François Chollet:数据与算力扩展只是泛化能力的乘数
Keras 之父 Chollet 发了一条推文,谈算力投资和真正的智能方法之间的区别,观点很尖锐,做研究的朋友可以看看。
François Chollet 在 X 上发文称,数据扩展和算力扩展只是对方法泛化能力的乘数,而非能力本身。他表示数万亿美元的投资能让泛化能力低下的低效方法显得够用,但不能使其成为 AI 的终极路线。他建议研究者选择可扩展的方法以享受这一乘数红利,同时强调更聪明的方法会扩展得更好。
Data scaling and compute scaling are *multipliers* for the generalization power of an approach. Trillions of dollars in investment can make very inefficient, low-generalization approaches look adequate, but that doesn't make them the ultimate end-all-be-all of AI.
You should work on methods that *can* scale, because if you don't benefit from this multiplier you simply can't compete. But you should keep in mind scaling is only a multiplier: a cleverer approach will scale *better*.