“The machine learning community, which has unwelcomed me since I started vocally critiquing the scaling approach, has finally started to come around.”
Our reading, the sentence that was checked:
The machine learning community has started to accept critiques of the scaling approach.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, “Scale Is All You Need” is dead, paragraph 12.
- 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-8a7e73-a1
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
- the machine learning community
- has finally started to come around
- to the critique of the scaling approach
Our interpretation
- Plain reading: by late 2025, prominent ML researchers and venues were publicly voicing the critique that scaling LLMs is not enough (for reliability, generalization, or AGI). They need not credit Marcus personally.
- Target period: 2025 up to the post date. The claim describes a current shift with no stated period.
- The aside that the community 'unwelcomed' Marcus is biographical and not graded.
Counts as supported if: Prominent ML figures, major conferences, or community surveys in 2025 voice the view that scaling alone has key limits and that new approaches are needed.
Counts as contradicted if: Evidence that the ML community in 2025 broadly rejected critiques of scaling, with no notable figures or venues embracing them.
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
'Started to come around' needs prominent ML voices and venues in 2025 to be voicing the scaling critique. They were. A write-up of NeurIPS 2025 said the community's 'center of gravity is moving away from “just scale it”'. It also said 'What got us here won’t get us there' had become a common refrain about LLMs. In November 2025 Ilya Sutskever said 100x more scale would not transform things. The same NeurIPS write-up called the field 'openly polarized', and some top papers still backed scaling. But 'started to' does not require consensus. The evidence shows the shift, though not that anyone credited Marcus.
Limitations
- Rests on one industry write-up plus individual statements. No measured trend.
- 'Come around' is read as accepting limits to scaling, not endorsing Marcus specifically.
Evidence (4 items)
Supports the claim: Lambda (company blog),
In 2025, that belief is shifting. The center of gravity is moving away from “just scale it” and toward something more constraint-aware and capability-driven:
Why it matters here: An industry write-up of NeurIPS 2025 says the community is moving away from 'just scale it'. Published a week after the post, but it describes the conference held before it.
- Where
- opening paragraph
- Publisher
- Lambda (company blog)
- Kind of source
- News report
- Describes events of
- Public from
Supports the claim: Lambda (company blog),
“What got us here won’t get us there” has become a common refrain for LLMs. Models trained primarily to mimic language are inefficient and fundamentally limited
Why it matters here: Describes the critique of LLM limits as a common refrain in the community at NeurIPS 2025.
- Publisher
- Lambda (company blog)
- 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 scaling-era researcher publicly accepted a central point of the scaling critique.
- Publisher
- Dwarkesh Podcast
- Kind of source
- News report
- Describes events of
- Public from
Context only: Lambda (company blog),
What’s interesting is how openly polarized the field is now.
Why it matters here: Counter-context. The field was polarized, not unified, and some award-winning work still backed scaling along the right dimensions.
- Publisher
- Lambda (company blog)
- 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, “Scale Is All You Need” is dead
The machine learning community, which has unwelcomed me since I started vocally critiquing the scaling approach, has finally started to come around.
- Where
- paragraph 12
- Published
- 2025-12-08
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-breaking-news-scale-is-all-you-need-8a7e73-a1: 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-8a7e73-a1.