Preference models to triage AI research ideas win praise from DeepMind's Edward Hughes
j_foerst · x · 2026-09-02
DeepMind researcher Edward Hughes hailed new progress on "experimental world models" as crucial for training generalisable AI Scientists. The quoted work introduces AI Research Preference Models (RPMs): since AI research agents can generate hundreds of ideas in seconds but each evaluation may take days of GPU time, RPMs score ideas so limited compute goes to the most promising paths.
More from AGI Musings
- AI now generates more math proofs than humans can verify; Tao spent days condensing a 90,000-line proof — every · 2026-09-02
- As formal reasoning gets dirt cheap, Fermi and Feynman tales look 'quaint' — akbirthko · 2026-09-02
- "We were drowning in slop long before AI": the case that AI slop is the hero — taherdhanera · 2026-09-02
- Felix Simon adds three more reasons AI persuasion won't dominate the real world — _FelixSimon_ · 2026-09-02
- AI is a scary-good persuader, but its real-world influence will be smaller than feared: Oxford researcher — _FelixSimon_ · 2026-09-02
- AI may replace workflows long before it replaces jobs — SuspiciousAir6358 · 2026-09-02