“He only fully changed his mind a few weeks later, after Galactica tanked and ChatGPT ate his lunch.”
Our reading, the sentence that was checked:
LeCun only fully changed his mind about LLMs a few weeks after November 2022, after Meta's Galactica tanked and ChatGPT outcompeted Meta.
Assessment: Contradicted Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, The False Glorification of Yann LeCun, paragraph 18.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-the-false-glorification-of-yann-lecun-024a9f-a1
We recorded 13 claims from this post. See every claim from this post.
The wording above is the version this site observed on 2026-10-28. The source may have been edited after 2025-11-18. This page does not claim the wording was captured then.
What the claim commits to
- Kind
- Observation
- Deadline, as the speaker wrote it
- None stated. This site does not add one.
- Deadline state when assessed
- No horizon
- Deadline state on 2026-10-29
- No horizon (the release’s reference date)
Conditions
- a few weeks after his November 2022 praise of LLMs (qualifier, stated by the speaker)
From the speaker
- only fully changed his mind
- a few weeks later
- after Galactica tanked
- ChatGPT ate his lunch
Our interpretation
- In context, the relevant belief is whether current LLMs can reach human-level AI, rather than whether they can be useful or impressive.
- The historical comparison covers the context's 2019 critiques through the alleged late-2022 reversal. December 31, 2022 is an approximate outer boundary for a few weeks after November.
- An explicit rejection before November 2022 contradicts the proposed timing. Praise for a model does not establish belief that it can reach AGI.
- After is assessed as chronology. A causal claim about competitive disappointment would require additional evidence.
Counts as supported if: Contemporaneous statements establish a substantive move from accepting LLMs as a sufficient route to human-level AI to rejecting that route only in the weeks after November 2022.
Counts as contradicted if: A clear public rejection of current LLMs as a sufficient route to human-level AI predates the alleged late-2022 change.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
Outcome: Contradicted
The dated evidence goes against what the claim committed to, under the criteria frozen before grading.
Scored on a plain reading. The speaker named no measure for this claim, so it was judged on the reading an informed reader would take, stated below under our interpretation. Both reviews had to find that reading fair.
- Evidence as of
- Deadline state when assessed
- No horizon
- Assessment recorded
- Policy version
- 2026-10-28.7
Rationale
LeCun publicly rejected scaling current LLMs as a route to human-level AI in June 2022. A September interview also described them as insufficient. Both predate the claimed change after November. Later praise for a particular model would not establish a reversal on this separate question.
Limitations
- The finding concerns the timing of his publicly stated position, not his private state of mind.
- The evidence comes from outlets' original interviews because the underlying research-paper text was not retrievable.
- A narrower claim that his criticism became more frequent or forceful would require a separate assessment.
Evidence (2 items)
Goes against the claim: MIT Technology Review,
no direct experience of the world
Why it matters here: The same paragraph explicitly records LeCun rejecting the proposition that scaling current large language models will produce human-level AI. His rejection was public months before the claimed late-2022 reversal; this is original interview reporting.
- Publisher
- MIT Technology Review
- Kind of source
- News report
- Describes events of
- Public from
Goes against the claim: ZDNET,
necessary but not sufficient
Why it matters here: The interview identifies large language models among the approaches LeCun regarded as missing essential components of intelligence. It independently confirms that this position preceded November 2022.
- Where
- Introduction, immediately before the paragraph identifying GPT-3 and other large language models
- Publisher
- ZDNET
- Kind of source
- News report
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, The False Glorification of Yann LeCun
He only fully changed his mind a few weeks later, after Galactica tanked and ChatGPT ate his lunch.
- Where
- paragraph 18
- Published
- 2025-11-18
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-the-false-glorification-of-yann-lecun-024a9f-a1: Current. Outcome: Contradicted. Evidence as of 2026-10-28. Public date not recorded.
Recorded events
Assessment publishedBy this site
First assessment published after both reviews and the final-text check passed.
About assessment marcus-the-false-glorification-of-yann-lecun-024a9f-a1.