Yu Yi

Yu Yi is an AI-first practitioner and What’s Next｜科技早知道 S10E18 guest in 1 人公司，扛 5 个人的活，还要管 50 个 Agents？｜S10E18. The source describes him as someone who has spent about 5,000 hours treating himself as an AI-first experiment, with most daily work routed through AI while he preserves final judgment over high-quality outcomes.

Yu Yi’s distinctive contribution is the partner-like agent frame. He wants agents to accumulate experience, push back on his views, and work toward the user or customer rather than merely obeying his immediate instruction. At the same time, he keeps explicit red lines around deletion, protocol changes, spending, and socially damaging actions, making his practice a concrete example of Human-Agent Collaboration plus Agent Permission Boundaries.

Connections

• One-Person Company — Yu Yi frames OPC as a way to test and start, not necessarily a permanent one-person endpoint.
• Building Public and Trust As Business Asset — he argues that founders need self-media, public proof, and trust as AI makes content easier to fake.
• AI Use Pacing — his account of many parallel windows and tasks shows how agent abundance can create attention overload.
• Context Engineering and Human Judgment Under AI — his workflow depends on instructing, reviewing, and correcting AI rather than fully abdicating judgment.