I recently chatted with some startup teams in Hong Kong. They mentioned Meta, which is spending hundreds of millions of dollars poaching talent from various AI giants like a football player. Why is 40% of its hires coming from OpenAI? Anthropic alone has poached several former Meta employees. This is because Anthropic's corporate culture constantly values its employees, insisting on labels like "safety first," "developer-friendly," and "people before technology." This ensures that employees, while offering high salaries, are not likely to be poached for higher salaries. Meanwhile, OpenAI Amid the debates over for-profit vs. public welfare, open source vs. closed source, security vs. freedom, and board dismissal, the company has lost its distinct value proposition, which is the primary reason its remaining employees are easier to poach. The higher the level of talent, the harder it is to retain them with money alone. This is due to the diminishing marginal utility of money. Values are a primary consideration, but the founder's personal charm, blood ties, marriage, hometown/regional connections, and revolutionary affiliations also play a role. Many academics are researching equity distribution/option incentives for startup teams. This is an additional layer beyond financial commitment, but truly securing key individuals requires more. In Shenzhen, I chatted with my friends at @Trickle_HQ about these companies and the stories of Figma, Canva, and Vercel, which sparked these reflections.