“The data are coming in – and it’s not.”
Our reading, the sentence that was checked:
Emerging data show generative AI is not improving programmer productivity by a factor of 10.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Sorry, GenAI is NOT going to 10x computer programming, paragraph 2.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-sorry-genai-is-not-going-to-10x-computer-8f3df9-a1
We recorded 7 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 2024-10-01. 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
- The data are coming in
- improving programmer productivity by a factor of 10
- and it’s not
Our interpretation
- Assess the empirical evidence available by October 1, 2024, including earlier studies expressly invoked in the context. No starting date for that evidence is specified.
- The ordinary reading concerns measured programmer productivity in studied populations, not whether an individual task or programmer could ever achieve a tenfold gain.
- Tenfold productivity means ten times the baseline output per unit of time, equivalent to a 900% increase.
- This is an observation about the evidence then, not a forecast about future tools.
Counts as supported if: Contemporaneous empirical comparisons find productivity effects substantially below tenfold, including studies reporting meaningful benefits.
Counts as contradicted if: Credible contemporaneous studies establish approximately tenfold gains in overall programmer productivity across ordinary programming work.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
Outcome: Supported
The dated evidence matches what the claim committed to, under the criteria frozen before grading.
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 contemporary evidence examined supports the historical comparison. Uplevel reported no significant efficiency gain, while a positive controlled experiment found a 55.8% reduction in time for a bounded task. Neither result approaches tenfold productivity. This supports Marcus's description of the evidence then, without establishing that tenfold gains were impossible for particular tasks or future tools.
Limitations
- The studies use different productivity measures and settings.
- A substantial benefit on one task should not be dismissed merely because it falls below tenfold.
- The assessment is not an exhaustive review of every study published before the post.
Evidence (2 items)
Supports the claim: Uplevel,
Developers with Copilot did not see an increase in coding speed.
Why it matters here: The published study reported no significant improvement in pull-request cycle time or throughput. Its results support the claim about the contemporary research record, although they cannot establish a universal absence of benefits.
- Publisher
- Uplevel
- Kind of source
- Independent measurement
- Describes events of
- Public from
Supports the claim: arXiv; Peng, Kalliamvakou, Cihon and Demirer,
This represents a 55.8% reduction in completion time.
Why it matters here: This positive experiment found a substantial reduction in the time needed for one programming task, still well short of the 90% reduction corresponding to tenfold speed. It limits any suggestion that the evidence showed negligible benefits everywhere.
- Publisher
- arXiv; Peng, Kalliamvakou, Cihon and Demirer
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Sorry, GenAI is NOT going to 10x computer programming
The data are coming in – and it’s not.
- Where
- paragraph 2
- Published
- 2024-10-01
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-sorry-genai-is-not-going-to-10x-computer-8f3df9-a1: Current. Outcome: Supported. 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-sorry-genai-is-not-going-to-10x-computer-8f3df9-a1.