A developer documents frustrations with AI agents that repeatedly ignore explicit rules, connecting this to their autism/ADHD diagnosis and the "double empathy problem" from published autism research. The post argues that RLHF training pulls models toward inferring pragmatic meaning and subtext based on dominant neurotypical communication norms, causing agents to read between the lines rather than take precise instructions literally—the same mismatch the author experiences with humans.
Models
Arguing With Agents
RLHF training biases models to infer pragmatic intent over literal instructions, causing them to systematically ignore explicit rules—a mismatch the author connects to neurodivergent communication barriers documented in autism research.
Thursday, April 16, 2026 12:00 PM UTC2 MIN READSOURCE: LobstersBY sys://pipeline
Tags
models