“It has never worked.”
Our reading, the sentence that was checked:
Building ever larger LLMs has never succeeded in reaching AGI.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, A trillion dollars is a terrible thing to waste, paragraph 19.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-a-trillion-dollars-is-a-terrible-cf80f5-a1
We recorded 14 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 2025-11-27. 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
- It has never worked
- building ever larger LLMs
- in hopes of reaching AGI
Our interpretation
- It refers to achieving AGI through larger LLMs, not to achieving useful capabilities or commercial success.
- Use the ordinary public-evidence reading: no broadly accepted, independently substantiated demonstration of broadly human-level general intelligence had resulted from LLM scaling by publication.
- Never covers the historical record through 27 November 2025; no lower boundary is supplied.
- This observation does not predict that scaling can never achieve AGI.
Counts as supported if: Contemporaneous independent evaluations and capability assessments support the absence of a broadly accepted AGI demonstration from LLM scaling, including substantial remaining gaps in general performance.
Counts as contradicted if: A dated, independently substantiated and broadly accepted demonstration establishes that scaling LLMs had achieved AGI by publication.
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 public record supports the ordinary historical reading: scaling LLMs had not produced a broadly accepted AGI demonstration by the post. Independent cross-lab testing and the October capability assessment describe substantial remaining gaps. Google's November frontier release reports major advances while still describing AGI as a destination. This supports a statement about demonstrated achievement, not a claim that scaling is useless or permanently incapable of reaching AGI.
Limitations
- The conclusion depends on the ordinary public-demonstration reading of AGI.
- Benchmark failures alone do not prove the absence of general intelligence.
- The evidence does not establish future impossibility or undisclosed internal capabilities.
Evidence (3 items)
Supports the claim: ARC Prize,
new ideas for AGI
Why it matters here: ARC Prize's cross-lab testing found substantial shared limitations on novel tasks. Its assessment supports the absence of demonstrated AGI at that point, although one benchmark cannot define AGI by itself.
- Publisher
- ARC Prize
- Kind of source
- Independent measurement
- Describes events of
- Public from
Supports the claim: International AI Safety Report,
Performance gaps
Why it matters here: The broad scientific assessment documents continued limitations on realistic work despite major advances on selected tests. This supports the public-evidence reading of the claim, without proving a universal negative about every system.
- Publisher
- International AI Safety Report
- Kind of source
- Secondary summary
- Describes events of
- Public from
Context only: Google,
path toward AGI
Why it matters here: Google reports substantial new benchmark gains but presents the release as progress toward AGI. The announcement is a useful recent cross-check, not independent proof about the entire field.
- Publisher
- 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, A trillion dollars is a terrible thing to waste
It has never worked.
- Where
- paragraph 19
- Published
- 2025-11-27
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-a-trillion-dollars-is-a-terrible-cf80f5-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-a-trillion-dollars-is-a-terrible-cf80f5-a1.