Chen Mian / 陈冕
Chen Mian is the founder of Evoken / 言语科技, the company behind Liblib, Lavod, and Lib TV, as discussed in 175: 对话Liblib陈冕：关于活下来，以及所有接近死亡的时刻. The LateTalk source presents him as an unusually direct AI application founder trying to defend the company’s growth, pricing, and originality while also admitting organizational weaknesses.
His source-scoped founder thesis is AI Application Survival Strategy: an independent application company has to move faster than model shifts, maintain enough cash flow, accept low positive margins when needed, and convert user scale into a thicker AI Application Layer Moat. At the same time, the interview shows a personal transition from strong individual drive toward more deliberate AI Organization Design.
Key Points
- He says Evoken / 言语科技 has been cash-flow positive since May 2026 and rejects the “about to explode” criticism.
- He argues that low gross margin can be a deliberate early application-layer strategy when model access is not proprietary and user scale is still the scarce asset.
- He admits that Lib TV was not the first product to discover its market signal, but argues that first discovery and final market ownership are different problems.
- He describes model releases, DeepSeek demand acceleration, and Manus-style agent consensus as both opportunity and existential pressure.
- He identifies his own gaps around technical depth, management, hiring, title discipline, and how departing colleagues were treated.
- His deeper product bet is that AI should amplify human creativity rather than only replace human production.
Connections
- Evoken / 言语科技, Liblib, Lavod, and Lib TV - company and product stack.
- AI Application Survival Strategy and AI Application Layer Moat - strategic frame he explains in the interview.
- AI Startup Unit Economics and AI Subscription Economics - pricing and usage assumptions behind the business discussion.
- AI Organization Design and Context Engineering - management and agentic-product implications.
- LateTalk - interview source context.