#英伟达ceo连线特朗普淡化ai风险
👉 AI安全的分歧

Ren Yinxun and the CEOs of two model companies publicly go head-to-head on AI safety 🖥️

He says slowing down won’t solve the problem; what should be done is to advance safety capabilities in parallel. Mark Zuckerberg of Meta stands on his side—their views are closer to Ren Yinxun’s.

On the other side is Dario Amodei from Anthropic. A few days ago, he published a long post arguing for putting the brakes on the development of cutting-edge models. The statements from both sides landed in the same week.

The core of their disagreement comes down to money. Slowing down would mean delaying compute investment, and the party that sells the chips won’t be happy. The two model companies call out the risks, while the hardware side calls it an opportunity. On the same day, their stock prices moved in completely opposite directions.

The market reaction was just as direct. AI safety and bond yields were repriced together, with capital raising the weight assigned to risk.

For crypto, the disagreement in the AI narrative will spill over. The lines for compute, energy, and tokenized compute will all sway along. When the direction isn’t clear, volatility comes first.

Who do you think is more convincing? Let’s discuss in the comments.