Changes, page 35 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 35 of 37: changes 6801 to 7000 of 7383, newest first. Dates are when the change was recorded.

2026-10-29
Assessment publishedBy this siteGary Marcus
AI agents will be a major focus of attention in the AI field in 2025.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-still-lacks-common-sense-70-years-303ff1-a2.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-still-lacks-common-sense-70-years-0f7941
The final-text check failed: rationale_supported_by_evidence; overclaims: It is contradicted if agents are shown to reliably complete realistic multi-step tasks while no major commonsense advance is documented..
2026-10-29
Assessment publishedBy this siteGary Marcus
Sora's videos are fundamentally more like visual collages than game-engine renderings, which is why they sometimes have a surreal or magic-trick quality.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-still-lacks-common-sense-70-years-0771b0-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
There is a serious need to have independent scientists involved in decisions about managing AI risk.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-risks-short-term-and-longand-what-cee5a8-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
AI poses serious risk in both the short term and the long term, a view Marcus says he and Hinton now share.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-risks-short-term-and-longand-what-a0a037-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
LLM-enhanced terrorism, perhaps coupled with mass AI-generated propaganda, could lead to nuclear war or to the deliberate spread of pathogens worse than covid-19, and many, many people could die and civilization could be utterly disrupted.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-risk-agi-risk-f9d4bb-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-risk-agi-risk-f92470
hedged language is not scored, even on a plain reading
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-risk-agi-risk-eb9455
The openai screen reads the claim differently.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-risk-agi-risk-cc9796
The anthropic screen reads the claim differently.
2026-10-29
Assessment publishedBy this siteGary Marcus
AI does not need to be superintelligent to create serious problems.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-risk-agi-risk-c1e258-a2.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-risk-agi-risk-bfbde1
e5 is dated 2023-03, which is not inside the period described (2023-03-01 to 2023-03-28); it can only be context
2026-10-29
Assessment publishedBy this siteGary Marcus
Marcus does not see computers rapidly and radically improving themselves as an immediate possibility.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-risk-agi-risk-87e51c-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-risk-agi-risk-8711d5
The anthropic screen reads the claim differently.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-risk-agi-risk-633189
The anthropic screen reads the claim differently.
2026-10-29
Assessment publishedBy this siteGary Marcus
We should treat LLMs as a dress rehearsal for future AI, which may be even harder to control.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-risk-agi-risk-5cad52-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
We should think much more about criminal and terrorist misuse of LLMs and how to stop it, not just about Skynet scenarios.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-risk-agi-risk-568684-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Current AI already poses enormous risks we are ill-prepared for, and future AI could well make things worse.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-risk-agi-risk-447613-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-risk-agi-risk-2c2806
The anthropic screen reads the claim differently.
2026-10-29
Assessment publishedBy this siteGary Marcus
We need to get serious about AI risk now, given rapid widespread deployment and unclear use cases and mitigations.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-risk-agi-risk-294b7f-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Adept.AI just raised $350 million to let large language models access pretty much everything, despite LLMs' clear tendencies toward hallucination and unreliability.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-risk-agi-risk-0b2e13-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
We have let too much slide and done too little to face the risks of AI, which is Marcus's reason for signing the global call for AI red lines.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-red-lines-we-should-not-cross-541ada-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
The long-feared moment of generative AI influencing a democratic election may have arrived, through an AI-generated video in Argentina's May 2025 election.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-may-have-just-influenced-argentinas-b6c7b3-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-literacy-hallucinations-and-the-bc3597
The openai review found the criteria unfair to the speaker's words.
2026-10-29
Assessment publishedBy this siteGary Marcus
If educated people like lawyers still haven't gotten the message that LLMs hallucinate, and that they can be fined or publicly humiliated if they rely on LLMs, something is fundamentally wrong.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-literacy-hallucinations-and-the-b8a4d4-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
LLM hallucinations persist and will not go away anytime soon.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-literacy-hallucinations-and-the-a3f7cf-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Publicized cases of AI-fabricated legal citations are probably a tiny fraction of the total incidence.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-literacy-hallucinations-and-the-a2a65e-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Those in the media who have downplayed hallucinations and oversold current AI have done the public a disservice.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-literacy-hallucinations-and-the-922fc7-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
The media should do more to alert people to the shortcomings of LLMs, including how regularly hallucinations happen.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-literacy-hallucinations-and-the-845cea-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Charlotin's database lists 112 cases, an average of a bit less than one publicly reported case a week since ChatGPT became popular, and if May is representative, the problem is getting worse.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-literacy-hallucinations-and-the-663c38-a2.
2026-10-29
Assessment publishedBy this siteGary Marcus
LLMs cannot sanity-check their own work, stick purely to known sources, or flag when they have invented something.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-literacy-hallucinations-and-the-31881c-a2.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-literacy-hallucinations-and-the-2fa8a6
The openai review does not accept e2 as supports.
2026-10-29
Assessment publishedBy this siteGary Marcus
Hallucination is a deep, serious, unsolved problem with generative AI.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-literacy-hallucinations-and-the-2d9d2c-a2.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-literacy-hallucinations-and-the-26186d
The final-text check failed: citations_correct.
2026-10-29
Assessment publishedBy this siteGary Marcus
AI hypesters are currently winning the public narrative.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-literacy-hallucinations-and-the-17241a-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-layoffs-productivity-and-the-klarna-df6dd9
The openai review found the criteria unfair to the speaker's words.
2026-10-29
Assessment publishedBy this siteGary Marcus
Stopping hiring of junior employees in favor of AI risks destroying the pipeline to expertise.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-layoffs-productivity-and-the-klarna-d48dc3-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-layoffs-productivity-and-the-klarna-d47e6b
The final-text check failed: rationale_supported_by_evidence; overclaims: He gives no measure, threshold, population or period..
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-layoffs-productivity-and-the-klarna-d16902
The final-text check failed: rationale_supported_by_evidence; overclaims: So AI did a real share of the work at expert quality, but not most of it..
2026-10-29
Assessment publishedBy this siteGary Marcus
Employers replacing employees with AI may end up reversing course as Klarna did.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-layoffs-productivity-and-the-klarna-7cb59b-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-layoffs-productivity-and-the-klarna-72c829
e2 is dated 2025, which is not inside the period described (2016 to 2025-08-23); it can only be context; e3 is dated 2025, which is not inside the period described (2016 to 2025-08-23); it can only be context
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-layoffs-productivity-and-the-klarna-29ca78
The openai screen reads the claim differently.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-layoffs-productivity-and-the-klarna-181434
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-ai-influencer-hype-gone-wild-bf52a3
The final-text check failed: rationale_supported_by_evidence; overclaims: CNBC reported on its Technology Executive Council survey (e1, dated 2024-10-21)..
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-influencer-hype-gone-wild-756228
The openai review found the criteria unfair to the speaker's words.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-influencer-hype-gone-wild-62c61f
The openai review does not accept e3 as supports.
2026-10-29
Assessment publishedBy this siteGary Marcus
People are not ignoring generative AI; they are waiting to see whether it will work.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-influencer-hype-gone-wild-2d8f86-a2.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-has-sort-of-passed-the-turing-f86ebd
The anthropic screen finds the claim checkable, so it must be graded.
2026-10-29
Assessment publishedBy this siteGary Marcus
Nobody has fully implemented Marcus's proposed comprehension challenge test.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-has-sort-of-passed-the-turing-f44fd7-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Few people in the AI community take the Turing Test seriously, and few should.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-has-sort-of-passed-the-turing-ed0c86-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
To do well in the new study, the system had to be coached to adopt the persona of a young, introverted, slang-using person knowledgeable about internet culture, so its success is still in no small part about tricks.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-has-sort-of-passed-the-turing-d36200-a2.
2026-10-29
Assessment publishedBy this siteGary Marcus
If a careful comprehension challenge test were built, no current AI system would come close to human performance on it.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-has-sort-of-passed-the-turing-cef6a3-a3.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-has-sort-of-passed-the-turing-cc79d7
The final-text check failed: proposition_faithful, citations_correct.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-has-sort-of-passed-the-turing-a26607
The openai screen reads the claim differently.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-has-sort-of-passed-the-turing-864ee4
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-ai-has-sort-of-passed-the-turing-840ee7
The anthropic review found the criteria unfair to the speaker's words.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-has-sort-of-passed-the-turing-77d4ea
The anthropic review found the criteria unfair to the speaker's words.
2026-10-29
Assessment publishedBy this siteGary Marcus
None of the Turing Test chatbots, from ELIZA to ChatGPT 4.5, are anywhere near as intelligent as naive audiences might think.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-has-sort-of-passed-the-turing-6ad3f1-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Fooling people for a few minutes is not the measure of intelligence, and never was.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-has-sort-of-passed-the-turing-63c079-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-has-sort-of-passed-the-turing-581ef1
The openai review does not accept e3 as supports.
2026-10-29
Assessment publishedBy this siteGary Marcus
The Turing Test is a test of human gullibility, not a test of intelligence.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-has-sort-of-passed-the-turing-4b7619-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Today's Turing Test-taking AI systems are better than the earlier generation, but that does not mean anything is 'over'.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-has-sort-of-passed-the-turing-43e4ed-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-has-sort-of-passed-the-turing-415d07
The openai screen reads the claim differently.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-has-sort-of-passed-the-turing-2ed6e5
The openai screen reads the claim differently.
2026-10-29
Assessment publishedBy this siteGary Marcus
In a faceoff, Eugene Goostman might actually beat ChatGPT.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-has-sort-of-passed-the-turing-26cf73-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-has-become-saurons-ring-d31fab
The openai screen reads the claim differently.
2026-10-29
Assessment publishedBy this siteGary Marcus
AI has gone down a path of enabling and accelerating surveillance capitalism, misinformation, nonconsensual deepfake porn and IP theft, which was not inevitable.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-has-become-saurons-ring-7c7863-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
The science of AI has been corrupted by money.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-has-become-saurons-ring-6f6f4c-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
AI development should go back to being genuinely in the public interest.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-has-become-saurons-ring-2c6b01-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Dario Amodei's claim that nearly all code would be written by AI within the next year was made without evidence.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-coding-fantasy-meets-pac-man-eebde9-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Coders and software architects are not on their way out due to AI.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-coding-fantasy-meets-pac-man-e10c00-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-coding-fantasy-meets-pac-man-d41a82
The anthropic screen doubts the attribution.
2026-10-29
Assessment publishedBy this siteGary Marcus
Marcus thinks that most or all of the points he made last week, about the challenges of building and debugging novel systems with sufficient generality, will stand the test of time.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-coding-fantasy-meets-pac-man-bfc56d-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-coding-fantasy-meets-pac-man-b825bc
The openai review found the criteria unfair to the speaker's words.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-coding-fantasy-meets-pac-man-943a21
The final-text check failed: citations_correct.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-coding-fantasy-meets-pac-man-9174f8
a causal claim cannot be assessed on the forecast_outcome axis; only a conditional forecast records a condition status
2026-10-29
Assessment publishedBy this siteGary Marcus
Current AI coding tools do not resolve the challenges of maintaining complex code bases over time.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-coding-fantasy-meets-pac-man-8a2f07-a2.
2026-10-29
Assessment publishedBy this siteGary Marcus
AI coding tools, acting as a fantastic, souped-up version of autocomplete, somewhat increase productivity.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-coding-fantasy-meets-pac-man-756940-a2.
2026-10-29
Assessment publishedBy this siteGary Marcus
AI will serve as a tool to help people write code but will not soon replace people who know how to conceive of, write, and debug code.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-coding-fantasy-meets-pac-man-70c2d9-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Most or all trained coders will soon use AI as part of their workflow, if they are not already.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-coding-fantasy-meets-pac-man-6bcef4-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Marcus guesses that current AI coding tools would struggle to use 6502 assembly well, for multiple reasons including a lack of high-level libraries, fewer Stack Overflow queries to rely on, and a greater need to understand high-level architectural concerns that might be side-stepped in higher-level languages.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-coding-fantasy-meets-pac-man-591df3-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
In The Guardian's test, people using AI tools without coding knowledge did not succeed well at building a Pac-Man clone.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-coding-fantasy-meets-pac-man-4e64bb-a2.
2026-10-29
Assessment publishedBy this siteGary Marcus
If his analysis is correct, LLM dominance may soon begin to decline.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-cartoon-of-the-year-and-five-rereadings-c0819f-a2.
2026-10-29
Assessment publishedBy this siteGary Marcus
If his analysis is correct, newer, sounder alternatives to LLMs may soon start to emerge.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-cartoon-of-the-year-and-five-rereadings-c0819f-2-a2.
2026-10-29
Assessment publishedBy this siteGary Marcus
LLMs are not like humans and never will be.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-cartoon-of-the-year-and-five-rereadings-b39f78-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Generalization is, and always has been, the vital question in AI.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-cartoon-of-the-year-and-five-rereadings-93205b-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
AI desperately needs world models.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-cartoon-of-the-year-and-five-rereadings-9098cc-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Neurosymbolic AI will eventually displace pure LLMs.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-cartoon-of-the-year-and-five-rereadings-836b91-a3.
2026-10-29
Assessment publishedBy this siteGary Marcus
LLMs are unable to stop hallucinating.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-cartoon-of-the-year-and-five-rereadings-15925b-a2.
2026-10-29
Assessment publishedBy this siteGary Marcus
If verified-yet-anonymous proof-of-human is required for high-reach interactions, it will make it mathematically difficult and financially ruinous for one operator to secretly run ten thousand accounts.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-bot-swarms-threaten-to-undermine-f4601e-a2.
2026-10-29
Assessment publishedBy this siteGary Marcus
Venture capital is already helping industrialize astroturfing; for example, Andreessen Horowitz-backed Doublespeed advertises a way to orchestrate actions on thousands of social accounts and to mimic natural user interaction on physical devices so the activity appears human.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-bot-swarms-threaten-to-undermine-ee8fd6-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Policymakers should incentivize cryptographic attestations and reputation standards to make fake personas expensive.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-bot-swarms-threaten-to-undermine-e8af7f-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-bot-swarms-threaten-to-undermine-e34d67
The final-text check failed: rationale_supported_by_evidence; overclaims: Several dated reports show influence operations using LLMs to produce disinformation within about 18 months of ChatGPT's launch in November 2022. | A second AI developer documents actors using LLMs for influence campaigns, which confirms the trend is broad..
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-bot-swarms-threaten-to-undermine-d3b635
hedged language is not scored, even on a plain reading
2026-10-29
Assessment publishedBy this siteGary Marcus
Legislation must guarantee vetted academic and civil society researchers free, privacy-preserving access to platform data.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-bot-swarms-threaten-to-undermine-bdba92-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-bot-swarms-threaten-to-undermine-b95028
hedged language is not scored, even on a plain reading
2026-10-29
Assessment publishedBy this siteGary Marcus
Even so-called 'thinking' (reasoning) AI models are vulnerable to training-data poisoning by swarms.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-bot-swarms-threaten-to-undermine-b204d7-a2.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-bot-swarms-threaten-to-undermine-95adbb
hedged language is not scored, even on a plain reading
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-bot-swarms-threaten-to-undermine-8b5436
The two reviews differ: anthropic causal_assessment/not_applicable/mixed, openai causal_assessment/not_applicable/contradicted.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-bot-swarms-threaten-to-undermine-89a5b6
The final-text check failed: conditions_shown, rationale_supported_by_evidence; overclaims: So it does not meet the early-resolution rule, which needs Breakout Scale Category 4 or higher or measured opinion shifts at population scale..
2026-10-29
Assessment publishedBy this siteGary Marcus
An independent, non-governmental AI Influence Observatory of academic groups and NGOs should be created to detect coordinated swarms.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-bot-swarms-threaten-to-undermine-85ce38-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Purely reactive defenses will fail; defenders should proactively stress-test using agent-based simulations of AI swarms.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-bot-swarms-threaten-to-undermine-7e7066-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-bot-swarms-threaten-to-undermine-770b18
The openai review reads the claim differently.
2026-10-29
Assessment publishedBy this siteGary Marcus
Social media platforms must abandon episodic whack-a-mole takedowns in favor of continuous monitoring for improbable coordination.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-bot-swarms-threaten-to-undermine-5052c3-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Emerging AI bot swarms can coordinate fleets of synthetic personas that are hard to distinguish from real communities.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-bot-swarms-threaten-to-undermine-21e95b-a2.
2026-10-29
Assessment publishedBy this siteGary Marcus
When it costs pennies to coordinate a fake mob and takes moments to counterfeit a human identity, the public square is left wide open to attack.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-bot-swarms-threaten-to-undermine-130fd3-a2.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-bot-swarms-threaten-to-undermine-0e867c
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-ai-bot-swarms-threaten-to-undermine-02d8e5
The anthropic screen reads the claim differently.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-as-reality-television-f38175
The anthropic screen doubts the attribution.
2026-10-29
Assessment publishedBy this siteGary Marcus
Negative press coverage of Sam Altman and OpenAI, once seen as able to do no wrong, keeps coming.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-as-reality-television-d6d3db-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Gary Marcus wishes that AI these days were more about research than about money, politics, and ego.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-as-reality-television-8d3c76-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-as-reality-television-82fbde
The openai review reads the claim differently.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-as-reality-television-82fbde-2
The anthropic review abstained.
2026-10-29
Assessment publishedBy this siteGary Marcus
OpenAI's ChatGPT task-scheduling agent will not work reliably.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-agents-hype-versus-reality-redux-d3f1d2-a3.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-agents-hype-versus-reality-redux-b8e3b1
the conclusive event's date differs from its evidence
2026-10-29
Assessment publishedBy this siteGary Marcus
Eventually, though not soon, AI agents will do a huge fraction of human knowledge work, perhaps physical labor too, and be worth trillions.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-agents-hype-versus-reality-redux-a70782-a3.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-agents-hype-versus-reality-redux-9e0e0e
The final-text check failed; overclaims: In late 2025, agents automated very little end-to-end remote work..
2026-10-29
Assessment publishedBy this siteGary Marcus
AI agents probably will not do a huge fraction of human knowledge work this decade, except in narrow use cases.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-agents-hype-versus-reality-redux-9e0e0e-2-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-agents-hype-versus-reality-redux-775074
The two reviews differ: anthropic forecast_outcome/elapsed/supported, openai forecast_outcome/elapsed/unresolved.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-agents-hype-versus-reality-redux-6d5392
The openai review found the criteria unfair to the speaker's words.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-agents-hype-versus-reality-redux-65244e
The openai review found the criteria unfair to the speaker's words.
2026-10-29
Assessment publishedBy this siteGary Marcus
As of January 2025 there is no principled, trustworthy solution to the hallucination problem.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-agents-hype-versus-reality-redux-65244e-2-a3.
2026-10-29
Assessment publishedBy this siteGary Marcus
In 2025, AI agents will exist only as demos.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-agents-hype-versus-reality-redux-5a7e45-a3.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-agents-hype-versus-reality-redux-4af992
The final-text check failed: rationale_supported_by_evidence; overclaims: The source text does not settle which. | Whether the claim refers to any agent or only to reliable general agents is unresolved..
2026-10-29
Assessment publishedBy this siteGary Marcus
In the long term, everyone will have their own AI agents and companies will have armies of them.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-agents-hype-versus-reality-redux-2c1aed-a3.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-agents-hype-versus-reality-redux-0d2294
The final-text check failed; overclaims: This is the strongest 2025 counterpoint, though they still failed nearly a quarter of tasks..
2026-10-29
Assessment publishedBy this siteGary Marcus
Someday, current agents' failure to produce much of lasting value will be seen as a poignant reflection of a massive intellectual and economic mistake: investing almost solely in LLMs as a kind of hopeful shortcut.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-agents-have-so-far-mostly-been-e6c362-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
AI agents in 2025 have largely not delivered the boundary-pushing capabilities that Kevin Roose promised.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-agents-have-so-far-mostly-been-e4c832-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-agents-have-so-far-mostly-been-e153fc
The anthropic screen doubts the attribution.
2026-10-29
Assessment publishedBy this siteGary Marcus
Yet more good money, in staggeringly large amounts, is being poured after bad into generative AI.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-agents-have-so-far-mostly-been-dce74f-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
He suspects it will be typical that AI agent errors compound over time, as in Penrose's account-balance benchmark.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-agents-have-so-far-mostly-been-d4bf7b-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-agents-have-so-far-mostly-been-d441b1
The anthropic screen finds the claim checkable, so it must be graded.
2026-10-29
Assessment publishedBy this siteGary Marcus
The gap between AI agent demos and real-world performance will probably persist for years to come.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-agents-have-so-far-mostly-been-b08233-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-agents-have-so-far-mostly-been-ab9cd9
The anthropic review found the criteria unfair to the speaker's words.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-agents-have-so-far-mostly-been-a29037
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-ai-agents-have-so-far-mostly-been-9f0741
The final-text check failed: rationale_supported_by_evidence; overclaims: No accepted measure exists of the economic value generated by AI agents..
2026-10-29
Assessment publishedBy this siteGary Marcus
Because LLMs are fundamentally unreliable, multi-step agentic tasks accumulate errors that sometimes lead to catastrophe.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-agents-have-so-far-mostly-been-9a2d36-a3.
2026-10-29
Assessment publishedBy this siteGary Marcus
AI agents should not be trusted with your calendar, let alone your credit cards.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-agents-have-so-far-mostly-been-95bc55-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Perhaps, after enough failures of AI agents, reality may finally start to settle in.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-agents-have-so-far-mostly-been-95bc55-2-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-agents-have-so-far-mostly-been-8603b3
The anthropic review found the criteria unfair to the speaker's words.
2026-10-29
Assessment publishedBy this siteGary Marcus
Generative AI should be penalized for failing to meet agent expectations, yet investors and big tech keep increasing investment.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-agents-have-so-far-mostly-been-79ce93-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
LLMs are unlikely ever to provide the substrate needed for working AI agents.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-agents-have-so-far-mostly-been-762555-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
ChatGPT agent is unlikely to produce wildly different results on Penrose's account-balance benchmark.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-agents-have-so-far-mostly-been-7262b8-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
To his knowledge, none of the big companies' AI agents are reliable, except maybe in very narrow use cases.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-agents-have-so-far-mostly-been-5f4295-a2.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-agents-have-so-far-mostly-been-5c92a3
The anthropic review found the criteria unfair to the speaker's words.
2026-10-29
Assessment publishedBy this siteGary Marcus
Venture capitalists style themselves as brave funders of innovation, but in the current climate they are anything but brave funders of innovation.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-agents-have-so-far-mostly-been-4dd8eb-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Current AI agents are grossly vulnerable to cyberattacks because of their superficial understanding of their tasks.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-agents-have-so-far-mostly-been-452803-a3.
2026-10-29
Assessment publishedBy this siteGary Marcus
An important, publicly visible system that can be beaten in .001% of the times it is attacked is a mess.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-agents-have-so-far-mostly-been-3e81d8-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
The sentiment that AI agents' reality doesn't match the hype has grown from March to August 2025.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-agents-have-so-far-mostly-been-3c3c51-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Hallucinations frequently occur in AI agents.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-agents-have-so-far-mostly-been-38b9ed-a2.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-agents-have-so-far-mostly-been-22a997
e1 is decisive for a financial claim but states no metric scope; e2 is decisive for a financial claim but states no metric scope; e3 is decisive for a financial claim but states no metric scope
2026-10-29
Assessment publishedBy this siteGary Marcus
LLMs and current agents are driven by mimicry rather than deep understanding.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-agents-have-so-far-mostly-been-1fbf99-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-ai-agents-have-so-far-mostly-been-1e53ce
The openai review found the criteria unfair to the speaker's words.
2026-10-29
Assessment publishedBy this siteGary Marcus
Mistakes and errors have greatly limited the utility of agents like ChatGPT agent.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-agents-have-so-far-mostly-been-0c896c-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Current AI systems do not understand what they are doing and have no concept of actions like deleting a database.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-agents-have-so-far-mostly-been-0bef71-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
All the big AI companies have introduced AI agents as of August 2025.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-ai-agents-have-so-far-mostly-been-026bb0-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
LLMs are a detour, and the field's patience for them will likely run out.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-will-not-happen-in-your-lifetime-f95c16-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
What AGI development is missing is new ideas and large-scale databases of machine-interpretable knowledge.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-will-not-happen-in-your-lifetime-f4c6e5-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-agi-will-not-happen-in-your-lifetime-f02fe0
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-agi-will-not-happen-in-your-lifetime-eeb229
The anthropic review reads the claim differently.
2026-10-29
Assessment publishedBy this siteGary Marcus
Current AI should not depend on random details of training corpora that scientists cannot inspect; this is a bad idea.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-will-not-happen-in-your-lifetime-edb284-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Marcus thinks better AI, in which we could directly instill values, might actually make us safer than the current world, in which so much depends on the vagaries of what happens to be in this or that training set.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-will-not-happen-in-your-lifetime-ec5161-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-agi-will-not-happen-in-your-lifetime-d913eb
e1 is dated 2023-01, which is not inside the period described (2022-11-01 to 2023-01-22); it can only be context; e2 is dated 2023-01, which is not inside the period described (2022-11-01 to 2023-01-22); it can only be context; e4 is dated 2023-01, which is not inside the period described (2022-11-01 to 2023-01-22); it can only be context
2026-10-29
Assessment publishedBy this siteGary Marcus
The neurosymbolic AI community is growing quickly, with conferences going from dozens to thousands of attendees.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-will-not-happen-in-your-lifetime-d33833-a2.
2026-10-29
Assessment publishedBy this siteGary Marcus
Nothing is guaranteed: we might not get to AGI because we blow ourselves up, or because we waste so much time on large language models that other important challenges such as semantics, causality, commonsense reasoning and abduction never get the attention they deserve.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-will-not-happen-in-your-lifetime-d2122b-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Even if a new chip type is required for AGI, that could take about thirty years rather than a hundred.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-will-not-happen-in-your-lifetime-cc3e40-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-agi-will-not-happen-in-your-lifetime-c5100a
The anthropic review found the criteria unfair to the speaker's words.
2026-10-29
Assessment publishedBy this siteGary Marcus
Bias is likely to be a huge issue with large language models like ChatGPT.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-will-not-happen-in-your-lifetime-c07a21-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-agi-will-not-happen-in-your-lifetime-bff791
The openai screen reads the claim differently.
2026-10-29
Assessment publishedBy this siteGary Marcus
Marcus takes Meta's Cicero as a hopeful sign that the scaling-über-alles narrative is losing force and that more and more people are open to new things.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-will-not-happen-in-your-lifetime-bce14d-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Minor abuses of ChatGPT have already been seen, and plagiarism, for example, will never be the same.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-will-not-happen-in-your-lifetime-bc870c-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-agi-will-not-happen-in-your-lifetime-b85960
hedged language is not scored, even on a plain reading
2026-10-29
Assessment publishedBy this siteGary Marcus
LLMs are certainly not reliable in the way general intelligences like the Star Trek computer would be.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-will-not-happen-in-your-lifetime-b4017e-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
2023 will see the first death attributable to large language models.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-will-not-happen-in-your-lifetime-b1a88d-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-agi-will-not-happen-in-your-lifetime-aa3d11
The anthropic screen reads the claim differently.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-agi-will-not-happen-in-your-lifetime-a3a02c
e1 is dated 2023-01, which is not inside the period described (2023-01-01 to 2023-01-22); it can only be context; e2 is dated 2023-01, which is not inside the period described (2023-01-01 to 2023-01-22); it can only be context
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-agi-will-not-happen-in-your-lifetime-9f6905
The openai screen reads the claim differently.
2026-10-29
Assessment publishedBy this siteGary Marcus
A decent fraction of the AI community has already recognized that the primary forte of large language models is generating bullshit, and many also see that the current path to autonomous vehicles is not making the progress that some expected.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-will-not-happen-in-your-lifetime-98aaf5-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
ChatGPT is a dress rehearsal for more potent AI, and Marcus would not say society has gone through it with flying colors.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-will-not-happen-in-your-lifetime-965c4c-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-agi-will-not-happen-in-your-lifetime-94b41f
The anthropic screen reads the claim differently.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-agi-will-not-happen-in-your-lifetime-7b3991
The anthropic screen reads the claim differently.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-agi-will-not-happen-in-your-lifetime-78cb9d
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-agi-will-not-happen-in-your-lifetime-6ccb63
The anthropic screen reads the claim differently.
2026-10-29
Assessment publishedBy this siteGary Marcus
Marcus is still not seeing a principled argument that constructing flexible, adaptive systems with commonsense reasoning, deeper semantics and so on will take us past the end of the current century.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-will-not-happen-in-your-lifetime-6c3453-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
ChatGPT will lead to more circles of fake websites created to sell advertising.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-will-not-happen-in-your-lifetime-6ae4f4-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-agi-will-not-happen-in-your-lifetime-69594e
The openai screen reads the claim differently.
2026-10-29
Assessment publishedBy this siteGary Marcus
There is unlikely to be a single magic singularity moment for AGI.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-will-not-happen-in-your-lifetime-68381c-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Achieving AGI will require integrating deep learning with explicit knowledge and explicit cognitive models.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-will-not-happen-in-your-lifetime-5e6cf0-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Large language models are not remotely close to AGI; they lack semantics, reasoning, common sense and theory of mind.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-will-not-happen-in-your-lifetime-49f087-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
AGI is unlikely to be exactly humanlike; adaptivity is key and probably achievable by multiple means.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-will-not-happen-in-your-lifetime-3aa6df-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
We might not get to AGI if we can't come up with something better than current systems, which are so greedy in data and energy consumption that they could asymptote before anything really general emerges.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-will-not-happen-in-your-lifetime-39ca18-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
ChatGPT is in many ways a far inferior intelligence and cannot be trusted.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-will-not-happen-in-your-lifetime-2b78a1-a2.
2026-10-29
Assessment publishedBy this siteGary Marcus
AI development should include many voices beyond technologists.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-will-not-happen-in-your-lifetime-28cf48-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
The flexible, general intelligence that defines AGI does not require sentience.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-will-not-happen-in-your-lifetime-23de23-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-agi-will-not-happen-in-your-lifetime-22191e
The final-text check failed: proposition_faithful, citations_correct.
2026-10-29
Assessment publishedBy this siteGary Marcus
Marcus thinks the field is wasting funding and bright young minds on an approach that probably is not on the right path.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-will-not-happen-in-your-lifetime-1e834f-a1.
2026-10-29
HeldBy this siteGary Marcus
Held record marcus-agi-will-not-happen-in-your-lifetime-1abf63
The openai review found the criteria unfair to the speaker's words.
2026-10-29
Assessment publishedBy this siteGary Marcus
Within 5 years, a substantial fraction of AI effort will look beyond the current LLM-focused approaches.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-will-not-happen-in-your-lifetime-1649d2-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
AI systems will gradually become more reliable and more general.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-will-not-happen-in-your-lifetime-143ce1-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
Troll farms may start using ChatGPT to create misinformation at unprecedented scale.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-will-not-happen-in-your-lifetime-13231a-a1.
2026-10-29
Assessment publishedBy this siteGary Marcus
LLMs often use words without deeply understanding what they mean.
First assessment published after both reviews and the final-text check passed.
Assessment marcus-agi-versus-broad-shallow-intelligence-ff3887-a1.