Here is the article:
Dario Amodei, CEO of Anthropic, believes that recent developments with China’s DeepSeek are a reminder that export controls are crucial in the development of artificial intelligence (AI). In a blog post, Amodei wrote that he does not view DeepSeek as adversaries, but rather as smart and curious researchers who want to create useful technology. However, he believes that they are also beholden to an authoritarian government that has committed human rights violations and aggressive behavior on the global stage.
Amodei’s comments come after DeepSeek’s sudden emergence in the US and its rapid rise to the top of Apple’s App Store, causing a significant selloff in Nvidia’s stock and a market capitalization loss in the US. DeepSeek’s new AI model, R1, is a fraction of the cost to create compared to models built by US competitors.
This development is concerning for Anthropic, which is competing with OpenAI, Google, and others to develop the most advanced generative AI model. Amodei emphasized that the rise of DeepSeek highlights the need for the US to stay on top and avoid ceding too much of the market to China. He believes that the US needs to stay ahead in AI development and that export controls are crucial to achieve this goal.
Amodei also noted that the rise of DeepSeek makes controls on chip exports to China “even more existentially important” than before. The model was created despite a slew of Biden administration controls restricting the sale of certain advanced chips to China that could enhance military capabilities.
Amodei emphasized that the competition between the US and China is not about “ducking” the competition, but rather about staying ahead in the development of AI. He believes that US AI companies should strive to have better models than those in China if they want to prevail.
