In an essay, Dario Amodei, CEO of Anthropic, argues that U.S. export controls on AI chips are hindering Chinese companies like DeepSeek from achieving maximum potential. He notes that DeepSeek's models, while competitive, still lag behind U.S. models in terms of performance. Amodei praises the engineering talent at DeepSeek but emphasizes that while the U.S. is experiencing cost reductions in AI training, strategic export policy decisions will significantly shape the future competitive landscape in AI, especially in light of upcoming policy changes under the Trump administration.
DeepSeek produced a model close to the performance of U.S. models 7-10 months older, for a good deal less cost (but not anywhere near the ratios people have suggested).
U.S. companies [are also] achieving the usual trend in cost reduction. The efficiency innovations DeepSeek developed will soon be applied by both U.S. and Chinese labs.
Collection
[
|
...
]