Thread by @cauchyfriend

[scattered thoughts]

It's amazing how much of our understanding of the world is downstream from the contemporary philosophy of science we use within mathematics and the sciences.

so much of the 19th/20th century was an attempt to build some sturdy foundation

for us to build mathematics and science. while in a strict sense we failed, the amount of formalization within math and physics was still incredible.

then so much of applied social sciences tried to build using that same approach towards understanding reality.

for most of us, this is still how the world was taught to us. but it seems we're moving away from it now. the new world is much stranger, it's building vast machines that we in some sense don't really understand.

despite huge success, it's sort of an admission of failure.

presumably, latent in the architecture of the networks -- hidden from us -- is an elegant and beautiful piece of code, that expresses the logic of burgeoning intelligence, which humans couldn't write.

we're moving to a world where our engineered constructs "know" and operate on scientific knowledge that, for the first time, no human could write down or hold in their head.

that's super weird.