AI Fund’s GP, Andrew Ng: LLMs as the Next Geopolitical Weapon & Do Margins Still Matter in AI?
Andrew Ng argues open-weight LLMs are a geopolitical weapon, chip export controls backfired on the US, and AGI timelines are wildly overhyped.
- China releasing open-weight models builds geopolitical soft power: the country of origin shapes answers to politically sensitive questions for global users.
- US chip export controls backfired — they accelerated China’s domestic semiconductor industry rather than slowing it down.
- Engineers with 10–20 years of experience who have mastered AI tools are the most productive cohort; fresh grads who ignore AI are the cohort most at risk.
- Ng considers AGI in 2 years “ridiculous” and places useful-AGI timelines at decades away; Karpathy’s claim that useful agents are a decade away is also wrong — agentic workflows are delivering ROI now.
- AI Fund’s venture studio model: seed idea before finding a founder, enter at ~$4M cap (~20% ownership) plus common stock for sweat equity.
- Ng hopes for 5–6%+ GDP growth from AI (closer to Masa Son than Karpathy’s 2%), driven by making intelligence cheap for everyone.
- The biggest enterprise AI barrier is change management, not data; data bottleneck is overhyped.
- Advising people not to learn to code because AI will replace it is the worst career advice possible — coding skills compound with AI assistance.
2025-11-17 · Watch on YouTube
