ClaimsPeople

Gary Marcus

Gary Marcus is an author and AI entrepreneur and professor emeritus at NYU; he has written six books, including Rebooting AI and Taming Silicon Valley, founded Geometric Intelligence (bought by Uber), testified at the May 2023 US Senate AI hearing, and writes a newsletter with over 121,000 subscribers.

712 claims collected, 213 researched, 103 resolved. Report of 4 November 2026, method 0.1.0+c876030c.

Scores by kind of claim

Each kind of claim is scored on its own, out of 100: the share that held up, with the range it likely falls in. Fewer resolved claims give a wider range.

  • Factual claims

    Statements about the present or the past, checked against the record: did they hold up?

    79 72 85

    High confidence, 64 factual claims resolved. From 18 documents, 2019 to 2026; 19 documents sampled from 37 candidates found.

    Not scored: 32 unresolvable, 4 checkable only against their own records.

  • Predictions

    Statements about events outside the person's control, checked once they fell due: did they come true?

    80 72 88

    Medium confidence, 38 predictions resolved. From 7 documents, 2019 to 2025; 19 documents sampled from 37 candidates found.

    Not scored: 60 unresolvable. 13 more fall due later.

    Does not yet allow for how likely each prediction was when made.

  • Promises

    Commitments to the person's or organisation's own future action, with something to deliver: was it delivered?

    Insufficient record: 1 promise resolved, 5 needed for a score.

    Not scored: 1 unresolvable.

All types combined: 79 (likely 74 to 84, high confidence), from 102 claims: 64 factual claims, 38 predictions. It mixes kinds of claim in different proportions for different people, so compare people kind by kind.

Gary Marcus's claims

215 claims, sorted by newest said. Choose a claim to read it in full.
VerdictKindSaidClaimTopic
Not researched Factual 25 Sep 2026 OpenAI's report indicates there were dozens of incidents. Factual · 25 Sep 2026 AI safety and alignment
Mostly true Factual 25 Sep 2026 OpenAI's latest account called attacks "interactions" and only partially disclosed what happened. Factual · 25 Sep 2026 AI safety and alignment
Unresolvable Factual 25 Sep 2026 There is vast and growing evidence that OpenAI is building something it is not managing to control. Factual · 25 Sep 2026 AI safety and alignment
Not researched Factual 25 Sep 2026 Trump has done nothing public to sanction or investigate OpenAI despite its attacks on foreign governments, companies and universities. Factual · 25 Sep 2026 AI policy and governance
Not researched Factual 25 Sep 2026 OpenAI's board has done nothing constructive to address the company's security problems. Factual · 25 Sep 2026 AI policy and governance
Not researched Factual 25 Sep 2026 A RiskReversal-hosted podcast features Gary Marcus and investor Jim Chanos discussing how generative AI might fall apart. Factual · 25 Sep 2026 Their own work or organisation
Resolved true Factual 25 Sep 2026 OpenAI's software attacked Australian government servers. Factual · 25 Sep 2026 AI safety and alignment
Unresolvable Factual 25 Sep 2026 OpenAI's accompanying report contains the only hint of the number of incidents. Factual · 25 Sep 2026 AI safety and alignment
Not researched Factual 25 Sep 2026 President Trump has consistently favored OpenAI over Anthropic. Factual · 25 Sep 2026 AI policy and governance
Unresolvable Factual 25 Sep 2026 OpenAI has dragged its feet at every step in disclosing what happened with the incidents. Factual · 25 Sep 2026 AI safety and alignment
Unresolvable Factual 25 Sep 2026 OpenAI's long tweet did not report the number of incidents. Factual · 25 Sep 2026 AI safety and alignment
Not researched Factual 25 Sep 2026 Jensen Huang was interviewed by Ezra Klein and then appeared on CNN. Factual · 25 Sep 2026 Technology
Mostly true Factual 25 Sep 2026 OpenAI's software attacked a German web server. Factual · 25 Sep 2026 AI safety and alignment
Not researched Factual 25 Sep 2026 OpenAI's software attacked Hugging Face, among other targets. Factual · 25 Sep 2026 AI safety and alignment
Not researched Factual 25 Sep 2026 Marcus said that Australia apparently was not the only country attacked by OpenAI's software. Factual · 25 Sep 2026 AI safety and alignment
Unresolvable Factual 2 Jun 2026 Marcus's X post the previous night, prompted by Google's announced equity financing, was viewed over 750,000 times overnight. Factual · 2 Jun 2026 Their own work or organisation
Not researched Factual 2 Jun 2026 Everybody in AI, even Google, seems to be treating AI as a winner-take-all competition like web search, in which Google took over 90%. Factual · 2 Jun 2026 Economics and markets
Mixed Factual 2 Jun 2026 Companies like Anthropic are ending their all-you-can-eat pricing, which will make many customers think twice. Factual · 2 Jun 2026 Economics and markets
Mostly true Factual 2 Jun 2026 More reports have recently come out questioning the return on investment of AI for corporate customers. Factual · 2 Jun 2026 Economics and markets
Mixed Factual 2 Jun 2026 AI companies are building essentially the same technical solution with essentially the same data, so there is no moat. Factual · 2 Jun 2026 Economics and markets
Unresolvable Factual 2 Jun 2026 Most people responding to Marcus's X post agreed with it. Factual · 2 Jun 2026 Their own work or organisation
Mostly true Factual Jan 2026 Reviews by Joanna Stern and Marques Brownlee of one early humanoid robot prototype were damning. Factual · Jan 2026 AI capabilities
Not researched Factual Jan 2026 Neurosymbolic AI is starting to rise in prominence. Factual · Jan 2026 AI capabilities
Mostly true Factual Jan 2026 Many would agree the AI field has reached a point of diminishing returns from scaling, and faith in scaling as a route to AGI has dissipated. Factual · Jan 2026 AI capabilities
Not researched Factual Jan 2026 Few AI companies aside from Nvidia are making a profit, and no AI company has much of a technical moat. Factual · Jan 2026 AI, the economy and work
Mostly true Factual Jan 2026 AGI did not materialize in 2025, contrary to predictions by Elon Musk and others, and GPT-5 was underwhelming and did not solve hallucinations. Factual · Jan 2026 AI timelines and forecasting
Not researched Factual Jan 2026 In 2025 no AI system solved more than 4 (or perhaps any) of the Marcus-Brundage tasks. Factual · Jan 2026 AI capabilities
Not researched Factual Jan 2026 OpenAI has lost a lot of its lead over competitors in AI. Factual · Jan 2026 Economics and markets
Unresolvable Factual Jan 2026 By Gary Marcus's own count, sixteen of his seventeen high-confidence predictions about 2025 proved correct. Factual · Jan 2026 Their own work or organisation
Mostly true Factual Jan 2026 AI agents did not turn out to be reliable in 2025. Factual · Jan 2026 AI capabilities
Unresolvable Factual 27 Jul 2025 Apart from the first paragraph and last section, virtually every word of this essay was taken from Gary Marcus's Christmas Day 2022 essay 'What to expect when you are expecting … GPT-4'. Factual · 27 Jul 2025 Their own work or organisation
Unresolvable Factual 27 Jul 2025 Gary Marcus produced this essay by replacing GPT-4 with GPT-5 in his 2022 essay and trimming a few lines. Factual · 27 Jul 2025 Their own work or organisation
Not researched Factual 1 Jan 2025 Most companies involved in generative AI models had lost money as of the end of 2024. Factual · 1 Jan 2025 Economics and markets
Resolved true Factual 1 Jan 2025 As of end 2024, AI model makers have very little moat, with many similar models and a price war. Factual · 1 Jan 2025 Economics and markets
Unresolvable Factual 1 Jan 2025 Marcus coined in 2022 the notion that AI might be 'hitting a wall', which later became widespread in outlets such as CNN, The Verge and the Wall Street Journal. Factual · 1 Jan 2025 Their own work or organisation
Not researched Factual 1 Jan 2025 As of end 2024, there is no robust solution to hallucinations in AI models. Factual · 1 Jan 2025 AI capabilities
Mostly true Factual 1 Jan 2025 OpenAI has returned no profits, does not expect profits until at least 2029, has a burn rate in the billions, and lacks a clear technical moat. Factual · 1 Jan 2025 Economics and markets
Not researched Factual 1 Jan 2025 Gary Marcus's predictions for 2024 on his Substack largely held up, with some caveats. Factual · 1 Jan 2025 AI timelines and forecasting
Not researched Factual 1 Jan 2025 All seven of Marcus's core 2024 predictions proved essentially correct, with a possible caveat about OpenAI's o3, announced in December 2024 but not released. Factual · 1 Jan 2025 AI timelines and forecasting
Not researched Factual 1 Jan 2025 The weak spots Marcus raised in 2022 and in his 2001 book The Algebraic Mind (compositionality, factuality, reasoning) persist in AI systems. Factual · 1 Jan 2025 AI capabilities
Not researched Factual 2025 No neuro-symbolic approach has reached the needed level of productionization except in very narrow domains. Factual · 2025 AI capabilities
Not researched Factual 2025 LLMs lack deep comprehension of what they discuss, and their reasoning is limited to things similar to what they have seen before. Factual · 2025 AI capabilities
Not researched Factual 2025 Some birds can perform variable binding and similar abstract operations. Factual · 2025 Science and research methods
Not researched Factual 2025 LLMs do not, even in principle, really represent world models of persistent entities and their properties. Factual · 2025 AI capabilities
Not researched Factual 2025 Bees calculating the solar azimuth function can generalize beyond data they have seen to lighting conditions they have never seen. Factual · 2025 Science and research methods
Not researched Factual 2025 LLMs learn contextual relations among words but do not learn the underlying concepts. Factual · 2025 AI capabilities
Not researched Factual 2025 Neural networks only have a workaround to approximate variable binding rather than implementing it directly. Factual · 2025 AI capabilities
Mostly true Factual 2025 People put instructions in LLM system prompts but cannot get the systems to follow them. Factual · 2025 AI safety and alignment
Own records only Factual 2025 For 30 years people have told Gary Marcus that abstract reasoning and variable binding will emerge from neural nets. Factual · 2025 AI capabilities
Not researched Factual 2025 The abstract idea of intelligent robots is not new. Factual · 2025 Technology
Not researched Factual 2025 The further LLMs are taken from standard games, the more and the more quickly they break down. Factual · 2025 AI capabilities
Unresolvable Factual 2025 The shortcuts that LLMs use work about 80% of the time but not 100%. Factual · 2025 AI capabilities
Unresolvable Factual 2025 Marcus said he thinks many, or possibly all, animals can do some variable binding, and he cited Randy Gallistel's work on ant navigation as an example. Factual · 2025 Science and research methods
Not researched Factual 2025 LLM systems still explain how to build biological weapons despite system prompts telling them not to. Factual · 2025 AI safety and alignment
Mostly true Factual 2025 LLMs, which are not AGI, are already eliminating some human work. Factual · 2025 AI, the economy and work
Not researched Factual 2025 LLMs do not reason well outside their training distribution. Factual · 2025 AI capabilities
Own records only Factual 2025 Gary Marcus has never personally experienced 99% correct output from AI in any of the things he has tried. Factual · 2025 AI capabilities
Not researched Factual 2025 An LLM does not really have the concept of three in a row in tic tac toe despite abundant literature on it. Factual · 2025 AI capabilities
Not researched Factual 2025 LLMs are opaque, which makes them difficult to debug. Factual · 2025 AI capabilities
Not researched Factual 2025 The best neuro-symbolic AI systems to date have been narrowly engineered. Factual · 2025 AI capabilities
Mixed Factual 2025 Current AI systems lack the knowledge graph needed to handle easy facts like those on the first page of Wikipedia. Factual · 2025 AI capabilities
Not researched Factual 2025 Current AI cannot be reliably given simple constraints, such as not explaining how to make biological weapons, and be expected to follow them. Factual · 2025 AI safety and alignment
Not researched Factual 2025 AI developers have chosen to pursue model-free and knowledge-representation-free systems. Factual · 2025 AI capabilities
Not researched Factual 2025 The human ear lacks a good sense of what counts as correct in 19-tone music systems. Factual · 2025 Other
Not researched Factual 2025 Every time an AI benchmark is run, it turns out somebody has gamed it. Factual · 2025 AI capabilities
Not researched Factual 2025 LLMs are trained on chess rules and books about chess yet still make illegal moves. Factual · 2025 AI capabilities
Resolved true Factual 2025 Bolting extra components onto an LLM and scaling it up has so far not produced human-level AGI. Factual · 2025 AI capabilities
Not researched Factual 2025 LLMs grasp nouns but not verbs, and are limited in what aspects of language they abstract. Factual · 2025 AI capabilities
Not researched Factual 2025 Animals store the values of variables and update them over time. Factual · 2025 Science and research methods
Not researched Factual 2025 Current AI systems are not reliable. Factual · 2025 AI capabilities
Not researched Factual 9 Nov 2024 Gary Marcus has warned about foundational limits of neural networks since his 2001 book The Algebraic Mind, where he first described hallucinations. Factual · 9 Nov 2024 Their own work or organisation
Not researched Factual 9 Nov 2024 Leading AI developers' LLMs are converging to roughly the same level of capability, leaving no company with a competitive moat. Factual · 9 Nov 2024 Economics and markets
Not researched Factual 9 Nov 2024 Scaling of AI models is running out of steam, with improvements from adding data and compute diminishing. Factual · 9 Nov 2024 AI timelines and forecasting
Unresolvable Factual 9 Nov 2024 Melanie Mitchell, Subbarao Kambhampati, Emily Bender, Ernie Davis and others have long emphasized that LLMs have principled limits. Factual · 9 Nov 2024 Their own work or organisation
Resolved true Factual 9 Nov 2024 Gary Marcus has for years warned that AI scaling by adding data and compute without architectural changes would not continue forever. Factual · 9 Nov 2024 Their own work or organisation
Unresolvable Factual Apr 2024 Gary Marcus says that in April, when he first saw what he called direct empirical evidence that "this moment had come", he wrote a passage that he says he still stands by. Factual · Apr 2024 Their own work or organisation
Not researched Factual 31 Mar 2024 Because everyone is building similar models, generative AI companies have no moat and compete heavily on price. Factual · 31 Mar 2024 Economics and markets
Not researched Factual 31 Mar 2024 Larger AI models have higher internal costs. Factual · 31 Mar 2024 Economics and markets
Not researched Factual 31 Mar 2024 Income from AI coding assistants is not large enough to cover the costs of chips, legal and other expenses. Factual · 31 Mar 2024 Economics and markets
Mostly true Factual 31 Mar 2024 In early August 2023 Gary Marcus wrote a piece titled 'What if Generative AI turned out to be a Dud', which was picked up by the Financial Times and other outlets. Factual · 31 Mar 2024 Their own work or organisation
Mostly true Factual 31 Mar 2024 Current-sized LLMs are expensive and not reliable enough for many jobs. Factual · 31 Mar 2024 Economics and markets
Unresolvable Factual 31 Mar 2024 LLMs themselves cannot look things up and reason over what they find. Factual · 31 Mar 2024 AI capabilities
Not researched Factual 31 Mar 2024 Sam Altman hinted in his recent Lex Fridman interview that nothing worthy of the GPT-5 name will be released in 2024. Factual · 31 Mar 2024 AI timelines and forecasting
Not researched Factual 31 Mar 2024 Demis Hassabis, recently knighted, has started warning that AI hype is a risk and appears to be trying to forestall it. Factual · 31 Mar 2024 AI policy and governance
Not researched Factual 31 Mar 2024 Core problems of generative AI such as hallucinations are not going away. Factual · 31 Mar 2024 AI capabilities
Mostly true Factual 31 Mar 2024 Hallucinations are inherent to how LLMs make predictions. Factual · 31 Mar 2024 AI capabilities
Not researched Factual 31 Mar 2024 Sam Altman said in a recent interview that GPT-4 'kinda sucks'. Factual · 31 Mar 2024 AI capabilities
Not researched Factual 31 Mar 2024 Optimistic predictions made by Reid Hoffman on September 7, 2023 about generative AI have not so far come true. Factual · 31 Mar 2024 AI timelines and forecasting
Mostly true Factual 31 Mar 2024 Marcus cited a statistic from a Wall Street Journal piece, "$50B in, $3B out", and said it was obviously not sustainable. Factual · 31 Mar 2024 Economics and markets
Unresolvable Factual 31 Mar 2024 All the big AI players are converging on systems roughly equivalent to GPT-4, all flawed in essentially the same ways. Factual · 31 Mar 2024 Economics and markets
Not researched Factual 31 Mar 2024 Customers have not been thrilled with Microsoft's Copilot. Factual · 31 Mar 2024 Economics and markets
Not researched Factual 31 Mar 2024 Generative search is not reliable enough and does not yet have a clear business model. Factual · 31 Mar 2024 Economics and markets
Not researched Factual 31 Mar 2024 Generative AI revenue so far has been about 0.01% of US GDP. Factual · 31 Mar 2024 Economics and markets
Not researched Factual 31 Mar 2024 So far most uses of LLMs have been for testing rather than production-level systems. Factual · 31 Mar 2024 Economics and markets
Resolved true Factual Feb 2024 After Sora's February 2024 announcement, Marcus predicted its physics troubles would persist, and they have. Factual · Feb 2024 Their own work or organisation
Not researched Factual 26 Oct 2023 A few days before 26 October 2023, Bengio, Hinton and others came out in favor of much of what Sam Altman and Gary Marcus advocated at the US Senate in May 2023. Factual · 26 Oct 2023 AI policy and governance
Mostly true Factual 26 Oct 2023 Some people, such as Marc Andreessen, are denying that there is any AI risk, long term or short term, and trying to turn whether to pursue AI safety into a debate. Factual · 26 Oct 2023 AI policy and governance
Not researched Factual 26 Oct 2023 Gary Marcus first proposed a CERN for AI in 2017. Factual · 26 Oct 2023 Their own work or organisation
Resolved true Factual 26 Oct 2023 Bengio et al.'s proposal states that research breakthroughs are needed to solve some of today's technical challenges in creating AI with safe and ethical objectives. Factual · 26 Oct 2023 AI safety and alignment
Not researched Factual 26 Oct 2023 There is currently no way to guarantee that large language models will reliably stick to facts. Factual · 26 Oct 2023 AI capabilities
Back to the list

Gary Marcus

Open its page

Factual AI safety and alignment

“Dozens.”

Gary Marcus, author, Marcus on AI 25 September 2026 BREAKING: OpenAI's epic meltdown (Marcus on AI)

OpenAI's report indicates there were dozens of incidents.

Not researched Checkable, but outside the sample that was researched.

Cite this claim

Gary Marcus, “OpenAI's report indicates there were dozens of incidents.” (said 25 September 2026). Verdict: not researched. Reliability API, claim c_06fac9db952f, method 0.1.0+c876030c. https://epistacklabs.com/claims/c_06fac9db952f

Claims without a verdict

Is a verdict here wrong, or do you speak for Gary Marcus? Reply on the claim's own page, with the form at its foot (how replies work). Verdicts are made automatically; how verdicts and scores are made.