> It's my opinion that a lot of science today is theoretical BS Well, at least in computer science (ML) it's not what I see. It's very much empirically driven

> It's my opinion that a lot of science today is theoretical BS Well, at least in computer science (ML) it's not what I see. It's very much empirically driven and with a heavy engineering vibe. A little bit more science would often help imho, but papers without strong result are hard to publish, even if they improve our understanding of something. It sometimes feels a bit like a large corporate research lab, which is only a small part of science. We do not only create, we want to understand!