Storia: Anthropic urges top AI labs to slow development over self-improvement risks — Warptech Lab News