Current LLM agents have severe jaggedness: they excel at verifiable, RL-trained tasks (code generation, tool use) but fail catastrophically at soft/unverifiable tasks (like telling jokes), because they're only optimized along dimensions where rewards can be computed, leaving trained-off domains untouched.
factualpending
Speaker
Andre KarpathyEvidence Quote
“The joke that apparently all the LLMs like love the most is why do scientists not trust atoms?... this is the joke you still get today.”
Source
Skill Issue: Andrej Karpathy on Code Agents, AutoResearch, and the Loopy Era of AI— No Priors: AI, Machine Learning, Tech, & StartupsCreated: 8/12/2026, 6:22:17 PM
Your notes
Loading notes...