the end of the dream of founding a little company that becomes the next Google or Meta.

  41. Assessment: UnresolvedSaid

    “While OpenAI is no better, Anthropic is offensive in that it resembles everything that’s ruined the tech industry — a company with a product that costs billions of dollars that can only be sold by talking about what it might do in the future, a masterpiece of grift and hubris that I believe will stumble and crumble in the future.”

    What we tested: Anthropic will stumble and crumble in the future.

Held from this post (14)

These records did not pass every check, so nothing about them is published. Held is not a finding.