“Hedging with probably makes it sounds like we think there is a viable competitor out there for human general intelligence in the AI world. There isn’t.”
Our reading, the sentence that was checked:
There is currently no AI system that is a viable competitor to human general intelligence.
Assessment: Unresolved Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Three baffling claims about AI and machine learning in four days, statistical errors in top journals, and claims from Yann LeCun that you should not believe., paragraph 16.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-too-much-benefit-of-the-doubt-1b6e98-a1
We recorded 13 claims from this post. See every claim from this post.
The wording above is the version this site observed on 2026-10-29. The source may have been edited after 2022-10-16. 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
- a viable competitor
- human general intelligence
- in the AI world
- There isn’t
Our interpretation
- A viable competitor means an existing system with credible human-comparable breadth in learning, reasoning, language, and practical adaptation.
- The claim does not require an AI to exceed every human on every task. Conversely, superhuman performance in one specialty is insufficient.
- The surrounding child examples make flexible acquisition of unfamiliar abilities part of the comparison.
- For the ongoing state, I use October 1–16, 2022, the current reporting month through the post. Earlier system evaluations are context unless they establish the relevant October state.
Counts as supported if: Broad contemporary evaluations or a well-supported field assessment establish that existing leading systems fall materially short of human-comparable breadth and adaptability.
Counts as contradicted if: An existing system credibly demonstrates human-comparable breadth and adaptability across the relevant domains, including unfamiliar tasks.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
Outcome: Unresolved
The record was checked, and the evidence available by the cutoff does not settle it.
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 retrieved research points toward substantial limitations relative to humans, while Gato demonstrates meaningful breadth across domains. Neither source establishes a human-level general competitor. But failure to establish a competitor in these papers is insufficient to verify the categorical October claim. Their dates, coverage, and comparisons leave a material gap.
Limitations
- The evidence is stronger about particular systems than about the entire field.
- Expert benchmark performance is an imperfect measure of ordinary human general intelligence.
- Later AI advances cannot contradict this historical observation.
Evidence (2 items)
Context only: arXiv,
poorly
Why it matters here: The broad benchmark documents a substantial gap between the evaluated language models and expert human raters. It supports the direction of Marcus's assessment, but does not evaluate every October system or practical general intelligence.
- Publisher
- arXiv
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Context only: arXiv,
task learning
Why it matters here: Gato's experiments span text, vision, and robotic control, so it is relevant counterevidence to treating all AI as confined to one specialty. The paper does not establish human-comparable general adaptability.
- Publisher
- arXiv
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
What would settle it
A broad assessment covering systems existing by October 16, 2022, with explicit human comparisons across learning, reasoning, language, and practical adaptation, or a verified contemporary system meeting those criteria.
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Three baffling claims about AI and machine learning in four days, statistical errors in top journals, and claims from Yann LeCun that you should not believe.
Hedging with probably makes it sounds like we think there is a viable competitor out there for human general intelligence in the AI world. There isn’t.
- Where
- paragraph 16
- Published
- 2022-10-16
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-too-much-benefit-of-the-doubt-1b6e98-a1: Current. Outcome: Unresolved. 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-too-much-benefit-of-the-doubt-1b6e98-a1.