Changes, page 20 of 37

Every substantive change to the register, newest first. Speaker events record what the people covered said or changed. Site events record what this site published, corrected, flagged, or withdrew. The two never share a label.

Substantive changes (7383)

Page 20 of 37: changes 3801 to 4000 of 7383, newest first. Dates are when the change was recorded.

2026-10-29
Assessment publishedBy this siteGary Marcus
Inferring sentience in AI systems directly from their outputs is naive.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-richard-dawkins-and-the-claude-delusion-0c9460-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Remote driving is unlikely to work well enough as a long-term solution on public roads because data communications are not stable enough.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-rethinking-driverless-cars-fdd09a-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Waymo is probably better, but how much better is anybody's guess.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-rethinking-driverless-cars-e4dd88-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
No AGI fantasy so far has held up against the real world.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-rethinking-driverless-cars-e06ccb-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
It is quite possible no company will profit from remote-assisted driving anytime soon.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-rethinking-driverless-cars-d62cc3-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Remotely-assisted driving is semi-autonomy, not true autonomy; it might save money, though the author doubts it; and it has literally nothing to do with AGI.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-rethinking-driverless-cars-d0c9c3-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
LLMs in safety-critical domains like medicine may disappoint, requiring humans in the loop for a very long time.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-rethinking-driverless-cars-96cc26-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-rethinking-driverless-cars-6a5895
The two reviews differ: anthropic forecast_outcome/open_horizon/unresolved, openai forecast_outcome/open_horizon/not_adjudicable.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-rethinking-driverless-cars-683b11
The final-text check failed: rationale_supported_by_evidence, citations_correct; overclaims: Some requests resolved automatically, and Cruise said human advisors generally supplied wayfinding guidance rather than driving remotely (e1). | The reported assistance figure of “2–4% of the time on average” has no defined denominator..
2026-10-29
Assessment publishedBy this siteGary Marcus
Driver-assist systems like Tesla's may persist while occasionally causing fatal accidents due to unreliability.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-rethinking-driverless-cars-6798de-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-rethinking-driverless-cars-50162b
The anthropic review found the criteria unfair to the speaker's words.
2026-10-29
Assessment publishedBy this siteGary Marcus
Within a few years the self-driving car industry will likely be seen as an epic failure, defeated by outliers and AI's inadequate reasoning.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-rethinking-driverless-cars-38dff4-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
If true self-driving is this hard despite the data collected and $100B+ invested, any dreams of AGI happening soon are likely wildly unrealistic.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-rethinking-driverless-cars-1e98ef-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-reports-of-the-birth-of-agi-are-greatly-f9550e
The final-text check failed: proposition_faithful, rationale_supported_by_evidence; overclaims: Rationale: "the April 2023 DINOv2 paper ... reports leading general-purpose visual-feature performance [e2]" — the quoted e2 text only shows a ViT-g backbone used for distillation, not benchmark leadership. | e2 explanation: "reports leading general-purpose feature performance across image and pixel benchmarks" — not shown in the quoted evidence..
2026-10-29
Assessment publishedBy this siteGary Marcus
Current next-token prediction models have remarkable abilities and weaknesses whose scope is not well understood.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-reports-of-the-birth-of-agi-are-greatly-c4dfd5-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-reports-of-the-birth-of-agi-are-greatly-b46388
e1 is dated 2023, which is not inside the period described (2023-01-01 to 2023-10-18); it can only be context; e2 is dated 2023, which is not inside the period described (2023-01-01 to 2023-10-18); it can only be context; e3 is dated 2023, which is not inside the period described (2023-01-01 to 2023-10-18); it can only be context
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-reports-of-the-birth-of-agi-are-greatly-98f0a0
The anthropic screen finds the claim checkable, so it must be graded.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-reports-of-the-birth-of-agi-are-greatly-965273
The final-text check failed: proposition_faithful; overclaims: e2 explanation: "the factuality comparison also favored GPT-4" — the quoted abstract only says LLM-generated summaries; GPT-4 is not named in the evidence shown..
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-reports-of-the-birth-of-agi-are-greatly-919692
e1 is dated 2023, which is not inside the period described (2023-01-01 to 2023-10-18); it can only be context
2026-10-29
Assessment publishedBy this siteGary Marcus
There is zero justification for claiming that the current technology has achieved general intelligence.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-reports-of-the-birth-of-agi-are-greatly-8eb1f4-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
No single AI technique is likely ever to be a silver bullet.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-reports-of-the-birth-of-agi-are-greatly-6a50dc-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
LLMs trained on next-token prediction doing a pretty good job at a variety of natural language tasks, and very recently a considerably less good job at tasks combining text with images, is certainly interesting but does not justify the claim that they constitute general intelligence.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-reports-of-the-birth-of-agi-are-greatly-5d18b5-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
The next generation of systems will, without doubt, have greater abilities and more extensive applications and will somewhat mitigate the weaknesses of current systems.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-reports-of-the-birth-of-agi-are-greatly-5c2e35-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
The problem with the Agüera y Arcas and Norvig paper is not just its hostility toward critics or the way it lowers the bar for AGI, but that its key arguments don't go through.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-reports-of-the-birth-of-agi-are-greatly-56153c-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
AGI is more likely than not still at least a few paradigm shifts away.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-reports-of-the-birth-of-agi-are-greatly-426395-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Transformers and generative AI are getting a lot of attention and have considerable utility, but they are nothing like the all-purpose tools some people seem to imagine them to be.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-reports-of-the-birth-of-agi-are-greatly-355e77-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
It is certainly plausible that some prerequisites of AGI may have been achieved, but it is considerable overreach to presume that all of the most important ones have already been invented.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-reports-of-the-birth-of-agi-are-greatly-2939f8-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Current systems are reliable enough to be useful in some applications, but far fewer than often claimed.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-reports-of-the-birth-of-agi-are-greatly-2108c3-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-reports-of-the-birth-of-agi-are-greatly-19efa3
The openai review does not accept e1 as supports.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-reports-of-the-birth-of-agi-are-greatly-0d5048
a observation claim cannot be assessed on the causal_assessment axis
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-reports-of-the-birth-of-agi-are-greatly-09e2b6
e1 is dated 2023, which is not inside the period described (2023-01-01 to 2023-10-18); it can only be context
2026-10-29
Assessment publishedBy this siteGary Marcus
The first author's own nine-year-old child can do five-digit multiplication better than an LLM.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-reports-of-the-birth-of-agi-are-greatly-02608f-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
If LLMs become commonly used for difficult questions before their invalid-answer problem is fixed, society will be in serious trouble.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-reports-of-llms-mastering-math-have-fc77f5-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Getting AIs to answer 'I don't know' is one of the most important unsolved challenges facing the field.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-reports-of-llms-mastering-math-have-fbfba1-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
The refusal of these kinds of AI (LLMs) to admit ignorance or incapacity, and their obstinate preference for generating incorrect but plausible-looking answers instead, are among their most dangerous characteristics.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-reports-of-llms-mastering-math-have-bba95d-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-reports-of-llms-mastering-math-have-a0ac22
The openai review found the criteria unfair to the speaker's words.
2026-10-29
Assessment publishedBy this siteGary Marcus
The really important challenge is not getting AIs to solve more USAMO problems but getting them to say 'I give up' when they can't solve them.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-reports-of-llms-mastering-math-have-76a9bc-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Additional user prompting may not substantially improve reasoning LLMs' performance on such problems.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-reports-of-llms-mastering-math-have-5e7bac-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
However you slice it, the AIs' need for special cajoling before they will think carefully about questions presented to them is a serious weakness, and the demand for special prompting tilts the scale.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-reports-of-llms-mastering-math-have-0fb2d7-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
AI needs systems that derive the meaning of wholes from their parts according to syntax, which no one yet knows how to build.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-race-statistics-and-the-persistent-dba2b0-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-race-statistics-and-the-persistent-c38cb0
The final-text check failed: citations_correct.
2026-10-29
Assessment publishedBy this siteGary Marcus
Multimodal ChatGPT's trouble telling time from watch images comes from the same canonical-versus-noncanonical statistics problem seen in DALL-E.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-race-statistics-and-the-persistent-c2f239-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Blind fealty to the statistics of arbitrary data is not AGI; it is a hack, and not one we should want.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-race-statistics-and-the-persistent-b1a991-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Multimodal ChatGPT has low real comprehension, and the few correctly labeled clock examples are probably drowned out by billions of watch ads showing the canonical time.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-race-statistics-and-the-persistent-a6f9da-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-race-statistics-and-the-persistent-765f6b
The anthropic screen reads the claim differently.
2026-10-29
Assessment publishedBy this siteGary Marcus
DALL-E 3 cannot distinguish the world from its training data statistics and has no cognitive concepts of doctors, patients, race, or similar.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-race-statistics-and-the-persistent-5658e7-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
About 18 months after DALL-E 2, the images are vastly better, presumably largely because of a larger training set, but the fundamental approach of mapping words to images rather than building intermediate world models is the same, and so the core compositionality issue remains.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-race-statistics-and-the-persistent-4e82f9-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-race-statistics-and-the-persistent-2ca679
The final-text check failed: citations_correct.
2026-10-29
Assessment publishedBy this siteGary Marcus
AI research should focus on building systems that can represent and reason about values instead of just perpetuating past data.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-race-statistics-and-the-persistent-071f56-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Trump's April 2025 tariff policy was quite possibly written partly by LLMs.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-questions-about-president-trump-from-b3e946-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
If we don't stand together, big tech will walk all over all of us.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-psa-microsoft-may-be-training-on-d00242-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
None of us stand any chance against big tech unless we stand together.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-psa-microsoft-may-be-training-on-b9561e-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Every big tech company is desperate for AI training data.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-psa-microsoft-may-be-training-on-930bb2-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Office users who have anything at all private or sensitive in their Office documents might want to switch off the 'optional connected experiences' setting, as Marcus did.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-psa-microsoft-may-be-training-on-70ee4a-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Too often, media companies platform predictions from people who stand to gain immensely from their narratives, with too little skepticism.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-promises-are-cheap-b7319e-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
If the AI boom collapses, careless media repetition of hype may turn out to have been an important contributing factor.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-promises-are-cheap-a7dcaf-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
According to the December 2025 Remote Labor Index, AI could do only 2.5% of human online tasks.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-promises-are-cheap-a44926-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-promises-are-cheap-77b3e7
The openai review reads the claim differently.
2026-10-29
Assessment publishedBy this siteGary Marcus
The public has not been well served by media companies' practice of platforming predictions from people who stand to gain from their narratives with too little skepticism.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-promises-are-cheap-5ad2cc-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-promises-are-cheap-52a602
The final-text check failed: rationale_supported_by_evidence; overclaims: That points to large benchmark gains, but the figure is reported by OpenAI itself. | These point toward large gains, but they are company-reported and do not settle a subjective comparison..
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-promises-are-cheap-4fc9e1
The openai review found the criteria unfair to the speaker's words.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-promises-are-cheap-4be39e
The openai review does not accept e2 as supports.
2026-10-29
Assessment publishedBy this siteGary Marcus
Elon Musk's promise of a million Tesla robotaxis, pushed back to 2026, will not be fulfilled in 2026.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-promises-are-cheap-46185b-a2.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-promises-are-cheap-45e99b
The two reviews differ: anthropic historical_observation/not_applicable/unresolved, openai historical_observation/not_applicable/mixed.
2026-10-29
Assessment publishedBy this siteGary Marcus
All the big tech CEOs have come to realize that making promises costs them nothing.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-promises-are-cheap-37dfcc-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Tesla's valuation of nearly 400 times earnings is due to Musk's hype rather than profits.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-promises-are-cheap-32b435-a2.
2026-10-29
Assessment publishedBy this siteGary Marcus
Media outlets like the FT should give context on conflicts of interest and track records when reporting AI executives' predictions, and seek independent opinions.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-promises-are-cheap-208ba6-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Although GenAI is marketed as near-AGI, its reliability problems persist and limit how well it can be used in practice.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-poor-roi-for-genai-ed58cd-a2.
2026-10-29
Assessment publishedBy this siteGary Marcus
NVidia's sharp stock decline may be partly because it is becoming clear that GenAI is not delivering the expected return on investment.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-poor-roi-for-genai-c6e69c-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
With GPT-5-level models still elusive, the poor ROI picture for GenAI may not change enormously in the near term.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-poor-roi-for-genai-c02ce8-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
As of April 2025, GPT-5-level models have not yet arrived.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-poor-roi-for-genai-c02ce8-2-a2.
2026-10-29
Assessment publishedBy this siteGary Marcus
NVidia's stock is down 23.4% in 2025 so far, significantly more than the S&P's 11.5% decline.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-poor-roi-for-genai-1b5172-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-poetry-for-our-agi-times-30115c
The openai screen doubts the attribution.
2026-10-29
Assessment publishedBy this siteGary Marcus
Marcus thinks smart people are spending too much time experimenting with GPT-3 and DALL-E when there are deeper problems in AI to solve.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-poetry-for-our-agi-times-21f29a-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
The apparent confidence of LLMs is an illusion that comes from how they are trained.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-please-dont-trust-your-chatbot-for-f290cb-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
A large fraction of the population has begun using chatbots for medical advice.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-please-dont-trust-your-chatbot-for-bf86d4-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
For now and until proven otherwise, people should not take LLMs' apparent confidence as reason to trust them with their lives.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-please-dont-trust-your-chatbot-for-9898e3-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Hallucinations, mistakes and overconfidence are entirely typical of LLMs and entirely problematic in medicine.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-please-dont-trust-your-chatbot-for-7ffa80-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
An expert doctor might use an LLM to better effect than lay users by asking the right questions.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-please-dont-trust-your-chatbot-for-73e6b3-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-please-dont-trust-your-chatbot-for-3dcde7
The anthropic screen doubts the attribution.
2026-10-29
Assessment publishedBy this siteGary Marcus
Large language models are still frequently wrong yet confident, producing authoritative bullshit that shouldn't be trusted, and this applies strongly in medicine.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-please-dont-trust-your-chatbot-for-3db366-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-please-dont-trust-your-chatbot-for-352372
The openai review found the criteria unfair to the speaker's words.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-please-dont-trust-your-chatbot-for-2e779a
e2 is dated 2026-04, which is not inside the period described (2026-01-01 to 2026-04-21); it can only be context
2026-10-29
Assessment publishedBy this siteGary Marcus
When the AI bubble ends, the outcome will be bad.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-peak-bubble-f45df8-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
The total value of the tech market as a whole far exceeds the value its companies are likely ever to deliver.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-peak-bubble-c8d2b1-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
The generative AI market is a bubble and has reached peak bubble, though Marcus does not know when it will end.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-peak-bubble-84b55a-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-peak-bubble-7c952e
The anthropic screen finds the claim checkable, so it must be graded.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-peak-bubble-5b4bb1
The anthropic screen finds the claim checkable, so it must be graded.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-peak-bubble-29d8f0
The anthropic screen finds the claim checkable, so it must be graded.
2026-10-29
Assessment publishedBy this siteGary Marcus
The AI market is well past peak bubble and has reached a peak game of musical chairs.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-peak-bubble-1120ad-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-peak-bubble-052511
The anthropic screen reads the claim differently.
2026-10-29
Assessment publishedBy this siteGary Marcus
Oracle's new market cap, near a trillion dollars and up nearly 50% this week, largely driven by an apparently non-binding deal with a party that lacks the money to pay for the services, seems more bonkers than most AI valuations.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-peak-bubble-0420c1-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-peak-absurdity-part-ii-f5a78c
The anthropic screen doubts the attribution.
2026-10-29
Assessment publishedBy this siteGary Marcus
As of the post date, even after a very strong week, Oracle stock stands at just $170, compared with about $308 on September 11.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-peak-absurdity-part-ii-ab1355-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Marcus may have spoken too soon in saying that the Oracle stock spike would be seen as the peak of absurdity, because a new example has now appeared.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-peak-absurdity-part-ii-a5ff19-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-peak-absurdity-part-ii-9a2c00
The final-text check failed: citations_correct.
2026-10-29
Assessment publishedBy this siteGary Marcus
GPT-5 was not tantamount to AGI, although OpenAI thought or claimed that it would be.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-peak-absurdity-part-ii-60174f-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-peak-absurdity-part-ii-5b4bb1
The final-text check failed: proposition_faithful, citations_correct; overclaims: OpenAI does not have the $300 billion needed for its Oracle deal..
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-peak-absurdity-part-ii-29d8f0
The final-text check failed: citations_correct.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-peak-absurdity-part-ii-052511
The anthropic screen reads the claim differently.
2026-10-29
Assessment publishedBy this siteGary Marcus
Because LLMs cluster similar things in a high-dimensional space, they become synonym and pastiche machines that regurgitate with slight paraphrase, adding little and understanding less.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-partial-regurgitation-and-how-llms-97b315-a2.
2026-10-29
Assessment publishedBy this siteGary Marcus
Getting to real AI will require a different approach from the partial-regurgitation approach of current LLMs.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-partial-regurgitation-and-how-llms-6c3a0b-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Storing weights does not mean neural nets know what they are talking about.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-partial-regurgitation-and-how-llms-689615-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Partial regurgitation, no matter how fluent, does not and will not ever constitute genuine comprehension.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-partial-regurgitation-and-how-llms-12c8c0-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Marcus leans toward the view that the Paris AI Summit draft statement should not be signed.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-paris-ai-summit-train-wreck-cd3557-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-paris-ai-summit-train-wreck-5bb924
The anthropic review does not accept e2 as supports.
2026-10-29
Assessment publishedBy this siteGary Marcus
Considering all the big names assembled, the Paris AI Summit feels like a wasted opportunity.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-paris-ai-summit-train-wreck-362b56-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
If the draft Paris AI Summit statement were signed, it would accomplish little and would let too many people off the hook, allowing them to say they had addressed AI risks without actually doing all that much.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-paris-ai-summit-train-wreck-3520ac-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Amid the complaints about the Paris AI Summit's leaked draft statement, regulatory capture seemed huge.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-paris-ai-summit-train-wreck-2dfa61-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
If your friend has OpenClaw installed, you should not use their machine, because any password you type there might be vulnerable.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openclaw-aka-moltbot-is-everywhere-e92894-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openclaw-aka-moltbot-is-everywhere-bf4799
The anthropic review does not accept e1 as supports.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openclaw-aka-moltbot-is-everywhere-b60f63
hedged language is not scored, even on a plain reading
2026-10-29
Assessment publishedBy this siteGary Marcus
People who care about device security or data privacy should not use OpenClaw.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openclaw-aka-moltbot-is-everywhere-afccd2-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
OpenClaw is basically a cascade of LLMs.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openclaw-aka-moltbot-is-everywhere-5b24d2-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openclaw-aka-moltbot-is-everywhere-2199be
The anthropic review does not accept e2 as supports.
2026-10-29
Assessment publishedBy this siteGary Marcus
Unlike sandboxed iPhone apps, OpenClaw left unfettered is positioned to cause serious harm.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openclaw-aka-moltbot-is-everywhere-208362-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-waterloo-fc7a90
The two reviews differ: anthropic historical_observation/not_applicable/supported, openai historical_observation/not_applicable/unresolved.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-waterloo-ea67a1
The final-text check failed: proposition_faithful.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-waterloo-d3e7ef
The two reviews differ: anthropic historical_observation/not_applicable/supported, openai historical_observation/not_applicable/unresolved.
2026-10-29
Assessment publishedBy this siteGary Marcus
LLMs will not be the approach that leads the way to AGI.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-waterloo-b64a45-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
LLMs do markedly better on coding benchmarks than in real-world coding practice.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-waterloo-b64a45-2-a2.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-waterloo-aeede7
the horizon on 2026-10-28 is not_due, not open_horizon; hedged language is not scored, even on a plain reading
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-waterloo-9c03a7
The anthropic review found the criteria unfair to the speaker's words.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-waterloo-8dacbd
The openai screen reads the claim differently.
2026-10-29
Assessment publishedBy this siteGary Marcus
AI cannot be made better if you drown out the critics and keep pouring good money after bad.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-waterloo-749cae-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Almost no one was satisfied with GPT-5, and some customers wanted the old models back.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-waterloo-71964f-a2.
2026-10-29
Assessment publishedBy this siteGary Marcus
Hype and bragging alone cannot produce AGI.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-waterloo-46892d-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-waterloo-3908e7
The openai review found the criteria unfair to the speaker's words.
2026-10-29
Assessment publishedBy this siteGary Marcus
As the limits of scaling become recognized, scientists will pursue new approaches, one of which may produce trustworthy AI.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-waterloo-372737-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-waterloo-13ac30
The anthropic screen reads the claim differently.
2026-10-29
Assessment publishedBy this siteGary Marcus
Two years ago many people treated Altman like an oracle, and now, in the eyes of many, he looks more like a snake oil salesman.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-waterloo-0ea2ca-a2.
2026-10-29
Assessment publishedBy this siteGary Marcus
GPT-5 is starting to look like Altman's decisive defeat.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-waterloo-02dd34-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-slow-motion-train-wreck-ca18c7
The anthropic screen reads the claim differently.
2026-10-29
Assessment publishedBy this siteGary Marcus
The magazine cover of OpenAI's leaders from last September may soon become as iconic as the most famous magazine covers of our time.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-slow-motion-train-wreck-b154d8-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Investors should not put more money into OpenAI at higher valuations and should instead scrutinize what is going on at the company.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-slow-motion-train-wreck-500831-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Elon Musk's lawsuit against OpenAI may fail or even be dismissed.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-own-goal-c0eb62-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
The revealed facts suggest OpenAI never really intended to follow its charter.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-own-goal-bcbe0c-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Marcus applies the proverb 'leopards rarely change their spots' to OpenAI, expressing distrust of its character.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-own-goal-844bea-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-own-goal-644a39
The two reviews differ: anthropic historical_observation/not_applicable/mixed, openai historical_observation/not_applicable/unresolved.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-own-goal-4cdfa1
The anthropic screen reads the claim differently.
2026-10-29
Assessment publishedBy this siteGary Marcus
OpenAI knew it was misleading people into believing it meant open-sourced AI and failed to take an opportunity, at a May 2016 meeting, to correct that belief.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-own-goal-463360-a2.
2026-10-29
Assessment publishedBy this siteGary Marcus
Almost nothing is known about the precise internal mechanisms of GPT-4 or what it has been trained on.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-own-goal-3c5406-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-own-goal-28e78a
The anthropic screen reads the claim differently.
2026-10-29
Assessment publishedBy this siteGary Marcus
OpenAI should not turn a blind eye to covert racism, apparent copyright infringement, and violent or degrading outputs.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-own-goal-1329d2-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-own-goal-11234a
The anthropic review does not accept e2 as contradicts.
2026-10-29
Assessment publishedBy this siteGary Marcus
The apparent copyright infringements that Marcus and Reid Southen discussed in early January 2024 were continuing apace two months later, based on a quick experiment Marcus ran on the morning of the post.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-own-goal-03baf3-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Tyler Cowen's claim that OpenAI's o3 is AGI was refuted within a day of its release.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-o3-and-tyler-cowens-misguided-f34825-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-o3-and-tyler-cowens-misguided-31bb96
The openai screen reads the claim differently.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-o3-and-tyler-cowens-misguided-0589ab
The anthropic screen reads the claim differently.
2026-10-29
Assessment publishedBy this siteGary Marcus
While publicly asking the Senate to regulate it, OpenAI worked behind the scenes to water down regulations.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-lies-and-half-truths-cfd3b2-a2.
2026-10-29
Assessment publishedBy this siteGary Marcus
In his May 2023 Senate testimony, Sam Altman left out his indirect OpenAI stake through Y Combinator and his ownership of the OpenAI venture fund.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-lies-and-half-truths-c81e88-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
As OpenAI gears up to upload users' private email, calendar, and documents, possibly for eventual resale to advertisers, its lack of full candor is something worth considering.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-lies-and-half-truths-c09020-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-lies-and-half-truths-97fd17
The openai review reads the claim differently.
2026-10-29
Assessment publishedBy this siteGary Marcus
Given her role as CTO, it seems all but inconceivable to Marcus that Mira Murati had no idea what specific data Sora was trained on.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-lies-and-half-truths-8a8620-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Phrases in OpenAI's official filings such as 'Charitable' and 'For the public benefit' don't seem to have meant much at all.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-lies-and-half-truths-89ae41-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-lies-and-half-truths-519ac2
The openai screen reads the claim differently.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-lead-is-dwindling-fast-d52790
The anthropic screen reads the claim differently.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-lead-is-dwindling-fast-9ed742
The openai review does not accept e3 as contradicts.
2026-10-29
Assessment publishedBy this siteGary Marcus
The US government might succeed in knee-capping Anthropic, which would protect Josh Kushner's investment in OpenAI.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-lead-is-dwindling-fast-7f4134-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-lead-is-dwindling-fast-786cce
The openai review found the criteria unfair to the speaker's words.
2026-10-29
Assessment publishedBy this siteGary Marcus
OpenAI cannot keep multiplying its losses by around 8x a year for very long.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-lead-is-dwindling-fast-751563-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-lead-is-dwindling-fast-2f7af6
The openai review doubts the attribution.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-lead-is-dwindling-fast-2deb22
The anthropic review does not accept e3 as supports.
2026-10-29
Assessment publishedBy this siteGary Marcus
Copyright battles may cut massively into OpenAI's profits and may be unavoidable.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-got-99-problems-and-twitch-f8bbb1-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-got-99-problems-and-twitch-ec185a
The anthropic review found the criteria unfair to the speaker's words.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-got-99-problems-and-twitch-e59077
e7 is dated 2024-01, which is not inside the period described (2023-01-01 to 2024-01-27); it can only be context
2026-10-29
Assessment publishedBy this siteGary Marcus
OpenAI's licensing costs for copyrighted material may quickly mount.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-got-99-problems-and-twitch-e15e51-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-got-99-problems-and-twitch-de64ed
The final-text check failed: citations_correct; overclaims: OpenAI's answer explicitly includes licensed information among its training sources and acknowledges its AP partnership and further negotiations, contradicting the claimed omission of licensing..
2026-10-29
Assessment publishedBy this siteGary Marcus
Internal tensions at OpenAI are likely to persist in 2024.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-got-99-problems-and-twitch-d46504-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
The rising costs of bigger models trained on more data may not be accompanied by big enough gains in performance, at some point leading the big companies to step off the gas.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-got-99-problems-and-twitch-cf443d-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
The perception of OpenAI's unlimited potential could soon change.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-got-99-problems-and-twitch-c6beae-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
If LLMs plateau in the near term, enthusiasm may dissipate and investors may temper their investments.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-got-99-problems-and-twitch-c58975-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Marcus guesses that OpenAI, backed by and deeply tied to Microsoft, will survive, but that the bloom will fade from the rose.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-got-99-problems-and-twitch-afbae6-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Public Citizen's call could strain OpenAI and, if California investigates, lead to uncomfortable revelations.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-got-99-problems-and-twitch-a454eb-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Many generative AI companies are likely to face intense legal and public pressure over copyright in 2024.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-got-99-problems-and-twitch-a3cd7f-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-got-99-problems-and-twitch-9af2fc
The anthropic screen finds the claim checkable, so it must be graded.
2026-10-29
Assessment publishedBy this siteGary Marcus
OpenAI will likely face a rapid flurry of lawsuits in 2024.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-got-99-problems-and-twitch-9472d9-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Most high-quality data sources for LLM training have likely already been tapped.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-got-99-problems-and-twitch-911b4d-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
EU and other regulation may force OpenAI to confront unreliability, privacy, misinformation, and cybercrime problems.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-got-99-problems-and-twitch-8a5fee-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
OpenAI's request for a copyright exemption is likely in violation of common law and international IP conventions like Berne.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-got-99-problems-and-twitch-8710f3-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
2024 looks likely to be a much more challenging year for OpenAI and Sam Altman.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-got-99-problems-and-twitch-7f7229-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Large language models such as ChatGPT may quickly become commodities, price wars can be expected, and profits may continue to be elusive or modest at best.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-got-99-problems-and-twitch-62f02d-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
External investors putting money into OpenAI, most recently at a rumored $86 billion valuation, may find themselves disappointed and unclear on how profits to justify those valuations could ever be made.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-got-99-problems-and-twitch-53f9bf-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
As of January 2024, OpenAI has neither profits nor a competitive moat.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-got-99-problems-and-twitch-535fab-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-got-99-problems-and-twitch-3afee6
hedged language is not scored, even on a plain reading
2026-10-29
Assessment publishedBy this siteGary Marcus
It is unclear that ChatGPT adoption will be enough to justify OpenAI's very high valuation.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-got-99-problems-and-twitch-1942cd-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
OpenAI may spend a huge part of its energy in 2024 defending lawsuits.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-got-99-problems-and-twitch-166cb4-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
LLM technology may reach a plateau, perhaps in 2024 or 2025.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-got-99-problems-and-twitch-0ae95f-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Valuations of OpenAI and some peers may drop sharply.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-got-99-problems-and-twitch-090ea2-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-future-foretold-dbfb57
The final-text check failed: conditions_shown, citations_correct.
2026-10-29
Assessment publishedBy this siteGary Marcus
Generative AI has not always worked well enough to justify its costs.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-future-foretold-d8e08f-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-future-foretold-9afaee
The openai review found the criteria unfair to the speaker's words.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-future-foretold-99155a
The anthropic screen reads the claim differently.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-future-foretold-8c2779
The final-text check failed: rationale_supported_by_evidence, citations_correct; overclaims: Bloomberg linked the discrepancy with earlier estimates to different treatment of cloud-provider sales; that revision does not establish a revenue stall [e2]..
2026-10-29
Assessment publishedBy this siteGary Marcus
Schoolchildren using ChatGPT to do their homework are not enough to cover OpenAI's immense costs and cash burn.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-future-foretold-699479-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-disconcerting-hack-of-huggingface-fc03cb
The final-text check failed: rationale_supported_by_evidence; overclaims: On 2026-04-07, The Register reported Anthropic's statement that Mythos Preview could identify and exploit zero-day vulnerabilities in every major operating system and browser when directed to (e3)..
2026-10-29
Assessment publishedBy this siteGary Marcus
The AI industry might slow down only if companies are clearly held liable for the harms they cause.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-disconcerting-hack-of-huggingface-f49a5f-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
AI development should slow down or pause until security and AI safety are under control.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-disconcerting-hack-of-huggingface-e84e11-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-openais-disconcerting-hack-of-huggingface-e05b65
hedged language is not scored, even on a plain reading
2026-10-29
Assessment publishedBy this siteGary Marcus
OpenAI, and for that matter Anthropic, love playing the doom card.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-openais-disconcerting-hack-of-huggingface-d3abd4-a1.