Does Capitalism Doom Every Closed AI Model to Get Copied?

Summary

The panel analyzes the fight over open vs. closed AI models, sparked by Jensen Huang's letter signed by Nvidia, Meta, and others while Anthropic and Amazon hold out. Guests discuss whether Chinese open-weight models like DeepSeek pose a real threat, the economics of per-task cost vs. token cost, distilled IP, and enterprise security concerns. The consensus is that market forces, not regulation, should decide winners, but no clear public-market investment edges emerge.

• Nvidia's Jensen Huang posts first X letter supporting open AI models, signed by 50 companies.
• Anthropic and Amazon remain holdouts; OpenAI and Alphabet later signed.
• Nick Carter argues the US government owes large labs no business model.
• Chinese models like DeepSeek and Kimi K3 are raising at huge valuations (e.g., $70B) but face cost and trust hurdles.
• Lorenzo Valente notes per-task costs are not drastically lower due to higher token usage.
• Chris Perkins says Chinese open-weight models are not enterprise-ready due to data-breach risks.
• Ram Ahluwalia compares closed AI to the Wright brothers, suggesting distillation will erode moats.
• The debate centers on whether market competition or regulation should govern the open/closed balance.