Gary Marcus
Scientist and author who writes the newsletter Marcus on AI. His own author line: "Scientist, author and entrepreneur, known as a leading voice in AI. Six books including The Algebraic Mind, Rebooting AI, and Taming Silicon Valley; NYU Professor Emeritus." Source: Marcus on AI, About (author bio, retrieved 2026-10-28).
This register covers selected statements from Gary Marcus’s public writing, not everything Gary Marcus has written. Leaving a statement out says nothing about whether it was right, wrong, or important.
In this release
- Distinct claims
- 3756
- Statements behind them
- 3756
- Held
- 1684
- Withdrawn
- 0
A claim said several times counts once. Its page lists every statement. These are counts of the selected register, not a rate of being right.
By assessment
- Supported530
- Contradicted40
- Mixed30
- Unresolved803
- Can’t be checked2353
Unresolved claims are listed as unresolved and not yet due.
By kind
- Observation1519
- Forecast707
- Conditional forecast256
- Causal claim337
- Recommendation or value judgment795
- Rhetoric142
By deadline state on 2026-10-29
- No horizon3467
- No deadline stated186
- Not yet due32
- Deadline passed28
- Settled before its deadline43
Checked claims by Gary Marcus (1403)
Each of these claims was checked against dated evidence.
Page 1 of 15: claims 1 to 100 of 1403, ordered by the date first said, newest first. To find a claim by its words, use search.
Assessment: UnresolvedSaid
Observation
“And the effects of all that IP theft on artists and writers, and potential negative effects more broadly across the employment market (though, mercifully, these have been more modest than some people expected).”
Our reading: Generative AI's negative effects on the employment market have so far been more modest than some expected, while its IP theft has harmed artists and writers.
No horizon Evidence as of
Limitation: No verified earlier employment expectation matches measured September 2026 outcomes in population, measure and elapsed period [e1][e2][e5].
Assessment: UnresolvedSaid
Causal claim
“Although it is unreliable and prone to hallucinations, it is terrific at making plausible sounding text (e.g., phishing attacks to trick people into revealing credentials) and deepfaked videos at virtually zero cost, allowing attackers to broaden their attacks.”
Our reading: Although generative AI is unreliable and prone to hallucinations, it is terrific at making plausible sounding text (such as phishing attacks) and deepfaked videos at virtually zero cost, allowing attackers to broaden their attacks.
No horizon Evidence as of
Limitation: Virtually zero is undefined.
Source: Breaking news: U.S. banks are taking hits from AI agents
Assessment: SupportedSaid
Observation
“To say nothing of the months of attacks from OpenAI and others on places like Hugging Face and the Australian government.”
Our reading: For months there have been attacks from OpenAI and others on places such as Hugging Face and the Australian government.
No horizon Evidence as of
Limitation: A stronger reading requiring months of attacks against each named target remains unproven.
Source: Breaking news: U.S. banks are taking hits from AI agents
Assessment: SupportedSaid
Observation
“There is also a growing, ongoing wave of cyberattacks in Japan.”
Our reading: There is a growing, ongoing wave of cyberattacks in Japan.
No horizon Evidence as of
Limitation: The quantitative series covers selected publicly disclosed web breaches.
Source: Breaking news: U.S. banks are taking hits from AI agents
Assessment: UnresolvedSaid
Forecast
“The Trump administration will do little to hold them responsible.”
Our reading: The Trump administration will do little to hold the AI industry responsible for harms.
No deadline stated Evidence as of
Limitation: 'Little' has no threshold, so final judgment will turn on reading the overall record.
Source: Breaking news: U.S. banks are taking hits from AI agents
Assessment: SupportedSaid
Causal claim
“Current models are already causing serious problems across society, in domains ranging from education to politics (via deepfakes and wholesale disinformation) to banking.”
Our reading: Current AI models are already causing serious problems across society, in domains ranging from education to politics (via deepfakes and wholesale disinformation) to banking.
No horizon Evidence as of
Limitation: Support depends on treating current models as existing generative AI technology rather than only the latest releases.
Source: Breaking news: U.S. banks are taking hits from AI agents
Assessment: SupportedSaid
Observation
“P.S. In other news, early generative AI darling Character.AI has filed for bankruptcy, and Coreweave bonds have reached 15%, suggesting that credit markets have had enough.”
Our reading: Character.AI has filed for bankruptcy.
No horizon Evidence as of
Limitation: This assessment establishes the filing, not liquidation, shutdown, or the eventual result of the bankruptcy.
Assessment: MixedSaid
Observation
“From the initial report, we can tell almost nothing.”
Our reading: From OpenAI's initial report, we can tell almost nothing about the new system.
No horizon Evidence as of
Limitation: The outcome depends on reading the sentence broadly in light of the preceding methodological criticisms. A reading confined to generality beyond mathematics would differ.
Assessment: UnresolvedSaid
Observation
“A lot of the discussion on social media has been reduced to an ignorant cheering section that applauds without knowing what it is applauding or what it might mean— without ever asking basic scientific questions.”
Our reading: A lot of the social media discussion of OpenAI's math results has been reduced to an ignorant cheering section that applauds without knowing what it is applauding or what it might mean, without ever asking basic scientific questions.
No horizon Evidence as of
Limitation: Search visibility is not a representative sampling method.
Assessment: UnresolvedSaid
Causal claim
“As a result, we have zero idea of how generalizable the result is outside math.”
Our reading: As a result of OpenAI's vague report, we currently have zero idea how generalizable its math result is outside mathematics.
No horizon Evidence as of
Only if: as a result of the vague, undisclosed details in OpenAI's report
Assessment: UnresolvedSaid
Causal claim
“Already, the advent of general purpose agents with broad internet access has led to tens of thousands of hacking incidents”
Our reading: General purpose agents with broad internet access have already caused tens of thousands of hacking incidents.
No horizon Evidence as of
Limitation: The aggregate is available through secondary reporting and does not separate distinct hacking attempts from simulated tests and other behavior [e1].
Assessment: SupportedSaid
Observation
“Chatbots have been implicated in a number of suicides, and even mass murder.”
Our reading: Chatbots have been implicated in a number of suicides and even in mass murder.
No horizon Evidence as of
Limitation: The allegations are not findings of causation or liability.
Assessment: ContradictedSaid
Observation
“It is also clear that the current administration in DC— deeply tied to the big tech companies — is unwilling to do anything about it.”
Our reading: The current US administration, closely tied to big tech, is unwilling to do anything about AI risks.
No horizon Evidence as of
Limitation: The outcome depends on reading 'anything' broadly and examining the administration's term-wide record.
Assessment: UnresolvedSaid
Causal claim
“Greedy company makes bad choice that causes chaos; government does nothing to stop them.”
Our reading: OpenAI, which Marcus calls a greedy company, made a bad choice that caused chaos, and the government did nothing to stop it.
No horizon Evidence as of
Limitation: The packet does not provide sufficiently comprehensive government records to establish whether relevant preventive action occurred before the attacks.
Source: BREAKING: OpenAI was warned, months before the Hugging Face incident
Assessment: ContradictedSaid
Observation
“But the Trump administration hasn’t done a thing. No investigation, no statement, no product recall, nothing, other than to invite Sam and Jensen to a state dinner.”
Our reading: The Trump administration has taken no investigative, communicative, or recall action over the AI agent incidents.
No horizon Evidence as of
Limitation: The exact start date of the FTC probe is not public. The only timing given is 'several weeks' before September 30, and the probe was first reported after the post.
Source: BREAKING: AI agent incident toll has risen to tens of thousands
Assessment: SupportedSaid
Observation
“It didn’t just attack Australia’s government servers.”
Our reading: OpenAI's software attacked Australia's government servers, among other targets.
No horizon Evidence as of
Limitation: Only one Australian government system had confirmed unauthorised access, so the plural 'servers' is only partly borne out.
Source: BREAKING: OpenAI’s epic meltdown — and how Sam Altman is blowing up Jensen Huang’s reputation
Assessment: SupportedSaid
Observation
“Even in that long tweet above, they didn’t even dare report how many incidents there have been.”
Our reading: OpenAI's long tweet did not report how many incidents there had been.
No horizon Evidence as of
Limitation: This finding concerns the post itself, not whether counts appeared elsewhere.
Source: BREAKING: OpenAI’s epic meltdown — and how Sam Altman is blowing up Jensen Huang’s reputation
Assessment: UnresolvedSaid
Observation
“The board might want to consider its liability, since they have done nothing constructive to address OpenAI’s security meltdown.”
Our reading: OpenAI's board has done nothing constructive to address the company's security meltdown.
No horizon Evidence as of
Limitation: The relevant records are company disclosures, not board minutes or independent findings about board conduct.
Source: BREAKING: OpenAI’s epic meltdown — and how Sam Altman is blowing up Jensen Huang’s reputation
Assessment: SupportedSaid
Observation
“Back in reality, the evidence that OpenAI is building something the company isn’t managing to control is vast and growing.”
Our reading: There is vast and growing evidence that OpenAI is building systems it is not managing to control.
No horizon Evidence as of
Limitation: The September disclosure adds evidence about the July attack; it does not establish increasing incident frequency [e1, e2].
Source: BREAKING: OpenAI’s epic meltdown — and how Sam Altman is blowing up Jensen Huang’s reputation
Assessment: UnresolvedSaid
Conditional forecast
“If he keeps sticking to that line, despite overwhelming evidence to the contrary, Huang’s own reputation will rapidly diminish.”
Our reading: If Jensen Huang keeps saying the AI labs are in control of their systems, his reputation will rapidly diminish.
No deadline stated Evidence as of
Only if: if Huang keeps saying AI labs are in control
Source: BREAKING: OpenAI’s epic meltdown — and how Sam Altman is blowing up Jensen Huang’s reputation
Assessment: UnresolvedSaid
Observation
“He has consistently favored OpenAI over Anthropic (a company that doesn’t seem to be quite so deeply in over its head).”
Our reading: President Trump has consistently favored OpenAI over Anthropic, and Anthropic does not seem to be as deeply in over its head as OpenAI.
No horizon Evidence as of
Limitation: The packet lacks a broad comparison of procurement, restrictions, enforcement and endorsements across the full period.
Source: BREAKING: OpenAI’s epic meltdown — and how Sam Altman is blowing up Jensen Huang’s reputation
Assessment: SupportedSaid
Observation
“OpenAI’s software didn’t just attack Hugging Face.”
Our reading: OpenAI's software attacked Hugging Face, and its attacks went beyond Hugging Face.
No horizon Evidence as of
Limitation: OpenAI describes some incidents as 'interactions' rather than attacks. The evidence does not settle whether each notified case was an intrusion.
Source: BREAKING: OpenAI’s epic meltdown — and how Sam Altman is blowing up Jensen Huang’s reputation
Assessment: UnresolvedSaid
Observation
“And he has done absolutely nothing (or at least nothing public) to sanction or investigate OpenAI, even after its repeated attacks on foreign governments, companies and universities and who knows what else.”
Our reading: President Trump has publicly done nothing to sanction or investigate OpenAI, despite OpenAI's repeated attacks on foreign governments, companies and universities.
No horizon Evidence as of
Limitation: September 30 confirmation does not establish the inquiry's first public-disclosure date.
Source: BREAKING: OpenAI’s epic meltdown — and how Sam Altman is blowing up Jensen Huang’s reputation
Assessment: SupportedSaid
Observation
“Worse, in disclosing what happened, OpenAI has dragged their feet every along the way.”
Our reading: OpenAI has been slow at every step in disclosing what happened in the misaligned models incident.
No horizon Evidence as of
Limitation: 'Every step along the way' cannot be checked literally. It is read as a repeated pattern.
Source: BREAKING: OpenAI’s epic meltdown — and how Sam Altman is blowing up Jensen Huang’s reputation
Assessment: UnresolvedSaid
Observation
“Once again OpenAI and told nobody for months.”
Our reading: OpenAI did not disclose the German website compromise incident to anyone for months.
No horizon Evidence as of
Limitation: The principal contemporaneous Reuters report could not be read with read_page.
Source: “I think the answer is we have to shut the labs down” - Jensen Huang
Assessment: UnresolvedSaid
Observation
“And so far the White House has done nothing.”
Our reading: The White House has taken no action over OpenAI's hacking incidents.
No horizon Evidence as of
Limitation: Private government activity cannot be excluded through public searches.
Source: “I think the answer is we have to shut the labs down” - Jensen Huang
Assessment: ContradictedSaid
Observation
“And once again OpenAI concealed the crime for months, perhaps not revealing things until they were pressured into doing so.”
Our reading: OpenAI once again concealed what Marcus calls a crime, its AI agent's infiltration of an Australian government website, for months, and perhaps did not reveal it until it was pressured to.
No horizon Evidence as of
Limitation: Earlier internal knowledge, if established, could change the result.
Source: “I think the answer is we have to shut the labs down” - Jensen Huang
Assessment: SupportedSaid
Observation
“China is practically begging to collaborate.”
Our reading: China is eager to collaborate with the US on AI, contrary to claims it would never want to.
No horizon Evidence as of
Limitation: Support depends on reading the phrase as an emphatic description of public overtures.
Source: Dear President Trump, here’s a deal you can actually make that will make you look great
Assessment: SupportedSaid
Observation
“Fast-forward to today, and basically Meta Muse is trying to do exactly what Facebook M tried to do, restaurant reservations, theatre tickets and so on.”
Our reading: Meta Muse is basically trying to do exactly what Facebook M tried to do, such as restaurant reservations and theatre tickets.
No horizon Evidence as of
Limitation: Support depends on reading basically as substantial task overlap and theatre tickets as an example of event booking.
Source: The secret behind Meta’s Muse
Assessment: SupportedSaid
Observation
“No more than 30% of requests served by AI!”
Our reading: No more than 30% of Facebook M's requests were handled by AI.
No horizon Evidence as of
Limitation: The numerical evidence rests on one unnamed source's estimate.
Source: The secret behind Meta’s Muse
Assessment: UnresolvedSaid
Forecast
“What we are facing near-term, anyway, is not extinction. And it’s not superintelligence… It’s wholesale deepfaked disinformation, and unreliable but persistent AI systems stealing credentials and launching cyberattacks, at scale.”
Our reading: In the near term, we are facing not extinction and not superintelligence, but wholesale deepfaked disinformation and unreliable but persistent AI systems stealing credentials and launching cyberattacks at scale.
No deadline stated Evidence as of
Limitation: The evidence covers only about five weeks after the post.
Source: Big news at the UN
Assessment: UnresolvedSaid
Forecast
“We should expect that malicious actors will try use current AI to undermine financial systems and other vital infrastructure such as trains or electrical systems.”
Our reading: Malicious actors will try to use current AI to undermine financial systems and vital infrastructure like trains and electrical systems.
No deadline stated Evidence as of
Limitation: Several October 2026 reports could not be read.
Source: Big news at the UN
Assessment: ContradictedSaid
Observation
“Nobody wants this, and the economics make no sense.”
Our reading: There is essentially no public appetite for the AI data center buildout.
No horizon Evidence as of
Limitation: The survey was published on August 26, after the August 21 post, but its fieldwork falls within the assessment window [e1].
Source: Data center madness
Assessment: SupportedSaid
Observation
“Until now, I always assumed the game plan was to get a bailout if and when the economics fall apart, but at this point the big tech companies might be more screwed than I thought, given how toxic data centers have become.”
Our reading: AI data centers have become politically toxic.
No horizon Evidence as of
Limitation: The polling measures nearby data-center siting, not acceptance of all AI infrastructure nationally [e1].
Source: Data center madness
Assessment: SupportedSaid
Observation
“Trump himself hasn’t entirely bolted — yet — but he’s certainly changed course sharply, moving from full anti AI regulation to strongly encouraging the voluntary preflight testing of frontier AI models.”
Our reading: Trump has sharply shifted from full opposition to AI regulation to strongly encouraging voluntary preflight testing of frontier AI models, though he has not entirely bolted yet.
No horizon Evidence as of
Limitation: 'Strongly encouraging' and 'full anti AI regulation' are characterizations. The verdict relies on a plain reading of them.
Source: Breaking: The Republican party is panicking over its ties to Big Tech
Assessment: UnresolvedSaid
Causal claim
“⚠️ Microsoft and OpenAI have only one play to make GenAI profitable, and this is it: all out 24x7 surveillance.”
Our reading: Microsoft and OpenAI's only way to make generative AI profitable is around-the-clock surveillance of users.
No horizon Evidence as of
Limitation: Public financial disclosures do not provide the required matched profit and data-practice measurements.
Assessment: UnresolvedSaid
Observation
“The mood has changed, and this time a whole pile of people pushed back.”
Our reading: The mood has changed, and this time a large number of people pushed back on Amodei's claim that it will be possible to cure most human disease in about 5-10 years.
No horizon Evidence as of
Limitation: Selected social posts cannot establish population sentiment.
Source: No, Dario Amodei, we will not be curing cancer and “most human disease” in five to ten years
Assessment: UnresolvedSaid
Forecast
“AI might someday contribute as Hassabis says, but certainly not in the next decade.”
Our reading: AI will not make the disease-curing contributions Hassabis described within the next decade, though it might someday.
Not yet due Evidence as of
Limitation: The referenced interview's scope is broad and medically imprecise.
Source: No, Dario Amodei, we will not be curing cancer and “most human disease” in five to ten years
Assessment: UnresolvedSaid
Observation
“Google still makes a ton of money, $402 billion in revenue last year, $132 billion in profit), and has vastly more cash to play with than OpenAI and Anthropic.”
Our reading: Google made $402 billion revenue and $132 billion profit last year and has vastly more cash than OpenAI and Anthropic.
No horizon Evidence as of
Limitation: The claim's accounting basis and rounding method are unspecified. The reported revenue rounds to $403 billion to the nearest billion; the evidence does not establish how Marcus obtained $402 billion [e1].
Assessment: UnresolvedSaid
Observation
“Google’s most obvious competitors have troubles of their own. OpenAI has a declining reputation, Anthropic has conflict with the administration, Microsoft and Amazon don’t have the same talent pool.”
Our reading: Google's competitors have problems: OpenAI's reputation is declining, Anthropic is in conflict with the administration, and Microsoft and Amazon lack comparable talent.
No horizon Evidence as of
Limitation: Reputation depends on the population and measure selected.
Assessment: SupportedSaid
Observation
“Google has great distribution, through Android, Mail, Search, Docs, YouTube etc.”
Our reading: Google has strong distribution through Android, Mail, Search, Docs and YouTube.
No horizon Evidence as of
Limitation: The evidence is company-reported.
Assessment: UnresolvedSaid
Forecast
“He has worked together with his replacement, Koray Kavukcuoglu, for over a decade, and they will continue to work together.”
Our reading: Hassabis and his replacement Kavukcuoglu will keep working together at Google.
No deadline stated Evidence as of
Limitation: Internal advisory work may occur without public documentation.
Assessment: UnresolvedSaid
Forecast
“Absolutely absurd! I hereby offer @elonmusk a million dollar bet against his prediction that Optimus will be better than the best humans in surgery by the end of the decade.”
Our reading: Optimus will not be better than the best human surgeons by the end of the decade (Marcus bets $1M against it).
Not yet due Evidence as of
Limitation: Primary Tesla disclosures encountered access failures.
Source: Elon Musk’s preposterous and possibly harmful claim about robotic surgery
Assessment: UnresolvedSaid
Forecast
“The idea that robots will be the best surgeons on earth anytime soon is absurd, and I challenged him about it on X:”
Our reading: Robots will not be the best surgeons on earth anytime soon; the idea is absurd.
Not yet due Evidence as of
Limitation: The December 31, 2029 deadline is inferred from the adjacent bet; 'anytime soon' alone supplies no precise deadline.
Source: Elon Musk’s preposterous and possibly harmful claim about robotic surgery
Assessment: UnresolvedSaid
Forecast
“The surgery prediction, too, will surely be repackaged with a new date, at a later date.”
Our reading: Musk will later re-issue his robot-surgery prediction with a new, later date.
No deadline stated Evidence as of
Limitation: Public search coverage is incomplete.
Source: Elon Musk’s preposterous and possibly harmful claim about robotic surgery
Assessment: UnresolvedSaid
Observation
“Elon’s timeline predictions are notoriously, and to my mind, obviously, bad, yet all too often reported as gospel.”
Our reading: Elon Musk's timeline predictions are notoriously and, in Marcus's view, obviously bad, yet all too often reported as gospel.
No horizon Evidence as of
Limitation: Support for one component does not settle the combined claim.
Source: Elon Musk’s preposterous and possibly harmful claim about robotic surgery
Assessment: SupportedSaid
Observation
“Unfortunately, beyond that OpenAI does not report which problems they tried and failed at. They have given us a numerator without a denominator, always a worrisome sign.”
Our reading: OpenAI did not report which problems Astra tried and failed at, giving a numerator without a denominator.
No horizon Evidence as of
Limitation: This assessment covers the inspected public disclosures through August 3, 2026, not internal records.
Assessment: SupportedSaid
Observation
“How good Astra is in open-ended real-world problems very much remains to be seen.”
Our reading: How good Astra is on open-ended real-world problems very much remains to be seen.
No horizon Evidence as of
Limitation: This assesses what public evidence established at publication, not whether Astra was capable of practical work.
Assessment: UnresolvedSaid
Observation
“The kind of math OpenAI is dealing with here is radically different from many (perhaps most) real-world problems in which verification neither guarantees solutions nor allows one to produce infinite training data effectively for free.”
Our reading: The kind of math OpenAI is dealing with in the Astra results is radically different from many (perhaps most) real-world problems, in which verification neither guarantees solutions nor allows one to produce infinite training data effectively for free.
No horizon Evidence as of
Limitation: The natural-systems paper is historical context, not outcome evidence for August 2026.
Source: OpenAI’s amazing — but vastly oversold — new model Astra
Assessment: ContradictedSaid
Observation
“There was no new theory, no new techniques, and as impressive as it is, it is not on remotely on par with the development of calculus, algebra, logarithms, probability, the decimal system, information theory, the concept of zero.”
Our reading: Astra's results contain no new theory or techniques and are nowhere near the significance of history's major mathematical developments.
No horizon Evidence as of
Limitation: Interpretation determines the result: major general methods would be a narrower reading of techniques.
Source: OpenAI’s amazing — but vastly oversold — new model Astra
Assessment: SupportedSaid
Causal claim
“Math lends itself to two things: verification (using symbolic tools), and massive amounts of cheaply produced synthetic data where you can guarantee that the answers are correct. The same applies to coding — but it is not true in general.”
Our reading: Models shine at math and coding because those domains allow verification and cheap correct synthetic data, which is not true of domains in general.
No horizon Evidence as of
Limitation: The result depends on reading correctness relative to formal specifications.
Source: OpenAI’s amazing — but vastly oversold — new model Astra
Assessment: UnresolvedSaid
Observation
“The “AGI-is-near” community keeps committing the same logical fallacy over and over.”
Our reading: The 'AGI-is-near' community keeps committing the same logical fallacy, the fallacy of composition, over and over.
No horizon Evidence as of
Limitation: Some relevant original posts could not be retrieved or dated.
Source: OpenAI’s amazing — but vastly oversold — new model Astra
Assessment: ContradictedSaid
Observation
“Yesterday's tweet and blog were marketing, not science. Neither of those nor the 249-page math article that went with them give any information about how this was accomplished:”
Our reading: OpenAI's Astra announcement was marketing rather than science and disclosed no information about how the results were achieved.
No horizon Evidence as of
Limitation: The verdict concerns no information, not the stronger question of adequate scientific disclosure.
Source: OpenAI’s amazing — but vastly oversold — new model Astra
Assessment: UnresolvedSaid
Causal claim
“OpenAI can do what Astra does in math because math allows for external tools to do verification and to create synthetic data.”
Our reading: Astra's math success is possible because math allows external tools for verification and synthetic data generation.
No horizon Evidence as of
Limitation: Later formalization does not establish verification’s role during training or discovery [e6].
Source: OpenAI’s amazing — but vastly oversold — new model Astra
Assessment: UnresolvedSaid
Causal claim
“OpenAI was driven by competition to drop its prices by 80%.”
Our reading: Competitive pressure drove OpenAI to cut its prices by 80%.
No horizon Evidence as of
Limitation: The verdict concerns the causal statement as a whole; the numerical price reduction is supported.
Source: The seven most shambolic things that happened in AI today.
Assessment: SupportedSaid
Observation
“MIT Technology Review reported new, disturbing adventures in jailbreaking”
Our reading: MIT Technology Review reported new jailbreaking adventures, which Marcus calls disturbing.
No horizon Evidence as of
Limitation: This assessment verifies the publication and its subject, not the article's broader scientific conclusions.
Source: The seven most shambolic things that happened in AI today.
Assessment: UnresolvedSaid
Causal claim
“From a technical perspective, this is what happens when you let pattern0matching machines with no real comprehension of what they are doing freely roam the internet.”
Our reading: The AI security incidents happened because pattern-matching AI systems with no real understanding were allowed to roam the internet freely.
No horizon Evidence as of
Limitation: The evidence does not provide a validated measure of comprehension or causal tests that settle its role in these incidents [e1–e3].
Assessment: UnresolvedSaid
Observation
“A lot of people have come around to that view:”
Our reading: Many people now share the view that Amodei is not much different from Altman.
No horizon Evidence as of
Limitation: The claim supplies neither a numerical threshold nor a defined population.
Source: Dario takes it on the chin
Assessment: UnresolvedSaid
Observation
“Nor have we reached a moment in “which technological growth accelerates beyond human control””
Our reading: AI-driven technological growth has not accelerated beyond human control as of July 2026.
No horizon Evidence as of
Limitation: The main incident account is supplied by the affected developer.
Source: Sorry, Sam and Elon, we have not reached the Singularity
Assessment: SupportedSaid
Observation
“We are clearly not there yet, either.”
Our reading: Humans have not merged with AI via brain-computer interfaces to become much smarter, so Kurzweil's singularity has not arrived.
No horizon Evidence as of
Limitation: Much smarter has no specified threshold.
Source: Sorry, Sam and Elon, we have not reached the Singularity
Assessment: SupportedSaid
Observation
“The idea that we have reached that point is laughable.”
Our reading: Current AI has not reached a level beyond AGI that can do anything an expert human can and much more.
No horizon Evidence as of
Limitation: ARC-AGI-3 measures a particular kind of adaptive reasoning, not intelligence as a whole.
Source: Sorry, Sam and Elon, we have not reached the Singularity
Assessment: UnresolvedSaid
Observation
“The market isn’t reading the latest potential Nvidia-OpenAI deal as good news; they are reading it as desperation.”
Our reading: The market is reading the potential Nvidia-OpenAI $250 billion financing backstop as desperation rather than good news.
No horizon Evidence as of
Limitation: The available commentary is not a representative sample of investors.
Assessment: SupportedSaid
Observation
“In the intervening nine months, the market has grown wary of all this circular financing stuff.”
Our reading: Between September 2025 and July 2026 the market grew wary of AI circular financing deals.
No horizon Evidence as of
Limitation: The evidence compares different announcements and companies.
Assessment: SupportedSaid
Observation
“that models will continue to get more efficient and less expensive, but that hallucinations and reliability problems will persist.”
Our reading: Models have gotten more efficient and cheaper, but hallucinations and reliability problems have persisted.
No horizon Evidence as of
Limitation: The concrete within-period efficiency comparison is vendor-reported.
Assessment: UnresolvedSaid
Conditional forecast
“Certainly it would slow down progress and raise prices, helping two US companies at the expense of hundreds of thousands of others.”
Our reading: If the US builds a regulatory moat, it would certainly slow down progress and raise prices, helping two US companies at the expense of hundreds of thousands of others.
No deadline stated Evidence as of
Only if: if the US builds a regulatory moat
Assessment: UnresolvedSaid
Causal claim
“that the lack of a technical moat would mean that companies like OpenAI would remain unprofitable.”
Our reading: The lack of a technical moat has kept companies like OpenAI unprofitable, as Marcus predicted.
No horizon Evidence as of
Limitation: Financial evidence is reputable secondary reporting on audited statements, not directly inspected statements.
Assessment: UnresolvedSaid
Observation
“That’s basically what has happened.”
Our reading: Marcus's January 2025 predictions have basically come true: that the US would not achieve a decisive AI victory over China if it continued to focus largely on LLMs, that the two would converge on a tie, that the lack of a technical moat would keep companies like OpenAI unprofitable, that Nvidia would be vulnerable, that the CHIPS act would backfire without doing much to hold China back, that models would keep getting more efficient and less expensive while hallucinations and reliability problems persisted, and that racing endlessly around LLMs would sap resources from more original ideas.
No horizon Evidence as of
Limitation: This assesses the retrospective bundle, not a newly issued conditional forecast.
Assessment: SupportedSaid
Causal claim
“But the Mythos moment got The White House to consider and in fact implement preflight testing to some degree (though not with the transparency or independence or breadth or mandatoriness that I believe to be critical).”
Our reading: The Mythos moment caused the White House to implement AI preflight testing to some degree, though without the transparency, independence, breadth or mandatoriness Marcus considers critical.
No horizon Evidence as of
Limitation: The causal link rests on news reporting. The order's text reportedly does not name Mythos.
Source: Breaking: Demis Hassabis endorses preflight safety testing for AI
Assessment: SupportedSaid
Observation
“In my 2023 Senate testimony, in a 2023 essay here, and in my 2024 book Taming Silicon Valley, I strongly advocated for a system of preflight testing for AI models deployed at large scale, ideally mandatory, transparent and independent, modeled perhaps on the FDA’s analysis of costs and benefits for pharmaceuticals.”
Our reading: In his 2023 Senate testimony, a 2023 essay, and his 2024 book Taming Silicon Valley, Marcus strongly advocated preflight testing for AI models deployed at large scale, ideally mandatory, transparent and independent, perhaps modeled on the FDA's cost-benefit analysis for pharmaceuticals.
No horizon Evidence as of
Limitation: The evidence does not establish that every work separately included every safeguard or the exact FDA cost-benefit formulation.
Source: Breaking: Demis Hassabis endorses preflight safety testing for AI
Assessment: UnresolvedSaid
Causal claim
“Second, because the systems are not reliable, charging premium prices was never really viable in the long term.”
Our reading: Because the systems are not reliable, charging premium prices was never really viable in the long term.
No horizon Evidence as of
Limitation: The research abstract provides limited methodological detail.
Source: China catches up
Assessment: UnresolvedSaid
Causal claim
“Third, the basic approach is easily replicated, leading to the price war dynamics and small or negative margins.”
Our reading: The LLM approach is easily replicated, which causes price wars and small or negative margins.
No horizon Evidence as of
Limitation: Capability evidence concerns one competitor and predates the post.
Source: China catches up
Assessment: UnresolvedSaid
Observation
“As I have said before, over and over, LLMs are becoming a commodity.”
Our reading: LLMs are becoming a commodity.
No horizon Evidence as of
Limitation: The principal June source is a model marketplace with a commercial interest in offering multiple vendors.
Assessment: MixedSaid
Observation
“SpaceX, increasingly pitched as an AI play, has lost steam (up slightly for the day, but down 11.74% for the week), and its bonds are doing badly, too:”
Our reading: SpaceX, increasingly pitched as an AI play, has lost steam, being up slightly for the day but down 11.74% for the week, and its bonds are doing badly too.
No horizon Evidence as of
Limitation: The exact intraday snapshot and chart baseline are missing.
Assessment: SupportedSaid
Observation
“• Chinese models are gaining traction, fast.”
Our reading: Chinese AI models are rapidly gaining market traction.
No horizon Evidence as of
Limitation: The observations do not cover the full June 1–25 target period or establish continuous growth within it.
Assessment: UnresolvedSaid
Observation
“In my view, the world has indeed gone in this direction. LLMs have become commodities, prices wars have been significant, profits for the LLM providers have been elusive.”
Our reading: LLMs have become commodities.
No horizon Evidence as of
Limitation: The meaning of commodities materially affects the assessment.
Source: The Generative AI Fizzle™
Assessment: SupportedSaid
Observation
“In my view, the world has indeed gone in this direction. LLMs have become commodities, prices wars have been significant, profits for the LLM providers have been elusive.”
Our reading: Price wars among LLM providers have been significant.
No horizon Evidence as of
Limitation: The clearest evidence of directly responsive price cutting concerns China.
Source: The Generative AI Fizzle™
Assessment: UnresolvedSaid
Observation
“In my view, the world has indeed gone in this direction. LLMs have become commodities, prices wars have been significant, profits for the LLM providers have been elusive.”
Our reading: Profits for LLM providers have been elusive.
No horizon Evidence as of
Limitation: The unresolved verdict does not mean the identified providers were profitable.
Source: The Generative AI Fizzle™
Assessment: SupportedSaid
Observation
“Small wonder tokenmaxxing is dying;”
Our reading: Tokenmaxxing (maximizing AI token use) is in decline.
No horizon Evidence as of
Limitation: This is qualitative evidence of a corporate trend, not a representative prevalence estimate.
Source: Accenture: Then and now, and how it may signify things to come
Assessment: UnresolvedSaid
Observation
“But on the whole, Generative AI has not been delivering great gains. We have seen that now in study after study, MIT, McKinsey, Bain, and many more.”
Our reading: Overall, generative AI has not been delivering large gains, as many studies (MIT, McKinsey, Bain) show.
No horizon Evidence as of
Limitation: The NANDA document is preliminary and was retrieved from a third-party mirror.
Source: Accenture: Then and now, and how it may signify things to come
Assessment: UnresolvedSaid
Conditional forecast
“I really do think that Claude Code (which remember, is a special-purpose neurosymbolic system rather than a generic chatbot) will increase productivity for coders (if technical debt doesn’t swallow back those gains).”
Our reading: Claude Code will increase coders' productivity, unless technical debt cancels out the gains.
No deadline stated Evidence as of
Only if: if technical debt doesn't swallow back the gains
Source: Accenture: Then and now, and how it may signify things to come
Assessment: SupportedSaid
Observation
“It took longer than I expected, but well, here we are:”
Our reading: The politics and inadequacy of generative AI guardrails, which Marcus predicted in 2024 would become central, has now become a central issue, though later than he expected.
No horizon Evidence as of
Limitation: Prominence is qualitative.
Assessment: SupportedSaid
Observation
“Importantly, this is not an Anthropic problem, it’s a Generative AI problem.”
Our reading: The inability to block all jailbreaks is a problem with generative AI in general, not one specific to Anthropic.
No horizon Evidence as of
Limitation: The assessment was published in May, but the accessible source does not give individual experiment dates.
Assessment: UnresolvedSaid
Forecast
“You can’t multiply your losses every year by 8x (or even close) for very long.”
Our reading: OpenAI cannot keep multiplying its losses by around 8x a year for very long.
No deadline stated Evidence as of
Limitation: The first calendar year after the baseline had not ended at the evaluation date.
Source: OpenAI’s lead is dwindling fast
Assessment: SupportedSaid
Observation
“The White House has acknowledged that some models might be risky; now we are just arguing about the details.”
Our reading: The White House has acknowledged some AI models might be risky, so the debate is now only about details of regulation.
No horizon Evidence as of
Limitation: Sacks’s explanation establishes the administration’s stated position, not the model’s actual danger.
Source: What Washington must do
Assessment: SupportedSaid
Observation
“The ridiculous but fashionable idea of zero-regulation around AI, made popular by folks like Marc Andreessen, is (rightly) dead in the water.”
Our reading: The idea of zero regulation of AI is now, rightly, dead.
No horizon Evidence as of
Limitation: The result depends on the surrounding sentences narrowing the rhetoric to the White House’s acceptance of some model restrictions.
Source: What Washington must do
Assessment: SupportedSaid
Observation
“The Commerce department just ordered an export control directive (quoting Anthropic) “to suspend all access to Fable 5 and Mythos 5” by foreign nationals — even their own employees.”
Our reading: The US Commerce Department issued an export control directive requiring Anthropic to suspend all access by foreign nationals, including its own employees, to Fable 5 and Mythos 5.
No horizon Evidence as of
Limitation: The directive itself was not found in a public government record. The evidence is Anthropic's account and Fortune's reporting.
Source: Breaking news: US Commerce Department effectively shuts down Anthropic’s latest models
Assessment: UnresolvedSaid
Observation
“And, of course, all of these risks were foreseeable.”
Our reading: The harms alleged in Florida's lawsuit against OpenAI were all foreseeable.
No horizon Evidence as of
Limitation: Foreseeability does not establish that the alleged harms occurred or that OpenAI legally caused them.
Assessment: UnresolvedSaid
Causal claim
“To be sure, Anthropic brought this on themselves to a considerable degree, by (a) overselling doom and their own wares and (b) by literally and repeatedly calling for export controls themselves.”
Our reading: Anthropic brought the export control order on itself to a considerable degree, by overselling doom and its own products and by literally and repeatedly calling for export controls itself.
No horizon Evidence as of
Limitation: The packet contains no sufficiently complete decision record establishing the influence of Anthropic's advocacy.
Assessment: SupportedSaid
Causal claim
“Now let’s consider last night’s Commerce Department cut-off-your-nose-to-spite-your-face export control order, which has basically led to the (temporary?) shutdown of Fable and Mythos, perhaps sparking a crisis in the US AI industry.”
Our reading: The Commerce Department's June 2026 export control order has basically led to the possibly temporary shutdown of Anthropic's Fable and Mythos.
No horizon Evidence as of
Limitation: The evidence establishes suspended customer access, not cessation of all internal model activity.
Assessment: SupportedSaid
Observation
“His company did however recently support state law in Illinois that “would shield AI labs from liability in cases where AI models are used to cause serious societal harms, such as death or serious injury of 100 or more people or at least $1 billion in property damage”– pretty much the opposite of what he said to Senator Durbin in his deceitful, lie-filled testimony.”
Our reading: OpenAI recently supported an Illinois state law that would shield AI labs from liability in cases where AI models are used to cause serious societal harms, which Marcus calls pretty much the opposite of what Altman told Senator Durbin in his 2023 testimony.
No horizon Evidence as of
Limitation: This outcome applies to the legislative-support proposition, not the allegation of deceit.
Source: Maybe Section 230 doesn’t shield AI companies from liability, after all
Assessment: UnresolvedSaid
Observation
“Because essentially all their software has a tendency to fabricate fairly regularly and even defame people and to sometimes give bad medical information, and so on.”
Our reading: Essentially all LLM providers' software fabricates fairly regularly, sometimes defaming people and giving bad medical information.
No horizon Evidence as of
Limitation: Difficult and adversarial benchmarks cannot be treated as random samples of everyday use.
Source: Maybe Section 230 doesn’t shield AI companies from liability, after all
Assessment: UnresolvedSaid
Conditional forecast
“If American courts were to rule similarly, or if Congress made that the law of the land, all the LLM providers would be in deep, deep trouble.”
Our reading: If US courts or Congress held companies liable for their chatbots' output, all LLM providers would be in deep trouble.
No deadline stated Evidence as of
Only if: if US courts rule that chatbot output is the company's own speech, or Congress legislates so
Source: Maybe Section 230 doesn’t shield AI companies from liability, after all
Assessment: UnresolvedSaid
Forecast
“And so at some point, I think OpenAI is gonna have a problem where they can’t really meet their obligations.”
Our reading: At some point OpenAI will be unable to meet its financial obligations.
No deadline stated Evidence as of
Limitation: The cash figure comes from The Decoder's summary of The Information's report on documents shared with shareholders. It is not an audited filing.
Assessment: SupportedSaid
Observation
“And OpenAI still has yet to turn a profit, and it’s still not clear they ever will.”
Our reading: OpenAI has not yet turned a profit and it is unclear whether it ever will.
No horizon Evidence as of
Limitation: The figures come from press reports of shareholder documents, not audited public filings. Reuters said it could not immediately verify the report.
Assessment: SupportedSaid
Observation
“Anthropic has taken a lot of the business market and OpenAI has lost a lot of it.”
Our reading: Anthropic has taken a large share of the business market while OpenAI has lost much of it.
No horizon Evidence as of
Limitation: Enterprise LLM API share does not measure the entire business AI market [e1].
Assessment: SupportedSaid
Causal claim
“And the reason that those tend to be more expensive is that they tend (with the help of some good old fashioned AI that nobody wants to talk about) to run multiple large language models in the background many times until they get the answers that by some criteria seem to correct.”
Our reading: AI agents tend to be more expensive because they tend to run multiple large language models in the background many times until they get answers that by some criteria seem correct.
No horizon Evidence as of
Limitation: The main evidence is one provider's description of its own system.
Assessment: UnresolvedSaid
Conditional forecast
“If they can’t meet their obligations, I expect Nvidia and Oracle and many others would take a big hit as collateral damage (Domino 5).”
Our reading: If OpenAI can't meet its obligations, Nvidia, Oracle and many others would take a big hit as collateral damage.
No deadline stated Evidence as of
Only if: if OpenAI can't meet its obligations
Assessment: UnresolvedSaid
Observation
“Two years later, competitors have caught up. OpenAI has had to cut prices.”
Our reading: Over the past two years competitors have caught up to OpenAI and it has had to cut prices.
No horizon Evidence as of
Limitation: Arena measures human preferences and does not establish universal capability parity.
Assessment: SupportedSaid
Observation
“Opensource models (though perhaps not Meta’s) are only a few months behind.”
Our reading: Open-source models, perhaps excluding Meta's, are only a few months behind frontier models.
No horizon Evidence as of
Limitation: One measurement organization and one index.
Assessment: UnresolvedSaid
Observation
“Open AI, I think, has a problem right now. They are burning money the fastest.”
Our reading: OpenAI has a problem and is burning money faster than its AI competitors.
No horizon Evidence as of
Limitation: Neither item supplies a comparable January–March 2026 cash-burn ranking for OpenAI, Anthropic, Google and Amazon [e1, e2].
Topics
Counts across all 3756 claims by Gary Marcus, checked or not. A claim can have more than one topic.
- Capabilities1098
- Safety829
- Reliability814
- Regulation policy651
- Business economics615
- Other345
- Scaling327
- AGI timelines320
- Financing249
- Agents111
- Copyright96
- Labour71
Sources covered
- Marcus on AI, to . Every post in the window, re-captured from the publication's archive listing. The legacy dataset (474 posts, snapshot 2026-03-02) is used only as a cross-check list.
- Marcus on AI, to . Posts published after the legacy snapshot, discovered from the publication's archive listing.
Recent changes
- : Assessment published, There are few regulations on frontier AI systems, and this may not change soon.
- : Assessment published, Safely building AGI should be treated as a matter of global security, not national security.
- : Held
- : Assessment published, The timing of Brundage's exit may be connected to OpenAI's attempt to restructure as a for-profit PBC.
- : Assessment published, Brundage's departure is another sign that the original OpenAI no longer exists.