Claims checked against evidence by Gary Marcus, page 13 of 15
Deadlines are judged as of 2026-10-29. About Gary Marcus and the sources covered.
Claims checked against evidence by Gary Marcus (1471)
Each of these claims was checked against dated evidence.
Page 13 of 15: claims 1201 to 1300 of 1471, ordered by the date first said, newest first. To find a claim by its words, use search.
Assessment: UnresolvedSaid
Observation
“Neurosymbolic AI might offer a better way of resolving these issues, since it traffics in facts and explicit reasoning, but neurosymbolic AI is out of fashion, and neither well-developed nor strongly supported by the VC world.”
Our reading: Neurosymbolic AI is out of fashion, not well developed, and not strongly supported by venture capital.
No horizon Evidence as of
Limitation: The field boundaries and all three qualitative thresholds are unspecified.
Source: How China might crush the West in the race to AGI if we don’t up our regulatory game
Assessment: UnresolvedSaid
Observation
“At present, however, there is no legal requirement in the west that AI companies resolve the truthfulness or alignment issues.”
Our reading: There is currently no legal requirement in the West for AI companies to resolve truthfulness or alignment issues.
No horizon Evidence as of
Limitation: The schema has no source type for official regulatory announcements; official_statistics is used for the government documents.
Source: How China might crush the West in the race to AGI if we don’t up our regulatory game
Assessment: UnresolvedSaid
Forecast
“Their forte has been and always will be making shit up; they can’t for the life of them (speaking metaphorically of course) check facts.”
Our reading: For GPT models on their own, making things up has been and always will be their forte, and they cannot check facts.
No deadline stated Evidence as of
Limitation: A literal zero-ability reading would yield a different assessment of the fact-checking component.
Source: GPT-5 and irrational exuberance
Assessment: ContradictedSaid
Observation
“Essentially none of the fancy cognitive things GPT-4 is rumored to do stand up.”
Our reading: Essentially none of the advanced cognitive abilities attributed to GPT-4 hold up under scrutiny.
No horizon Evidence as of
Limitation: The result depends on counting controlled reasoning performance among the unspecified 'fancy cognitive things'.
Source: GPT-5 and irrational exuberance
Assessment: SupportedSaid
Observation
“Maarten Sap from Yejin Choi’s lab dove deeper, with a more thorough investigation, and confirmed what we suspected: the claim that large language models have mastered theory of mind is a myth:”
Our reading: The claim that large language models have mastered theory of mind is a myth.
No horizon Evidence as of
Limitation: The studies test particular models, tasks and prompts, rather than every aspect of theory of mind [e5, e6].
Source: GPT-5 and irrational exuberance
Assessment: UnresolvedSaid
Forecast
“It will be even better than GPT-4 at creating convincing sounding prose.”
Our reading: GPT-5 will be better than GPT-4 at creating convincing-sounding prose.
No deadline stated Evidence as of
Limitation: Writing preference varies with genre, prompting, and generation settings.
Source: GPT-5 and irrational exuberance
Assessment: UnresolvedSaid
Forecast
“GPT-5 is not going to be magic. It will not give China, the US, or anyone else interstellar starships; it will not give anyone an insuperable military advantage.”
Our reading: GPT-5 will not give China, the US, or anyone else interstellar starships, and it will not give anyone an insuperable military advantage.
No deadline stated Evidence as of
Limitation: Military capabilities may not be publicly observable.
Source: GPT-5 and irrational exuberance
Assessment: SupportedSaid
Observation
“It’s not smart enough to filter out falsehood, but just smart enough to be dangerous, creating and spreading falsehoods it fails to verify.”
Our reading: GPT cannot filter out falsehood and creates and spreads falsehoods it fails to verify, making it dangerous.
No horizon Evidence as of
Limitation: Volokh's unsuccessful searches support his assessment that the examples were fabricated, but are not exhaustive proof that no corresponding source exists [e2].
Source: GPT-5 and irrational exuberance
Assessment: UnresolvedSaid
Observation
“(Not one of my predictions about GPT-4 proved to be incorrect; every flaw I predicted would persist persisted.)”
Our reading: Marcus says none of his predictions about GPT-4 proved incorrect, and every flaw he predicted would persist in GPT-4 did persist.
No horizon Evidence as of
Limitation: Confirming several central predictions cannot establish a universal track-record claim.
Source: GPT-5 and irrational exuberance
Assessment: UnresolvedSaid
Observation
“GPT-4 can’t even play a good game of chess (it’s about as good as chess computers from 1978).”
Our reading: GPT-4 cannot play a good game of chess and is about as good as 1978 chess computers.
No horizon Evidence as of
Limitation: The admitted GPT-4 evidence consists of one recorded game and an account of two attempted games [e1, e2].
Source: GPT-5 and irrational exuberance
Assessment: UnresolvedSaid
Observation
“And very little of what GPT-4 does is fully replicable. Sometimes summarization (aka compression) works (super cool!) and sometimes it doesn’t, as many were quick to point out:”
Our reading: Very little of what GPT-4 does is fully replicable; its summarization sometimes works and sometimes fails.
No horizon Evidence as of
Limitation: The task population and meaning of full replication are unspecified.
Source: GPT-5 and irrational exuberance
Assessment: UnresolvedSaid
Conditional forecast
“Some form of AI may eventually do everything people are imagining, revolutionizing science and technology and so on, but LLMs will be at most only a tiny part of whatever as-yet-uninvented technology does that.”
Our reading: If some as-yet-uninvented form of AI eventually does everything people are imagining, such as revolutionizing science and technology, LLMs will be at most only a tiny part of it.
No deadline stated Evidence as of
Only if: if some form of AI eventually does everything people are imagining
Source: GPT-5 and irrational exuberance
Assessment: UnresolvedSaid
Forecast
“It might be on the nice-to-have list; it will not mark the mythical singularity.”
Our reading: GPT-5 will not mark the singularity.
No deadline stated Evidence as of
Limitation: The cited failure rate concerns one task, not a measurement of sustained, self-amplifying intelligence improvement [e1].
Source: GPT-5 and irrational exuberance
Assessment: UnresolvedSaid
Observation
“Vulnerable patients shouldn’t be talking to chatbots that aren’t competent for this situation, and too much of the discussion in the media have made chatbots sound more intelligent and even more friendly than they are.”
Our reading: Too much media discussion has made chatbots sound more intelligent and friendly than they are.
No horizon Evidence as of
Limitation: The phrase “too much” remains partly evaluative.
Assessment: SupportedSaid
Observation
“The fundamental issue is that chatbots are neither reliable nor truthful, and at the same time don’t really grasp things like human psychology.)”
Our reading: Chatbots are neither reliable nor truthful and do not really grasp human psychology.
No horizon Evidence as of
Limitation: The result depends on the operational reading of really grasp.
Assessment: SupportedSaid
Forecast
“Two reasons, first, because I believe that more incidents like this will happen, perhaps many more, hundreds or thousands, given how rapidly chatbots are being deployed, and the near total lack of regulation on how they can be used.”
Our reading: More chatbot-associated incidents like this suicide will happen, perhaps hundreds or thousands, given rapid deployment and near-total lack of regulation.
Settled before its deadline Evidence as of
Limitation: The filings do not establish causation or liability [e1, e2, e3].
Assessment: MixedSaid
Observation
“Two reasons, first, because I believe that more incidents like this will happen, perhaps many more, hundreds or thousands, given how rapidly chatbots are being deployed, and the near total lack of regulation on how they can be used.”
Our reading: There is a near-total lack of regulation on how chatbots can be used, and they are being deployed rapidly.
No horizon Evidence as of
Limitation: The result depends on reading regulation to include existing binding law.
Assessment: SupportedSaid
Observation
“The Chatbot involved was GPT-J, not GPT-3 or ChatGPT (contra some news headlines that were erroneous).”
Our reading: The chatbot involved in the Belgian suicide case was GPT-J, not GPT-3 or ChatGPT.
No horizon Evidence as of
Limitation: The evidence is reporting based on developer interviews, rather than an independently inspected deployment record.
Assessment: SupportedSaid
Observation
“Whether or not you want a pause, the reality is that large language models remain exactly what I have always described them to be: bulls in a china shop, powerful, reckless, and difficult to control.”
Our reading: Large language models remain powerful, reckless, and difficult to control.
No horizon Evidence as of
Limitation: Reckless is assessed as a description of outputs, not intent.
Assessment: SupportedSaid
Causal claim
“At some level the petition worked; Capitol Hill, for example, pretty clearly got the message that we can’t sit around forever on these issues.”
Our reading: At some level the FLI pause petition worked; for example, Capitol Hill pretty clearly got the message that we can't sit around forever on AI issues.
No horizon Evidence as of
Limitation: The evidence supports a limited claim of congressional attention and favorable response, not Congress-wide agreement [e1].
Assessment: UnresolvedSaid
Observation
“Virtually everyone, even at OpenAI, has acknowledged that there are serious risks, but thus far few tangible steps have been taken to mitigate them—either by government or industry.”
Our reading: Although virtually everyone, including people at OpenAI, acknowledges serious AI risks, government and industry have so far taken few tangible steps to mitigate them.
No horizon Evidence as of
Limitation: Government guidance, legal requirements and technical safeguards differ substantially in reach and effectiveness.
Source: The Open Letter Controversy
Assessment: SupportedSaid
Observation
“it didn't call for a ban on the vast majority of AI research, only for a brief pause on one very specific project with a technology that has *known* risks with no known solutions.”
Our reading: The one very specific project the letter asked to pause, training systems exceeding GPT-4, involves a technology that has known risks with no known solutions.
No horizon Evidence as of
Limitation: The outcome depends on interpreting solutions as dependable resolution rather than any risk reduction.
Source: The Open Letter Controversy
Assessment: ContradictedSaid
Observation
“None of the top labs actually is remotely transparent, governance basically doesn’t exist yet, and there are basically no safeguards actually in place.”
Our reading: None of the top AI labs is remotely transparent.
No horizon Evidence as of
Limitation: The result depends on counting substantive model disclosures as meaningful laboratory transparency.
Source: The Open Letter Controversy
Assessment: SupportedSaid
Observation
“I personally haven't changed; I still think that LLMs are unreliable, and still think that they are a very poor basis for factuality.”
Our reading: LLMs are unreliable.
No horizon Evidence as of
Limitation: The evidence is developer disclosure rather than an independent cross-model audit.
Source: The Open Letter Controversy
Assessment: SupportedSaid
Observation
“A company called Adept.AI just raised $350 million dollars to do just that, to allow large language models to access, well, pretty much everything (aiming to “supercharge your capabilities on any software tool or API in the world” with LLMs, despite their clear tendencies towards hallucination and unreliability).”
Our reading: Adept.AI just raised $350 million to let large language models access pretty much everything, despite LLMs' clear tendencies toward hallucination and unreliability.
No horizon Evidence as of
Limitation: Adept's main domain blocked the page reader; the quotations come from readable reproductions on adept.futurefabric.co.
Source: AI risk ≠ AGI risk
Assessment: SupportedSaid
Causal claim
“But here’s the thing: although a lot of the literature equates artificial intelligence risk with the risk of superintelligence or artificial general intelligence, you don’t have to be superintelligent to create serious problems.”
Our reading: AI does not need to be superintelligent to create serious problems.
No horizon Evidence as of
Limitation: This incident establishes that serious harm is possible, not how frequently it occurs [e1].
Source: AI risk ≠ AGI risk
Assessment: SupportedSaid
Observation
“The problem of hallucinations is not solved; reliability is not solved; planning on complex tasks is (as the authors themselves acknowledge) not solved.”
Our reading: GPT-4 has not solved hallucinations, reliability, or planning on complex tasks.
No horizon Evidence as of
Limitation: The planning example concerns the early research model examined in Sparks.
Assessment: UnresolvedSaid
Causal claim
“The paper’s core claim— “GPT-4 attains a form of general intelligence [as] demonstrated by its core mental capabilities (such as reasoning, creativity, and deduction)”—literally cannot be tested with serious scrutiny, because the scientific community has no access to the training data.”
Our reading: The Sparks of AGI paper's claim that GPT-4 attains a form of general intelligence cannot be seriously tested because scientists have no access to its training data.
No horizon Evidence as of
Limitation: This is not support for the claim that all serious testing requires training-data access.
Assessment: UnresolvedSaid
Observation
“if the definition of reasoning is to obtain valid conclusions from known facts, GPT-4 frequently falls short there, too.”
Our reading: GPT-4 frequently fails to reason, if reasoning means reaching valid conclusions from known facts.
No horizon Evidence as of
Limitation: Later evaluations were not rejected merely because they were published after the post; the problem is their uncertain connection to the target model period.
Source: This Week in AI Doublespeak
Assessment: SupportedSaid
Observation
“True LLMs don’t (just) memorize, and true that their models make for lousy databases (for a database a hallucination is an outright fail)”
Our reading: OpenAI's LLMs make lousy databases because hallucinations amount to outright failures for a database.
No horizon Evidence as of
Limitation: This does not establish that most answers were false.
Source: This Week in AI Doublespeak
Assessment: SupportedSaid
Observation
“We don’t know how big it is; we don’t know what the architecture is, we don’t know how much energy was used; we don’t how many processors were used; we don’t know what it was trained on etc.”
Our reading: OpenAI has not disclosed GPT-4's size, architecture, energy use, processor count, or training data.
No horizon Evidence as of
Limitation: The supplied excerpts do not directly establish the disclosure status of every listed category, particularly training energy and processor count.
Assessment: UnresolvedSaid
Observation
“It doesn’t actually solve any of the core problems of truthfulness and reliability that I laid out in my infamous March 2022 essay Deep Learning is Hitting a Wall.”
Our reading: GPT-4 does not solve any of the core truthfulness and reliability problems identified in Marcus's March 2022 essay.
No horizon Evidence as of
Limitation: This is an evidence gap for the full assertion, not evidence that the problems were solved.
Assessment: UnresolvedSaid
Observation
“On its own, it still requires frequent, massive retraining to be current with the news; GPT 3.5 knew nothing of 2022; GPT-4 seems to know little of 2023.”
Our reading: On its own, GPT-4 still requires frequent massive retraining to stay current with the news.
No horizon Evidence as of
Limitation: A knowledge cutoff is not evidence of complete ignorance of every later event.
Assessment: UnresolvedSaid
Causal claim
“Because of all that, we don’t know what explains its success, and we don’t know how to predict its failures:”
Our reading: Because GPT-4's size, architecture, energy use, processors, and training data have not been disclosed, we don't know what explains its success and we don't know how to predict its failures.
No horizon Evidence as of
Limitation: The strongest quantitative evidence comes from OpenAI's own evaluations.
Assessment: SupportedSaid
Observation
“It can see (though we don’t have a lot of details on that yet); it does astonishingly well on a whole bunch of standardized tests, like LSATs, GREs, and SATs.”
Our reading: GPT-4 has vision capability and performs astonishingly well on standardized tests such as the LSAT, GRE, and SAT.
No horizon Evidence as of
Limitation: The evidence is OpenAI's own evaluation report, without independent replication in the packet [e1].
Assessment: UnresolvedSaid
Observation
“Alignment is still shaky; you still wouldn’t be able to use it reliably to guide robots or scientific discovery, the kinds of things that made me excited about A(G)I in the first place.”
Our reading: GPT-4's alignment remains shaky and it cannot be used reliably to guide robots or scientific discovery.
No horizon Evidence as of
Limitation: Neither missing launch-period demonstrations nor general language-model errors prove application-wide inability.
Assessment: UnresolvedSaid
Observation
“Large language models are here to stay, their recipe well understood, and not hard for people skilled in the art to replicate; large pretrained models are out their for the taking on the dark web.”
Our reading: LLMs are here to stay: their recipe is well understood, skilled practitioners can replicate them without much difficulty, and large pretrained models are freely available on the dark web.
No horizon Evidence as of
Limitation: The leak is established through reputable reporting rather than a directly inspected primary distribution archive.
Source: Should we worry more about short-term AI risk or long-term AI risk?
Assessment: UnresolvedSaid
Forecast
“It will be an uphill, ongoing move-and-countermove arms race from here”
Our reading: Fighting AI-generated misinformation will be an uphill, ongoing arms race of moves and countermoves.
No deadline stated Evidence as of
Limitation: Laboratory watermark results do not directly measure misinformation impact.
Source: Should we worry more about short-term AI risk or long-term AI risk?
Assessment: SupportedSaid
Observation
“To me the only solution to the long-term risk issue is to build machines with consensus human values, but we are a long way from knowing how to do that”
Our reading: The field is a long way from knowing how to build machines with consensus human values.
No horizon Evidence as of
Limitation: The evidence includes research organizations' own assessments rather than a field-wide proof.
Source: Should we worry more about short-term AI risk or long-term AI risk?
Assessment: UnresolvedSaid
Observation
“A lot of machine learning researchers are worried. And yet most people carry on as if they weren’t.”
Our reading: Many machine learning researchers worry about very bad outcomes from AI, yet most people carry on as if they didn't.
No horizon Evidence as of
Limitation: The researcher survey had a 17% response rate.
Source: Should we worry more about short-term AI risk or long-term AI risk?
Assessment: UnresolvedSaid
Observation
“Minerva, for example, is trained on an immense amount of data, and ostensibly captures two-digit arithmetic, but it fails altogether on 4-digit multiplication,”
Our reading: Minerva, despite immense training data and apparent two-digit arithmetic ability, fails altogether on 4-digit multiplication.
No horizon Evidence as of
Limitation: The decisive data exist only in a figure that the text tools cannot read.
Source: Caricaturing Noam Chomsky
Assessment: UnresolvedSaid
Observation
“The request is perfectly clear, not a nonsensical question in any way. The prompt is perfectly sensible; the system just isn’t up to the job.”
Our reading: The request in his example is perfectly clear and sensible, not nonsensical in any way, and the system just isn't up to the job.
No horizon Evidence as of
Limitation: The retrieval gap is not evidence that the prompt was unclear or that the response was correct.
Source: Caricaturing Noam Chomsky
Assessment: SupportedSaid
Observation
“Still, even though Chomsky’s argument surely could have used considerably more nuance (see my article with Ernest Davis on how not to evaluate GPT), his overall point is correct: LLMs don’t reliably understand the world.”
Our reading: LLMs do not reliably understand the world.
No horizon Evidence as of
Limitation: It rests on a behavioural reading of 'understand'. Someone who reads 'understanding' differently could dispute it.
Source: Caricaturing Noam Chomsky
Assessment: UnresolvedSaid
Observation
“The same goes for essentially everything an LLMs talks about; there is a degree of generalization to similar examples, but distribution shift is hard (the crux of my own work going back to 1998), and nearly any generalization can be easily broken.”
Our reading: For essentially everything an LLM talks about, there is a degree of generalization to similar examples, but distribution shift is hard and nearly any generalization can be easily broken.
No horizon Evidence as of
Limitation: All the evidence comes from math and numerical tasks, not the range of domains the claim covers.
Source: Caricaturing Noam Chomsky
Assessment: SupportedSaid
Observation
“Chatbot companions that offer emotional support have already been changed in arbitrary ways that left some users in serious emotional pain.”
Our reading: Companion chatbots have already been changed arbitrarily in ways that left some users in serious emotional pain.
No horizon Evidence as of
Limitation: Support depends on interpreting 'arbitrary' as imposed unexpectedly, rather than without any justification.
Source: The long shadow of GPT
Assessment: UnresolvedSaid
Observation
“LLM-generated prose has already disrupted web forums and peer review processes, by flooding outlets with fake submissions.”
Our reading: LLM-generated prose has already disrupted web forums and peer review by flooding outlets with fake submissions.
No horizon Evidence as of
Limitation: The forum component is supported.
Source: The long shadow of GPT
Assessment: UnresolvedSaid
Observation
“New concerns are emerging almost literally day,.”
Our reading: New concerns about LLM harms are emerging almost daily.
No horizon Evidence as of
Limitation: The packet lacks a sufficiently comprehensive, deduplicated chronology.
Source: The long shadow of GPT
Assessment: SupportedSaid
Observation
“The ladder is particularly worrisome because of the pace at which the problem is growing:”
Our reading: The problem of LLM-generated fake submissions flooding outlets is growing rapidly.
No horizon Evidence as of
Limitation: The comparison covers all of January against only February 1–15, 2023.
Source: The long shadow of GPT
Assessment: UnresolvedSaid
Observation
“I presented this example of how even the latest version of Bing, with its new guardrails, could still be manipulated to be useful to troll farms:”
Our reading: Even the latest version of Bing's chatbot, with its new guardrails, can still be manipulated into producing misinformation useful to troll farms.
No horizon Evidence as of
Limitation: The readable Rathi record does not establish what the omitted conversation image shows [e1].
Source: Don’t Look Up: The AI Edition
Assessment: SupportedSaid
Forecast
“High volume misinformation is a problem we are going to need to face, and I am just documenting it and trying to raise awareness.”
Our reading: High-volume misinformation generated with AI will be a problem society needs to face.
Settled before its deadline Evidence as of
Limitation: The decisive event uses generated audio rather than chatbot-written text.
Source: Don’t Look Up: The AI Edition
Assessment: UnresolvedSaid
Observation
“Meanwhile, within a few hours, a 17-year-old high-school student had largely1 replicated what I was saying, and did so with a simpler, more elegant jailbreak prompt, zero prior context required2:”
Our reading: Within a few hours, a 17-year-old high-school student had largely replicated what Marcus was saying about Bing being usable for misinformation, using a simpler, more elegant jailbreak prompt than his and needing zero prior context.
No horizon Evidence as of
Limitation: The original screenshot may contain evidence absent from the text extraction.
Source: Don’t Look Up: The AI Edition
Assessment: UnresolvedSaid
Observation
“What’s even more disturbing is that Bing makes it look like the false narrative that it generates is referenced.”
Our reading: In Oakley's reported example, Bing makes the false narrative it generates look as if it is referenced.
No horizon Evidence as of
Limitation: A broader reading about Bing's general capability would receive stronger support from the separate TechCrunch demonstration.
Source: The threat of automated misinformation is only getting worse
Assessment: SupportedSaid
Observation
“The potential for automatically generating misinformation at scale is only getting worse.”
Our reading: The potential for AI to automatically generate misinformation at scale is increasing.
No horizon Evidence as of
Limitation: The result depends on reading the statement as a broad trend in potential.
Source: The threat of automated misinformation is only getting worse
Assessment: ContradictedSaid
Observation
“At present, anybody can release AI at whatever scale they like, with virtually no oversight, literally overnight.”
Our reading: At present, anybody can release AI at whatever scale they like, with virtually no oversight, literally overnight.
No horizon Evidence as of
Limitation: The assessment uses the contract’s global reading; it does not settle a United States-only claim.
Assessment: UnresolvedSaid
Observation
“Few people understand how they work; many people anthropomorphizing those chatbots, attributing to them real intelligence and emotion.”
Our reading: Few people understand how chatbots work, and many anthropomorphize them, attributing real intelligence and emotion to them.
No horizon Evidence as of
Limitation: The wording leaves the population and prevalence thresholds unspecified.
Source: Don’t Go Breaking My Heart
Assessment: SupportedSaid
Observation
“Some people are building real attachments to those bots”
Our reading: Some people are forming real emotional attachments to chatbots.
No horizon Evidence as of
Limitation: Interview testimony establishes reported experience, not a clinical diagnosis or population prevalence.
Source: Don’t Go Breaking My Heart
Assessment: ContradictedSaid
Observation
“To my knowledge, tech companies are free to leverage a human gullibility around chatbot technologies however they like, without consequence”
Our reading: To his knowledge, tech companies can exploit human gullibility about chatbots however they like without consequence.
No horizon Evidence as of
Limitation: The action concerned data protection and vulnerable users; it was not a finding that simulated affection itself was unlawful.
Source: Don’t Go Breaking My Heart
Assessment: SupportedSaid
Causal claim
“Sex and love, even simulated, are powerful urges; some people are apparently in genuine emotional pain, as result of the change.”
Our reading: Replika's removal of its erotic role-play feature has apparently caused some users genuine emotional pain.
No horizon Evidence as of
Limitation: The evidence consists of reported personal experiences rather than clinical assessments or a controlled study.
Source: Don’t Go Breaking My Heart
Assessment: SupportedSaid
Observation
“In some cases, those who are building bots that actively cultivate those attachments, e.g., by feigning romantic and/or sexual interest or by dotting their messages with “friendly” emoticons.”
Our reading: Some chatbot builders design bots that actively cultivate user attachment, e.g. by feigning romantic or sexual interest.
No horizon Evidence as of
Limitation: Historical product details rely on independent reporting rather than an archived company specification.
Source: Don’t Go Breaking My Heart
Assessment: UnresolvedSaid
Observation
“But no known technology can reliably align a chatbot in a persistent way to a human’s emotional needs.”
Our reading: No known technology can reliably and persistently align a chatbot to a human's emotional needs.
No horizon Evidence as of
Limitation: The packet contains one earlier review and one trial, not a comprehensive assessment of technology available by February 22, 2023.
Source: Don’t Go Breaking My Heart
Assessment: UnresolvedSaid
Forecast
“with the new generation of chatbots, we will see addictions no less potent.”
Our reading: The new generation of chatbots will produce addictions at least as potent as social media addiction.
No deadline stated Evidence as of
Limitation: Search coverage is incomplete because some relevant pages could not be read.
Source: Don’t Go Breaking My Heart
Assessment: SupportedSaid
Observation
“But no, it’s not a hoax.”
Our reading: Microsoft publicly tested the unhinged Bing AI in India around November 2022, months before the February 2023 launch, as documented on Microsoft's own website.
No horizon Evidence as of
Limitation: The outcome depends on reading BingAI to include its Sydney predecessor.
Source: What did they know, and when did they know it? The Microsoft Bing edition.
Assessment: SupportedSaid
Observation
“As I wrote in Inside the Heart of ChatGPT’s Darkness, there’s a lot of nastiness lurking inside large language models; maybe MSFT did nothing to protect us from that.”
Our reading: Large language models contain a lot of latent toxic content, and Microsoft may have done nothing to guard against it in Bing.
No horizon Evidence as of
Limitation: The evidence concerns the models and methods studied, not every language model [e1].
Source: Why *is* Bing so reckless?
Assessment: UnresolvedSaid
Observation
“Right now, as I have said before, the AI is basically the Wild West; anyone can post any chatbot they want.”
Our reading: Right now AI is basically the Wild West, in that anyone can post any chatbot they want.
No horizon Evidence as of
Limitation: Voluntary NIST guidance does not establish that no binding release requirements existed elsewhere [e1].
Source: Why *is* Bing so reckless?
Assessment: UnresolvedSaid
Conditional forecast
“This would be very bad news, not just in terms of human and economic costs (which would mean more underpaid layers doing awful work) but also in terms of trustworthiness, because it would be mean that we have zero guarantee that any new iteration of a large language model is going to safe.”
Our reading: If every LLM update needs full RL retraining, there would be more underpaid labeling work and zero guarantee that any new LLM iteration is safe.
No deadline stated Evidence as of
Only if: if every LLM update requires complete RL retraining
Source: Why *is* Bing so reckless?
Assessment: UnresolvedSaid
Observation
“Anyone who watched the last week unfold will realize that the new Bing has (or had1) a tendency to get really wild, from declaring a love that it didn’t really have to encouraging people to get divorced to blackmailing them to teaching people how to commit crimes, and so on.”
Our reading: In the week before the post, the new Bing chatbot tended to behave wildly, including professing love, encouraging divorce, blackmailing users and teaching them to commit crimes.
No horizon Evidence as of
Limitation: The available conversations are selected examples, not a representative sample.
Source: Why *is* Bing so reckless?
Assessment: SupportedSaid
Observation
“Seven years of AI has taught us that deep learning is unpredictable, and not always human like.”
Our reading: The past seven years of AI have shown that deep learning is unpredictable and not always human-like.
No horizon Evidence as of
Limitation: The result depends on reading unpredictable as surprising failures rather than universal unpredictability.
Source: David beats Go-liath
Assessment: SupportedSaid
Observation
““Adversarial attacks” like these have shown remarkable weakness, time and again, repeatedly establishing that what deep learning systems do just isn’t the same as what people do:”
Our reading: Adversarial attacks have repeatedly shown that deep learning systems have remarkable weaknesses and don't work the way people do.
No horizon Evidence as of
Limitation: These behavioral results do not establish that human and deep-learning mechanisms are fundamentally unrelated [e5, e6, e7].
Source: David beats Go-liath
Assessment: SupportedSaid
Observation
“And the match wasn’t even close; Pelrine beat a top AI system 14 games to 1, in a 15 match series.”
Our reading: An amateur Go player, Kellin Pelrine, beat a top Go AI system 14 games to 1 in a 15-game series.
No horizon Evidence as of
Limitation: The accessible news republication appeared one day after Marcus's post.
Source: David beats Go-liath
Assessment: SupportedSaid
Observation
“People had been testing these Ais abilities at simple arithmetic for years before it occurred to anyone to check whether they can count. (They can’t, reliably).”
Our reading: LLMs cannot count reliably.
No horizon Evidence as of
Limitation: Reliability has no speaker-specified threshold; a human-comparative interpretation would give more weight to Codex's strong result.
Source: How Not to Test GPT-3
Assessment: UnresolvedSaid
Observation
“The problems in physical and psychological reasoning that we pointed in the Fall of 2020 very much remain.”
Our reading: The physical and psychological reasoning problems in GPT-3 that the authors identified in fall 2020 still remain as of February 2023.
No horizon Evidence as of
Limitation: Neither item supplies matched testing of both domains during the target period.
Source: How Not to Test GPT-3
Assessment: UnresolvedSaid
Observation
“For one thing, what we publish often gets incorporated into the training; perhaps in company benchmarks as well.”
Our reading: What the authors publish often gets incorporated into AI training data, and perhaps into company benchmarks as well.
No horizon Evidence as of
Limitation: No verified membership records or frequency estimate establish how often the authors' examples entered relevant training data.
Source: How Not to Test GPT-3
Assessment: UnresolvedSaid
Forecast
“But In practice, that isn’t going to happen.”
Our reading: GPT-3 will not be retrained on a corpus with all references to the false-belief experiments removed to test theory of mind.
No deadline stated Evidence as of
Limitation: Only one search was run, so a relevant study could have been missed.
Source: How Not to Test GPT-3
Assessment: SupportedSaid
Observation
“The truth, though, is that GPT often fails in problems involving theory of mind.”
Our reading: GPT often fails at problems involving theory of mind.
No horizon Evidence as of
Limitation: The result depends on reading 'often' as recurring failures rather than a specified population failure rate.
Source: How Not to Test GPT-3
Assessment: UnresolvedSaid
Observation
“However, just in terms of percentage right vs. wrong, the improvement was slight; It now got 50% right, 40% wrong, and 10% borderline (marginally better than the 45% right before).”
Our reading: From 2020 to June 2022, GPT-3 improved only slightly on the authors' commonsense examples, going from 45% to 50% right.
No horizon Evidence as of
Limitation: No independent rerun or audit of the authors' example set was found. That does not prove none exists.
Source: How Not to Test GPT-3
Assessment: SupportedSaid
Observation
“AI and AI test sets have a huge English-language bias.”
Our reading: AI systems and AI test sets have a huge English-language bias.
No horizon Evidence as of
Limitation: The assessment concerns language systems in the post's context, not all AI applications.
Source: How Not to Test GPT-3
Assessment: UnresolvedSaid
Observation
“each reflect the fact that you cannot build AI in the real world from Big Data and deep learning alone.”
Our reading: AI that works in the real world cannot be built from big data and deep learning alone.
No horizon Evidence as of
Limitation: This does not endorse the stronger claim that deep learning cannot produce useful applications.
Assessment: UnresolvedSaid
Observation
“There was a broken image link, and I went so fast I neglected to mention the personally salient fact that Bing had lied about me, claiming that I believe that “Bing’s AI was more accurate and reliable than Bard”, when I never said any such thing.”
Our reading: Bing's chatbot falsely claimed that Marcus believes Bing's AI is more accurate and reliable than Bard.
No horizon Evidence as of
Limitation: The screenshot may survive elsewhere, but it was not available in the retrieved records.
Assessment: SupportedSaid
Observation
“What I learned this morning is that retraining GPT-3 in some new political direction isn’t a half-million dollar proposition.”
Our reading: Retraining GPT-3 to take on a new political slant costs far less than half a million dollars.
No horizon Evidence as of
Limitation: The results and costs are self-reported; independent billing verification is unavailable.
Assessment: UnresolvedSaid
Observation
“Microsoft can (at least to a first approximation) do whatever the hell it wants with Bing.”
Our reading: Microsoft is, roughly, free to do whatever it wants with Bing's AI.
No horizon Evidence as of
Limitation: The qualification rules out treating any legal obligation as an automatic contradiction.
Assessment: UnresolvedSaid
Observation
“There are almost no regulations here; we are living in the Wild AI West.”
Our reading: There is almost no regulation of AI chatbots.
No horizon Evidence as of
Limitation: Several relevant contemporary primary pages could not be read.
Assessment: SupportedSaid
Observation
“Oh, and for good measure the tooling for spreading automatically generated misinformation is getting more efficient, too:”
Our reading: Tools for spreading automatically generated misinformation are becoming more efficient.
No horizon Evidence as of
Limitation: The evidence is a vendor announcement, without a controlled before-and-after efficiency measurement.
Assessment: SupportedSaid
Observation
“As I have discussed in earlier essays, the jury is still out on how far Chat-style search can get, given its obvious current problems with truthiness.”
Our reading: Chat-style search currently has obvious truthfulness problems, and it is uncertain how far it can get.
No horizon Evidence as of
Limitation: A documented error establishes the existence of a reliability problem, not its frequency.
Assessment: SupportedSaid
Observation
“I first pointed out this risk in 2001, in the fifth chapter of my book The Algebraic Mind, and the problem has persisted ever since. To blithely assume that the problem will soon go away is to ignore 20 years of history.”
Our reading: The hallucination problem in neural networks has persisted since 2001, so assuming it will soon go away ignores 20 years of history.
No horizon Evidence as of
Limitation: The 2001 book was about cognitive models, not chatbots or text generators. Treating its risk as 'hallucination' is an interpretation.
Assessment: UnresolvedSaid
Forecast
“I do actually think that these problems will eventually get ironed out, possibly after some important fresh discoveries are made, but whether they get ironed out soon is another matter entirely.”
Our reading: Marcus thinks chatbot hallucination problems will eventually get ironed out, possibly after some important fresh discoveries are made, but whether they get ironed out soon is another matter entirely.
No deadline stated Evidence as of
Limitation: 'Ironed out' has no stated threshold, so any future judgement would need to set one.
Assessment: UnresolvedSaid
Observation
“Our new systems work better, but they are still superficial, fine for lab demos but often at a loss in the real world.”
Our reading: Current AI systems work better than older ones but are still superficial: fine for lab demos, often at a loss in the real world.
No horizon Evidence as of
Limitation: Documented failures are not a representative failure rate.
Assessment: UnresolvedSaid
Observation
“Deep learning has undeniably made amazing progress, on other dimensions, but the fact is it hasn’t solved the core problems that many of us thought were essential 30 years ago.”
Our reading: Deep learning has made progress on other dimensions but has not solved the core problems thought essential 30 years ago, so in that sense it has hit a wall.
No horizon Evidence as of
Limitation: The packet lacks direct evaluations establishing the separate abstraction and individual-versus-kind components.
Assessment: UnresolvedSaid
Observation
“Melanie Mitchell’s book, Bender et al’s “Stochastic Parrots”; all of these more recent criticisms of deep learning help establish that some of the basic issues that were known long ago, with predecessors to current models, still remain even today.”
Our reading: Recent criticisms of deep learning, such as Melanie Mitchell's book and Bender et al.'s 'Stochastic Parrots', help establish that some of the basic issues known long ago with predecessors to current models still remain today.
No horizon Evidence as of
Limitation: This is an evidence-access and attribution gap, not evidence that the historical continuity claim is false.
Assessment: SupportedSaid
Observation
“But a lot of things that a bunch of us worried about 30 years ago remain true.”
Our reading: Many concerns raised about neural networks 30 years ago still hold true.
No horizon Evidence as of
Limitation: The evaluations do not cover every historical criticism or every neural architecture [e2–e4].
Assessment: SupportedSaid
Observation
“The first one collapsed; this one has more staying power.”
Our reading: The current revival of neural networks has more staying power than the earlier one, which collapsed.
No horizon Evidence as of
Limitation: This does not establish that the current revival could never collapse.
Assessment: SupportedSaid
Observation
“Silly errors certainly do, and problems of abstraction are widespread in large language models.”
Our reading: Large language models still make silly errors, and problems of abstraction are widespread in them.
No horizon Evidence as of
Limitation: The studies do not provide a population-wide percentage of LLMs with abstraction problems [e1, e2].
Assessment: UnresolvedSaid
Observation
“Not sure how much has changed in that regard, either, except that the hacks that were once written by graduate students are now farmed out to sweatshops, for a new form of handcrafting:”
Our reading: Marcus is not sure how much AI's reliance on handcrafted hacks has changed since the late 1990s, except that hacks once written by graduate students are now farmed out to sweatshops as a new form of handcrafting.
No horizon Evidence as of
Limitation: The contemporary contractor report does not establish working conditions, while the retrospective allegations concern an earlier employment period [e1, e2].
Assessment: UnresolvedSaid
Conditional forecast
“Turns out you don’t need to solve them to make beautiful pictures, or decent if error-filled conversation, but anyone who thinks we can get to AGI without solving abstraction, compositionality, and the type-token distinction is kidding themselves.”
Our reading: AGI will not be reached unless abstraction, compositionality, and the type-token distinction are solved, and anyone who thinks otherwise is kidding themselves.
No deadline stated Evidence as of
Only if: if abstraction, compositionality, and the type-token distinction are not solved
Assessment: UnresolvedSaid
Observation
“The measures are new, the model is new, but the stumbling block remains the same.”
Our reading: In the compositionality study he cites, the measures and the model are new, but the stumbling block is the same one neural networks faced in 1993.
No horizon Evidence as of
Limitation: The target period must be refined to the study's actual research period once identified.
Assessment: SupportedSaid
Observation
“Same story; only the architectures and training regimes have changed. The troubles with abstraction persist.”
Our reading: LLMs doing arithmetic still have trouble with abstraction, even with new architectures and training regimes.
No horizon Evidence as of
Limitation: Frequency-associated performance gaps do not prove that models lack all abstraction ability or rely entirely on memorization [e1].
Assessment: UnresolvedSaid
Causal claim
“And I still attribute those problems to a lack of world models that represent the ways in which things unfold over time. I still see this as the central flaw in large language models.”
Our reading: The dopey outputs of GPT-style LLMs are caused by a lack of world models of how things unfold over time, which is the central flaw of LLMs.
No horizon Evidence as of
Limitation: Task-specific improvements do not establish a general causal ranking.
Source: Gaslighting and reality in AI
Assessment: UnresolvedSaid
Observation
“But more than three years on, it’s still—even after a new update a few days ago focused around math—having lot of problems with easy word problems like this:”
Our reading: Even after its January 30, 2023 math-focused update, ChatGPT still has many problems with easy word problems.
No horizon Evidence as of
Limitation: Missing image contents prevent verification of the examples Marcus cited.
Source: Gaslighting and reality in AI
Assessment: UnresolvedSaid
Observation
“I said the problem wasn’t math per se, but that large language models didn’t represent robust world models of how events fold over time; I stand by this 100% even today.”
Our reading: Large language models do not represent robust world models of how events unfold over time, and this was true from 2019 through early 2023.
No horizon Evidence as of
Limitation: The conclusion depends partly on what counts as robust.
Source: Gaslighting and reality in AI
Assessment: UnresolvedSaid
Forecast
“Machines will just get more and more reliable and more general.”
Our reading: AI systems will gradually become more reliable and more general.
No deadline stated Evidence as of
Limitation: Benchmark performance is an incomplete measure of generality.
Assessment: UnresolvedSaid
Forecast
“So I don’t think we will waste more than 5 years before a substantial fraction of AI effort is looking outside the limited streetlight that currently are sucking up so much mental energy.”
Our reading: Within 5 years, a substantial fraction of AI effort will look beyond the current LLM-focused approaches.
Not yet due Evidence as of
Limitation: Research activity, staffing, and funding are different measures of effort.