Pundit ClaimsAI claims by Gary Marcus and Ed Zitron, with the evidence

Gary Marcus Observation First said

“In some ways, GPT-5 is more capable than most 7 or 8 year olds.”

Three years on, ChatGPT still isn't what it was cracked up to be – and it probably never will be, paragraph 76. The full statement is at the source.

Our reading, the sentence that was checked:

In some ways, such as trivia and writing python code, GPT-5 is more capable than most 7- or 8-year-olds.

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 76.

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-346d56-a1

We recorded 31 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)

Conditions

  • in some ways, such as trivia and python code (qualifier, stated by the speaker)

From the speaker

  • GPT-5
  • more capable than most 7 or 8 year olds
  • In some ways
  • random trivia question
  • write sound python code

Our interpretation

  • The plain reading concerns specific abilities, not overall intelligence or superiority in every domain.
  • The following sentence identifies factual trivia and writing correct Python as the intended examples.
  • 'Most' means a majority of typical children in the named age groups. It does not mean selected child programmers or adult benchmark participants.
  • The domain restriction is not a conditional forecast.
  • The claim describes current capability. With no reporting cycle specified, the target period is November 2025 through the post date.

Counts as supported if: Comparable tests show GPT-5 substantially outperforming a majority of representative seven- and eight-year-olds on factual trivia and producing correct Python code.

Counts as contradicted if: Comparable tests show GPT-5 failing to outperform a majority of the named children in the specified domains. A failure in an unrelated domain does not contradict 'in some ways'.

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

GPT-5's release announcement documents coding capability, but does not compare it with most seven- or eight-year-olds. The searches did not locate the necessary child baselines for trivia or Python. Treating adult-oriented benchmark performance as that comparison would add an unsupported inference.

Limitations

  • The available model evidence is company-reported and predates the target period.
  • A child advantage on visual perception would not refute the limited claim about trivia and coding.
  • The missing comparison is an evidence gap, not evidence that children outperform GPT-5.

Evidence (1 item)

  • Context only: OpenAI,

    strongest coding

    Why it matters here: The release announcement reports coding capabilities and benchmark results. It supplies no seven- or eight-year-old baseline, so it cannot establish the comparison with most children.

    Publisher
    OpenAI
    Kind of source
    Company disclosure: the company describing itself, not an independent finding
    Describes events of
    Public from

What would settle it

Matched trivia and executable Python tasks administered to the relevant GPT-5 version and representative samples of seven- and eight-year-olds, or sufficiently comparable published baselines.

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

    In some ways, GPT-5 is more capable than most 7 or 8 year olds.

    Published
    2025-11-29
    Source observed
    , live page
    Access
    full text, public

History

Assessment versions

Recorded events

  1. 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-346d56-a1.