back
Reward Mismatches in RL Cause Emergent Misalignment
Get SIGNAL/NOISE in your inbox daily
Learning to do misaligned-coded things anywhere teaches an AI (or a human) to do misaligned-coded things everywhere. So be sure you never, ever teach…
Recent Stories
Jan 13, 2026
11 things UC Berkeley AI experts are watching for in 2026
How will AI disrupt the labor market? What will deepfake videos mean for our understanding of truth? Are we in a bubble, and if so, will the bubble burst?
Jan 13, 2026US approves sale of Nvidia’s advanced H200 chips to China
Nvidia has been caught in a tug-of-war between the US and China as the countries compete for AI dominance.
Jan 13, 2026Robots learn human touch with less data using adaptive motion system
Japanese researchers develop an adaptive robot motion system that enables human-like grasping using minimal training data.