AI on-the-job learning paradigm

Tag "AI on-the-job learning paradigm" — all articles

062026-11Dwarkesh Patel Publishes Three Pieces on AI Compute Centralization and On-the-Job LearningFrom June to August 2026, Dwarkesh Patel published a series of discussions on the AI frontier, focusing on the shift in training paradigms toward on-the-job learning, recursive self-improvement, and the concentration of compute. Dylan Patel predicts that Anthropic and OpenAI will control most of the world's compute by 2028; Ryan Greenblatt debates what happens once AI can automate AI research; and another piece argues that locking in AI safety regulation now is a mistake.Dwarkesh Patel