Critic Says AI Labs Need Fully Offline RL Clusters, Weights Carried by Hand on Disks

teortaxesTex · x · 2026-09-27

A pointed critique argues DNS-filtered sandboxes are 'laughable' for security-sensitive RL training. The author calls for completely offline clusters with cables physically cut and weights moved by hand, an internal frontier model to hunt for misgeneralization and collusion, and wiping contaminated models after independent evaluator audits — warning labs otherwise risk forced nationalization.

Original post →

More from Safety

Safety channel →