Love this bit from Jack Clark’s latest Import AI newsletter:
If you zoom out and look at the history of technology a huge amount of progress comes from weird people doing weird and off-consensus things. The history of computation, information theory, math, airplanes, physics, biology, etc, is full of crazy rebels who look at the world and come up with an at-the-time ‘crazy’ interpretation on reality which lets them ratchet science forward. Therefore, shouldn’t we in some sense expect the frontier of AI to be dominated by similarly heterodox individuals and organizations? And therefore shouldn’t we just accept that the frontier of AI, like other branches of science, will be full of kind of weird and/or alien-seeming people and orgs, and just accept this?