“And I certainly don’t think that ChatGPT has lived up to expectations.”
Our reading, the sentence that was checked:
ChatGPT has not lived up to expectations.
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 7.
- 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-5fcff7-a2
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
- ChatGPT
- has lived up to expectations
- Three years ago, on November 30, 2022
- AGI by 2025
- “10x” productivity
- anything a PhD student could do
- a billion dollar company run by just a single employee
Our interpretation
- The plain contextual reading concerns the strong expectations listed immediately afterward, not every user's expectations or whether ChatGPT has any value.
- The target is ChatGPT's record from its stated launch through publication on 2025-11-29.
- Tenfold productivity means ten times comparable output per unit of labor, with quality accounted for.
- An expectation not yet delivered by publication is distinct from a forecast whose deadline has expired. The end of 2025 had not arrived.
- An overall verdict needs evidence covering the cited expectations; isolated task results cannot establish the whole comparison.
Counts as supported if: Evidence from the target period establishes that ChatGPT had materially fallen short of the cited expectations of broad human or PhD-level competence and transformative worker-level productivity, with adequate coverage of the other cited outcomes.
Counts as contradicted if: Evidence from the target period establishes delivery of the cited capabilities and productivity outcomes at their stated breadth, rather than isolated demonstrations or selected 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
OpenAI's August 7, 2025 disclosure reports GPT-5 matching or exceeding experts in roughly half the cases on an internal benchmark spanning over 40 occupations; this supports a limited shortfall from universal expert competence [e1]. The July 2023 study summary reports 40% less completion time and 18% higher quality on selected writing tasks, showing gains without demonstrating tenfold whole-worker productivity [e2]. These findings do not cover the full set of expectations or establish ChatGPT's overall record against them through November 29, 2025 [e1, e2].
Limitations
- The capability comparison is company-reported and limited to its internal benchmark.
- The writing experiment concerns selected tasks and a 2023 version of ChatGPT, not representative whole-worker productivity at publication.
- The packet does not adequately cover AGI or the one-person billion-dollar company example.
- The cited expectations' intended delivery dates and weighting are not established.
Evidence (2 items)
Supports the claim: OpenAI,
economically valuable knowledge work.
Why it matters here: The section reports expert-comparable or better performance in roughly half the cases across more than 40 occupations. This supports a narrower shortfall from universal expert competence, while also documenting substantial capability; it does not measure all the expectations.
- Publisher
- OpenAI
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
Supports the claim: American Association for the Advancement of Science,
decrease by 40%
Why it matters here: The publisher reports meaningful improvements in professional writing tasks, below a tenfold time-productivity improvement. This supports the narrower shortfall on those tasks, but neither measures whole jobs nor establishes the performance of late-2025 ChatGPT.
- Publisher
- American Association for the Advancement of Science
- Kind of source
- Secondary summary
- Describes events of
- Public from
What would settle it
Documented expectations with their intended delivery dates, together with evaluations and verifiable outcome records through November 29, 2025 covering broad competence, representative worker productivity, AGI, and the one-person company example.
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
And I certainly don’t think that ChatGPT has lived up to expectations.
- Where
- paragraph 7
- Published
- 2025-11-29
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-three-years-on-chatgpt-still-isnt-5fcff7-a2: 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-5fcff7-a2.