“More than that, any normal seven-year old develops a rich model of the world, including the objects they see, the people they interact with and so on, and even how their bodies work. GPT is still just faking it.”
Our reading, the sentence that was checked:
Unlike a seven-year-old, GPT lacks a rich model of the world and is just faking it.
Assessment: Unresolved Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Three years on, ChatGPT still isn't what it was cracked up to be – and it probably never will be, paragraph 78.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-three-years-on-chatgpt-still-isnt-22462c-a1
We recorded 30 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 2025-11-29. 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
- any normal seven-year old
- a rich model of the world
- the objects they see
- the people they interact with
- how their bodies work
- GPT is still just faking it
Our interpretation
- The plain reading is that typically developing seven-year-olds possess coherent, transferable understanding of physical, social and bodily situations that contemporary GPT lacks.
- 'Faking it' is read as producing convincing answers without comparable robust understanding, not as a claim about conscious deception.
- The surrounding discussion identifies GPT-5 as the principal contemporary comparison.
- An isolated model error would not establish the absence of a rich world model. A narrow internal representation would not establish child-like breadth.
- The claim describes current capability. With no reporting cycle specified, the target period is November 2025 through the post date.
Counts as supported if: Comparative evidence shows typically developing seven-year-olds reasoning coherently across unfamiliar physical, social and bodily situations while contemporary GPT systematically fails to transfer the relevant understanding; representation studies provide consistent additional evidence.
Counts as contradicted if: Comparable evidence shows contemporary GPT possessing coherent representations and child-comparable transfer across the stated domains. Narrow successes alone do not meet this condition.
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 research shows why fluent performance alone cannot establish coherent understanding. It also shows that a GPT-style model can learn a usable internal representation. Neither finding settles whether contemporary GPT lacked the broad, transferable understanding Marcus attributes to seven-year-olds. The cited studies concern earlier systems and narrower tasks.
Limitations
- The behavioral reading of 'faking it' is an evaluator interpretation.
- Evidence of one representation cannot establish child-like understanding across domains.
- Failures on selected tasks cannot establish the absence of a world model.
Evidence (2 items)
Context only: arXiv; Keyon Vafa and colleagues,
Such incoherence creates fragility
Why it matters here: The experiments support distinguishing apparent task competence from coherent representations and transfer. They do not establish the claimed absence in November 2025 GPT or compare it with seven-year-olds.
- Where
- Abstract
- Publisher
- arXiv; Keyon Vafa and colleagues
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Context only: arXiv; Kenneth Li and colleagues,
control the output
Why it matters here: The Othello experiments found a usable internal board representation in a specially trained GPT variant. This challenges a blanket absence-of-representations reading, but does not demonstrate a contemporary GPT model with a child's broad understanding.
- Where
- Abstract
- Publisher
- arXiv; Kenneth Li and colleagues
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
What would settle it
Broad, controlled comparisons of the relevant GPT version with seven-year-olds on novel physical, social and bodily reasoning, combined with model-specific representation and transfer tests. A clearer threshold for 'rich' would reduce interpretive uncertainty.
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Three years on, ChatGPT still isn't what it was cracked up to be – and it probably never will be
More than that, any normal seven-year old develops a rich model of the world, including the objects they see, the people they interact with and so on, and even how their bodies work. GPT is still just faking it.
- Where
- paragraph 78
- Published
- 2025-11-29
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-three-years-on-chatgpt-still-isnt-22462c-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-three-years-on-chatgpt-still-isnt-22462c-a1.