“Hassabis is no longer on board either, and nor are Sutskever and LeCun. Nadella and Pichai have also hinted at skepticism around scaling.”
Our reading, the sentence that was checked:
Hassabis, Sutskever and LeCun no longer believe in scaling as the route to AGI, and Nadella and Pichai have hinted at skepticism about it.
Assessment: Mixed Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, BREAKING: Sam Altman concedes that we need major breakthroughs beyond mere scaling to get to AGI, paragraph 7.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-breaking-sam-altman-concedes-that-905284-a2
We recorded 6 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 2026-03-16. 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
- Hassabis is no longer on board either
- nor are Sutskever and LeCun
- Nadella and Pichai have also hinted at skepticism around scaling
- insiders are losing faith in pure scaling
Our interpretation
- The ordinary reading is about scaling existing AI methods alone being sufficient for AGI, not whether larger models can keep improving.
- The first sentence attributes a rejection of that sufficient-route thesis to Hassabis, Sutskever and LeCun. The second requires only expressions of doubt from Nadella and Pichai.
- Assess the public positions established before publication on 2026-03-16. Public statements have no regular reporting period, so the target covers the preceding public record through that date, with no invented starting date.
- Treat statements that new research or breakthroughs are needed as evidence against scaling alone. Continued enthusiasm for scaling is compatible with that position.
- The surrounding account of changing views also suggests a transition. An established skeptical position supports the stated position but does not establish a recent conversion.
Counts as supported if: Public statements before publication establish that each of Hassabis, Sutskever and LeCun rejected scaling existing methods alone as sufficient for AGI, and that Nadella and Pichai expressed at least some doubt about its sufficiency. A claimed change of position additionally needs evidence of the earlier position.
Counts as contradicted if: A named leader's public statements materially conflict with the position attributed to him. In particular, explicitly retaining scaling alone as a live route to AGI contradicts an unqualified attribution of rejection. Continued belief that scaling improves models does not by itself contradict skepticism about sufficiency for AGI.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
Outcome: Mixed
Part of the claim holds and part does not, and the record says which part is which.
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
In November 2025, Sutskever questioned whether increasing scale 100-fold would transform everything and called for renewed research, supporting skepticism about scaling alone [e2]. LeCun explicitly rejected scaling the current paradigm as a route to human-level intelligence in remarks reported in January 2026 [e3]. Nadella's November 2025 remarks and Pichai's December 11, 2024 interview described a need for breakthroughs, supporting the weaker attribution of doubt about scaling alone [e4, e5]. The Hassabis attribution does not hold: December 2025 reporting quoted him saying scaled current systems could be the entire AGI system, while also reporting his expectation of additional breakthroughs [e1]. LeCun had already rejected scaling existing techniques to human-level AI in 2023, so his cited remarks do not establish a recent change of position [e6].
Limitations
- The record does not establish earlier endorsement and subsequent rejection of scaling alone for everyone described as having changed position.
- Public statements establish expressed positions, not private beliefs.
- The Hassabis and January 2026 LeCun evidence relies on reporting rather than retrieved original transcripts.
- The evidence does not establish whether scaling alone can achieve AGI.
Evidence (6 items)
Supports the claim: Dwarkesh Patel,
back to the age of research
Why it matters here: In this original interview, Sutskever disputes the idea that multiplying current scale by 100 would transform everything and calls for renewed research. This supports Marcus's characterization of his departure from reliance on scaling alone, without implying that scaling produces no improvements.
- Publisher
- Dwarkesh Patel
- Kind of source
- News report
- Describes events of
- Public from
Supports the claim: The Indian Express,
simply making these systems bigger or better
Why it matters here: The report directly quotes LeCun rejecting enlargement of existing systems as sufficient for human-level intelligence. Reputable reporting is used because the attempted institutional primary sources did not yield a usable passage.
- Publisher
- The Indian Express
- Kind of source
- News report
- Describes events of
- Public from
Supports the claim: Dwarkesh Patel,
real science breakthroughs
Why it matters here: Nadella's stated need for scientific breakthroughs supports the weaker attribution of doubt about scaling alone. His preceding remarks explicitly endorse scaling's continued effectiveness, so this does not establish broad skepticism about scaling.
- Publisher
- Dwarkesh Patel
- Kind of source
- News report
- Describes events of
- Public from
Supports the claim: Semafor,
it’s achieving breakthroughs
Why it matters here: Pichai explicitly distinguishes breakthroughs from scaling alone, meeting the threshold of hinted skepticism. His answer remains optimistic about progress and does not declare AGI impossible through current approaches.
- Publisher
- Semafor
- Kind of source
- News report
- Describes events of
- Public from
Goes against the claim: AOL, Lakshmi Varanasi,
It could be the entirety of the AGI system.
Why it matters here: Hassabis explicitly retained scaled current systems as a possible complete route to AGI, contradicting an unqualified attribution of rejection. The report also says he suspected additional breakthroughs would be needed, so his position includes genuine doubt.
- Publisher
- AOL, Lakshmi Varanasi
- Kind of source
- News report
- Describes events of
- Public from
Context only: WIRED,
We're missing something big
Why it matters here: In this original interview, LeCun rejects obtaining human-level AI by scaling existing techniques. It establishes that his skeptical position was already public in 2023, limiting the post's suggestion of a recent change.
- Publisher
- WIRED
- 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, BREAKING: Sam Altman concedes that we need major breakthroughs beyond mere scaling to get to AGI
Hassabis is no longer on board either, and nor are Sutskever and LeCun. Nadella and Pichai have also hinted at skepticism around scaling.
- Where
- paragraph 7
- Published
- 2026-03-16
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-breaking-sam-altman-concedes-that-905284-a2: Current. Outcome: Mixed. 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-sam-altman-concedes-that-905284-a2.