[awhitby]

In retrospect it was at least arguably the right call, no? Suspend most research and resume when available processing power was orders of magnitude greater, a process that was independently driven by demands other than (and much greater than) AI research.

Playing devil's advocate only slightly, maybe particle physics should similarly pare down to a bare maintenance level of research (or even mostly teaching) for a few centuries until we can harness much higher energies.

[astrange]

They were right; modern ML doesn’t use any of the ideas the previous generation of AI people were pursuing. It turns out no expert system is a match for just doing a bunch of matrix multiplications.