The “bio-weapon version” of Mythos

The “bio-weapon version” of Mythos

Last Week in AI

The article discusses Andy Jones’s early research at Anthropic: training AI (e.g., GPT-3) to master scalable, simplified games as a preliminary step toward automating R&D—but it lacks technical depth and empirical metrics.

1. Andy Jones joined Anthropic as a founding employee based on research training AI 2. The study used a ‘tunable-complexity game’ as a testbed to evaluate model scalab 3. The work is framed as early exploration of AI-driven R&D automation, but no perf

AI research, reinforcement learning, Anthropic, scaling laws, automated R&D