Karina Nguyen at Princeton: judge AI research agents by taste and net value per joule

karinanguyen · x · 2026-11-18

Karina Nguyen gave a talk at Princeton on frontier LLM post-training, forecasting, and evals. Her framing: agents doing AI research should be judged like any researcher—by their taste in experiments and whether results justify the compute—plus a yardstick for where AI should run at all: net value per joule. Host Thoughtful Lab aims to bring RSI into the real world.

Original post →

More from AGI Musings

AGI Musings channel →