January has been notable for the variety of vital bulletins in AI. For me, two stand out: the US authorities’s help for the Stargate Challenge, an enormous information middle costing $500 billion, with investments coming from Oracle, Softbank, and OpenAI; and DeepSeek’s launch of its R1 reasoning mannequin, skilled at an estimated price of roughly $5 million—a big quantity however a fraction of what it price OpenAI to coach its o1 fashions.
US tradition has lengthy assumed that greater is best, and that costlier is best. That’s definitely a part of what’s behind the most costly information middle ever conceived. However we’ve to ask a really totally different query. If DeepSeek was certainly skilled for roughly a tenth of what it price to coach o1, and if inference (producing solutions) on DeepSeek prices roughly one-thirtieth what it prices on o1 ($2.19 per million output tokens versus $60 per million output tokens), is the US know-how sector headed in the correct route?
Be taught sooner. Dig deeper. See farther.
It clearly isn’t. Our “greater is best” mentality is failing us.
I’ve lengthy believed that the important thing to AI’s success can be minimizing the price of coaching and inference. I don’t consider there’s actually a race between the US and Chinese language AI communities. But when we settle for that metaphor, the US—and OpenAI specifically—is clearly behind. And a half-trillion-dollar information middle is a part of the issue, not the answer. Higher engineering beats “supersize it.” Technologists within the US must be taught that lesson.
