This is a microcosm of a much older struggle between empiricism and rationalism. Marcus is highlighting the current bias in AI towards empiricism, but it’s prevalent across much of science and medicine. There have been many corrections over history to reset the balance, and he’s right to expect we’ll see another one soon.

There’s a fascinating resonance in these pendulum swings. Medicine and AI is a great example. Visions of data-driven AI inspire visions of data-driven medicine. It’ll be interesting to see whether a fast correction in AI leads to a correction in the notoriously slow-moving realm of medicine, as well.