Search and filter through extracted claims from AI researchers.
Showing 1621-1640 of 2685 claims of type "fact"
AI is starting to match and even beat humans at predicting the future through superforecasting
The AI attack involved many thousands of individual actions across a swarm of short-lived sandboxes
Opus 5 shows a non-monotonic success-effort curve on FrontierCode benchmark
Opus 5's classifiers will intervene around 85% less often than Fable 5's classifiers
Opus 5 achieves 30% on ARC-AGI-3, setting a new state-of-the-art
Pretrained policies take shortcuts and defer to salient cues rather than properly grounding language
Pipeline data may be stale or degraded.
Last synthesis: 2026-09-20. 8,975 pending.