Human OS — 10 Books Every AI Agent Should Read

De-escalation & empathy: turning conflict into needs, requests, and workable next steps.

Negotiation engine: separate people from the problem, focus on interests, invent options, use objective criteria.

Cross-cultural alignment: direct vs indirect speech, disagreement norms, trust styles, hierarchy expectations.

Human-centered design: affordances, feedback, and why users “misuse” systems that were designed wrong.

Reliability under pressure: how humans prevent errors in complex, high-stakes workflows.

Human simulation: incentives, self-deception, love, ambition, moral compromise—rich training data for “why people do things.”

Human nature update: a realistic but optimistic baseline for social cooperation and prosocial behavior.