Skip to content
Slow Variables

Ask the data

Answers come from the same store as the site; every number is checked against the record it cites.

Indicators / series

watch.metr_feed.uplift.pt

Source: METR feed (watchlist) (METR) · Publisher terms; titles and links only · CSV

METR

idtextgradeas of datepublished datetieraudited vs reportedextraction methodreview statusretrieved athttp statuscontent hashsupersedes idurlflags
b45a53715a7fd5ddMeasuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivitygrade B, evidence tier 62025-07-102025-07-106reportedapiapproved2026-09-10T04:21:39.802316Z20092eb45fb9990source page
431be41f964168d0Analyzing coding agent transcripts to upper bound productivity gains from AI agentsgrade B, evidence tier 62026-02-172026-02-176reportedapiapproved2026-09-10T04:21:39.802316Z20092eb45fb9990source page
a5c66debc882a21bFive lessons from having helped run an AI-Biology RCTgrade B, evidence tier 62026-02-192026-02-196reportedapiapproved2026-09-10T04:21:39.802316Z20092eb45fb9990source page
fd47223741d27124We are Changing our Developer Productivity Experiment Designgrade B, evidence tier 62026-02-242026-02-246reportedapiapproved2026-09-10T04:21:39.802316Z20092eb45fb9990source page
3a2f45b81b75796eTask Substitution and Upliftgrade B, evidence tier 62026-05-082026-05-086reportedapiapproved2026-09-10T04:21:39.802316Z20092eb45fb9990source page
d8b60fd2e9243524Measuring the Self-Reported Impact of Early-2026 AI on Technical Worker Productivitygrade B, evidence tier 62026-05-112026-05-116reportedapiapproved2026-09-10T04:21:39.802316Z20092eb45fb9990source page
ead8cfa609343259Because 8 ≈ e², Anthropic's researcher uplift is plausibly >2xgrade B, evidence tier 62026-07-082026-07-086reportedapiapproved2026-09-10T04:21:39.802316Z20092eb45fb9990source page
Raw snippets and dispute text
  • b45a53715a7fd5dd Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity
  • 431be41f964168d0 Analyzing coding agent transcripts to upper bound productivity gains from AI agents
  • a5c66debc882a21b Five lessons from having helped run an AI-Biology RCT
  • fd47223741d27124 We are Changing our Developer Productivity Experiment Design
  • 3a2f45b81b75796e Task Substitution and Uplift
  • d8b60fd2e9243524 Measuring the Self-Reported Impact of Early-2026 AI on Technical Worker Productivity
  • ead8cfa609343259 Because 8 ≈ e², Anthropic's researcher uplift is plausibly >2x