Search and filter through extracted claims from AI researchers.
Showing 2001-2020 of 2685 claims of type "fact"
GPT-5.6 Sol is showing significant results in finding and fixing novel vulnerabilities
Moonshot AI announced Kimi K3 as their most capable model to date with 2.8 trillion parameters
A new fund will award at least $200,000 in grants and prizes for corrigibility research in 2026
Kimi K3 represents a significant performance jump from K2.5 in open model capabilities
Kimi K3 can autonomously edit full video reels with frontier-level quality
Hidden channels in GNCA self-organize into modular groups in parallel with visible form development
Pipeline data may be stale or degraded.
Last synthesis: 2026-09-20. 8,964 pending.