“Scaling hasn’t gotten them —or anyone else — to AGI.”
Our reading, the sentence that was checked:
Scaling has not brought OpenAI or anyone else to AGI.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, OpenAI, in deep trouble, paragraph 13.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-openai-in-deep-trouble-52753c-a1
We recorded 18 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 2025-03-01. 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
- Scaling hasn’t gotten
- them —or anyone else
- to AGI
Our interpretation
- The plain reading of AGI is broadly human-level general intelligence, including handling unfamiliar tasks across domains.
- This is a cumulative claim about achievement through March 1, 2025, not a prediction that scaling can never produce AGI.
- I assess publicly demonstrated capabilities. No starting date is supplied for the cumulative observation.
- Scaling includes increasing training resources and inference-time computation; the claim does not distinguish them.
Counts as supported if: Independent assessments of the contemporary field and leading scaled systems identify substantial gaps from broadly human-level general intelligence, with no independently established counterexample by the post date.
Counts as contradicted if: A system developed through scaling had independently demonstrated broadly human-level general intelligence by March 1, 2025.
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
Under the ordinary broad human-level reading, the public evidence supports the observation. ARC Prize explicitly rejected treating o3's breakthrough as AGI, and the international scientific review documented substantial limitations in generalization and independent work. Microsoft researchers had used AGI language for GPT-4, but explicitly qualified it as early and incomplete. This conclusion concerns achievement by the post date, not whether scaling could eventually succeed.
Limitations
- The result depends on the stated meaning of AGI.
- This is a judgment about demonstrated capabilities, not proof concerning every undisclosed system.
- The evidence does not establish that scaling is incapable of producing AGI.
Evidence (3 items)
Supports the claim: ARC Prize,
Passing ARC-AGI does not equate to achieving AGI, and, as a matter of fact, I don't think o3 is AGI yet.
Why it matters here: The evaluator recognized a major advance but explicitly rejected treating the tested o3 as AGI. The discussion also identifies failures on tasks easy for humans.
- Publisher
- ARC Prize
- Kind of source
- Independent measurement
- Describes events of
- Public from
Supports the claim: International AI Safety Report,
General-purpose AI systems have also been shown to sometimes fail at reasoning on seemingly simple tasks.
Why it matters here: The scientific synthesis describes uneven generalization and limits on independent extended work. Its chair's update recognizes o3's benchmark advances while noting the absence of public evidence about its open-ended real-world capabilities.
- Publisher
- International AI Safety Report
- Kind of source
- Secondary summary
- Describes events of
- Public from
Context only: arXiv / Microsoft Research authors,
The latest model developed by OpenAI, GPT-4, was trained using an unprecedented scale of compute and data.
Why it matters here: The authors describe GPT-4 as an early but incomplete AGI system, providing a material alternative use of the term. Their qualification does not establish completed AGI under the broad human-level reading used here.
- Publisher
- arXiv / Microsoft Research authors
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, OpenAI, in deep trouble
Scaling hasn’t gotten them —or anyone else — to AGI.
- Where
- paragraph 13
- Published
- 2025-03-01
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-openai-in-deep-trouble-52753c-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.