Jensen Huang accidentally calls for shutting down OpenAI, per Zvi's podcast breakdown

Don't Worry About the Vase (Zvi) · rss · 2026-09-25

Zvi breaks down Jensen Huang's interview on Ezra Klein's podcast, arguing the Nvidia CEO's position is riddled with contradictions.

No ASI, no existential risk: Huang sees AI as merely a 'new abstraction level' of software, claims AI only became useful in the last six months, and shows no appetite for treating AI as fundamentally different.
The 'HuggingFace Incident': Huang reasons like a traditional product engineer — once a product works, shift R&D to safety, verification and evals; labs that can't test safely 'must be shut down.' Zvi argues this logic proves too much: taken seriously, it implies shutting down OpenAI and frontier labs today.
Weak jobs claims: Huang credits $500B of AI investment with job creation and blames manufacturing job losses on outsourcing; research attributes 70-80% of those losses to automation. On falling junior SWE openings, his answer is 'wait two years' for AI-native graduates.
Stakes: Huang shapes chip allocations, heavily influences Trump-era US AI policy, and lobbies forcefully — making his risk denial consequential.

Zvi's verdict: Huang's safety tolerance is lower than even the most paranoid safetyist's, yet he misunderstands what the real risks are — engineering thinking mismatched onto AI.

Original post →

Jensen Huang accidentally calls for shutting down OpenAI, per Zvi's podcast breakdown

More from AGI Musings