Cognitive Revolution Highlights: Recursive Self-Improvement in the Era of Model Division of Labor

The Cognitive Revolution · rss · 2026-08-28

Nathan Labenz and Prakash revisit highlights from three live AI:AM mornings with guests including Louis Kirsch and Damon Falck of Inherent Laboratories, Vercel CTO Malte Ubl, Genesis Molecular AI CTO Sergey Edunov, Arm's Mohamed Awad, and others.

The central thread: the meaningful unit of AI work is becoming a division of labor between models — e.g., Inherent's 27B "scientist" model using GPT-5.5 Codex for implementation, and production pipelines routing taste, execution, security, and hardware coordination across different systems. The episode weighs implications for recursive self-improvement, verification of discoveries humans may not be able to check, model refusals in security work, and infrastructure for agents acting in the world.

Its sharpest stake: whether frontier labs are scaling RL and agentic workflows on reward environments and vendor pipelines that may be too rushed, noisy, or gameable to support the institutional self-improvement they pursue. Chapters also cover China's compute abundance, photonic compute stacks, and Astra vs. AGI.

Original post →

More from AGI Musings

AGI Musings channel →