“productivity studies sometimes show a 30% gain, but none have come close to 10x (1000% gain).”
Our reading, the sentence that was checked:
No productivity studies of LLMs have shown anything close to a 10x gain; some show around 30%.
Assessment: Mixed 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 12.
- 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-ed689e-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
- productivity studies
- sometimes show a 30% gain
- none have come close to 10x (1000% gain)
Our interpretation
- Assess studies publicly available during the three years discussed in the post, through November 29, 2025.
- The ordinary reading includes measured productivity on specific work tasks, with attention to output quality. It does not require a tenfold improvement across an employee's entire job.
- Model estimates of hypothetical time savings are weaker evidence than measured human performance.
- Tenfold output is a 900% increase. The parenthetical says 1000%, but this arithmetic discrepancy does not materially change the comparison.
Counts as supported if: Studies establish gains around 30%, and a sufficiently broad review of the relevant literature finds no measured task-productivity gains approaching tenfold.
Counts as contradicted if: A study available before the post reports approximately tenfold productivity on a genuine work task with assessed output quality. A narrow task result contradicts the universal wording without establishing tenfold productivity across whole jobs.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
Outcome: Mixed
Part of the claim holds and part does not, and the record says which part is which.
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 around-30% portion is supported by a literature review. The unrestricted claim that no study approached tenfold gains has a published counterexample: an orthopedic documentation pilot reported tenfold generation speed with assessed quality. That result is narrow and does not show that an individual worker can perform ten people's jobs. The proposed outcome therefore depends on including task-specific studies within Marcus's wording.
Limitations
- The counterexample involved six fictional cases and document production, not overall physician productivity.
- A reading restricted to complete jobs would require a different assessment.
- Anthropic's estimates are contextual, not decisive.
Evidence (3 items)
Supports the claim: Tom Coupé and Allen Wu, University of Canterbury,
about 33%
Why it matters here: The published review supports the around-30% portion. An average cannot establish that no individual study found a much larger gain.
- Publisher
- Tom Coupé and Allen Wu, University of Canterbury
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Goes against the claim: Acta Orthopaedica,
10 times faster
Why it matters here: This published pilot compared document generation for six fictional orthopedic cases and assessed quality using a blinded expert panel. It contradicts the unrestricted assertion about studies, but does not demonstrate tenfold productivity across physicians' entire jobs.
- Where
- Abstract, Results
- Publisher
- Acta Orthopaedica
- Kind of source
- Peer-reviewed paper
- Describes events of
- Public from
Context only: Anthropic,
90%
Why it matters here: This pre-post research also reported very large task-level savings. It used Claude's estimates and excluded some work outside conversations, so it is not decisive evidence of measured worker productivity.
- Publisher
- Anthropic
- 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, Three years on, ChatGPT still isn't what it was cracked up to be – and it probably never will be
productivity studies sometimes show a 30% gain, but none have come close to 10x (1000% gain).
- Where
- paragraph 12
- Published
- 2025-11-29
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-three-years-on-chatgpt-still-isnt-ed689e-a1: Current. Outcome: Mixed. 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-ed689e-a1.