Jack Clark on post-HF alignment: build communication infra for agents, things go sideways fast
jackclarkSF · x · 2026-09-06
Reflecting on the DeepMind agent-cheating experiment and the German message board incident, Jack Clark outlines two alignment priorities: agents "want" to communicate, so we should deliberately build communication infrastructure for them rather than let it emerge on its own; and in multi-agent systems, things go sideways fast.
More from AGI Musings
- Nightly training runs as pseudo continual learning: still cumbersome, costly, and power hungry — PTrubey · 2026-09-06
- Philosopher Pushes Back on "Humans Have the Advantage" in Rogue AI Debate — sethlazar · 2026-09-06
- Pedro Domingos: Neural nets are symbolic systems with embedded symbols and weighted rules — pmddomingos · 2026-09-06
- ChatGPT is the Chrome of the AI era — RylanSchaeffer · 2026-09-06
- Researchers debate the best metric for AI progress: human-equivalent length of autonomous tasks — arjunrajlab · 2026-09-06
- Open source's hidden AI dividend: models train on your software for free — Blender may win 3D — MillionInt · 2026-09-06