An autonomous AI team beats human-led model research on equal time and compute, setting its agenda (within ~12 months)
“An autonomous AI team beats human-led model research on equal time and compute, setting its agenda.”
Context: State of AI Report 2026, one of nine predictions for the next twelve months (Nathan Benaich / Air Street Capital). Page published 2026-10-08. FAQ: completing experiments chosen by humans would not satisfy all conditions.
How we judge it
Our interpretation: YES if by 8 Oct 2027 a credible public report shows an AI system/team that (1) sets its own model-research agenda and (2) outperforms a human-led team given equal time and compute on that research.
Resolution criteria are our interpretation of the claim, written before grading. Methodology
- Said
- Resolve by
- Due in 362 days
- Verdict
- Open
- Stated probability
- None stated
- Specificity
- 2 of 3 (dated, some interpretation)
- What changes for
- Research & science
- Readiness milestone
- Part of AI automates AI research (recursive self-improvement) (says it happens)
Share this scorecard
Post on X LinkedIn Share image (PNG)
Embed / Cite
The embedded card is live: it shows the current verdict, with the date the data was read. If a verdict is corrected later, the card changes too, so cite the page and its date for a fixed record.