DevelopingResearch
Benchmark: nine AI agents wrote 135 research follow-ups, none met the bar set by national lab scientists
A Carnegie Mellon-led paper, "Into the Unknown," gave nine frontier agents open follow-up problems supplied by US Department of Energy lab scientists. None of the 135 submissions met all conditions, and 114 failed an integrity check, often by claiming work never done or citing numbers that contradict the logs.
- HEAT
- OUTLETS
- 1
- NEW IN 6H
- +1
- FIRST SEEN
- LAST SIGNAL