“And now, at last, the field is starting to realize it.”
Our reading, the sentence that was checked:
The AI field is now starting to realize that scaling alone will not reach AGI.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, “Scale Is All You Need” is dead, paragraph 7.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-breaking-news-scale-is-all-you-need-ca56f9-a2
We recorded 9 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 2025-12-08. 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)
From the speaker
- now, at last
- the field
- is starting to realize it
- it = scaling would not get us to AGI
Our interpretation
- Plain reading: by late 2025, a growing share of AI researchers, including prominent figures, publicly said that scaling current LLM approaches alone would not reach AGI.
- Target period: 2025 up to the post date, the most recent period public evidence covers. 'Starting to' needs only a visible and growing trend, not consensus.
Counts as supported if: Surveys or prominent researcher statements from 2025 show substantial or growing acceptance that scaling alone will not reach AGI.
Counts as contradicted if: Evidence that in 2025 most researchers and leading figures still held that scaling alone would reach AGI, with no visible shift.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
Outcome: Supported
The dated evidence matches 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
The claim says the field is 'starting to' realize that scaling alone will not reach AGI. This needs a visible, growing view, not a consensus. Evidence from 2025 shows such a view. On 14 March 2025, The Decoder reported an AAAI survey in which 76% of respondents said scaling existing AI systems was 'unlikely' or 'very unlikely' to produce AGI (e1). On 25 November 2025, Ilya Sutskever said he did not think 100x more scale would transform everything, and that 'it's back to the age of research again' (e2). On 20 November 2025, NeurIPS announced that Turing Award winner Richard Sutton would give an invited talk on 3 December arguing that AI 'has lost its way' and needs agents that learn continually (e3). On 18 November 2025, Oriol Vinyals of Google DeepMind disagreed, saying Gemini 3 showed 'no walls in sight' (e4). In the same post, he described the view that scaling is over as a 'popular belief' (e4). So views were not unanimous, but by the claim's December 2025 date, the view it describes was widely voiced.
Limitations
- No earlier survey is available for comparison, so the growth of this view is inferred rather than measured.
- The AAAI survey sample may not reflect views inside frontier labs.
- The NeurIPS item (e3) is a pre-conference announcement of Sutton's talk. It does not record what he said on the day.
- The evidence covers views about scaling. It does not show whether scaling can or cannot reach AGI.
Evidence (4 items)
Supports the claim: The Decoder,
The study found that 76 percent of respondents rated this possibility as "unlikely" or "very unlikely."
Why it matters here: A survey of 475 AAAI community members found that 76% thought scaling current systems was unlikely to produce AGI.
- Where
- paragraph 2
- Publisher
- The Decoder
- Kind of source
- News report
- Describes events of
- Public from
Supports the claim: Dwarkesh Podcast,
But is the belief that if you just 100x the scale, everything would be transformed? I don’t think that’s true. So it’s back to the age of research again, just with big computers.
Why it matters here: A leading figure of the scaling era publicly said scaling alone would not transform things. That is a clear sign of the shift Marcus describes.
- Publisher
- Dwarkesh Podcast
- Kind of source
- News report
- Describes events of
- Public from
Supports the claim: NeurIPS,
Richard Sutton, recipient of the 2024 Turing Award, will argue that as “AI has become a huge industry, to an extent it has lost its way,” and proposes a return to fundamental principles
Why it matters here: The field's biggest conference gave a keynote slot to a Turing Award winner arguing that the field has lost its way and needs new approaches.
- Where
- paragraph 2
- Publisher
- NeurIPS
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
Context only: Oriol Vinyals (Google DeepMind) on X, via Thread Reader,
Contra the popular belief that scaling is over—which we discussed in our NeurIPS '25 talk with @ilyasut and @quocleix—the team delivered a drastic jump.
Why it matters here: Shows that some frontier-lab leaders still pushed back on limits to scaling. It also confirms that the 'scaling is over' view had become a 'popular belief'.
- Where
- tweet text
- Publisher
- Oriol Vinyals (Google DeepMind) on X, via Thread Reader
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, “Scale Is All You Need” is dead
And now, at last, the field is starting to realize it.
- Where
- paragraph 7
- Published
- 2025-12-08
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-breaking-news-scale-is-all-you-need-ca56f9-a2: Current. Outcome: Supported. 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-breaking-news-scale-is-all-you-need-ca56f9-a2.