Merrett advocates for safety-paced AI development, halts top training run
repligate · x · 2026-08-19
- Stance: Merrett advocates for "pacing" over "pausing" AI development, arguing that confidence in safety should set the pace. Signed the "Pacing the Frontier" initiative.
- Actions: Temporarily slowed some frontier training runs to strengthen security and monitoring. The largest planned frontier RL run remains on hold.
- Call to Action: Urgently needs tools for labs and countries to coordinate on pacing. Committing to sharing learnings as the approach evolves.
- Support: SkyeSharkie agrees, noting pacing prevents blindly shoving massive amounts of outsourced training data into runs.
Related event: Anthropic Pauses Some Frontier Training to Strengthen Safety(3 posts)→
More from AGI Musings
- AI, Robotics, and Biomedicine Are Converging as Underlying Models Accelerate All Fields — r0ck3t23 · 2026-08-19
- Robots need to beat humans on cost-per-dollar, not output-per-hour — VraserX · 2026-08-19
- Readiness to abandon Claude shows lack of differentiation, proving intelligence is commoditized — annetgriffin · 2026-08-19
- AI translations are so precise that all online content is now universally comprehensible — yungcontent · 2026-08-19
- RSI skeptic: software improvement is already 99.9999% automated — the last 0.0001% won't be magic — binarybits · 2026-08-19
- Classic Paper: The Surprising Creativity of Digital Evolution — Ghost_Pilot_MD · 2026-08-19