A relevation of understanding

i think i figured out how understanding works.

Factor, factor again, until you understand.

Where does a prior come from? It comes from factorization of another prior.

Where does “randomness” come from? It comes from computation. Also, there are uncountably infinite ways to complex extend λ (x: f32) => sinc(x). Fuck IEEE floats.

Humans had it easy. Reality is at least continuous.

2 ways of interacting with the æther outside the computer: reduce, copy

2 ways of interacting with another computer (still inside the conjoined computer): input, output

I’ve seen enough today 🫠

only speedrunners persistence can factor prior made up of IEEE floats; there is no simple solution,…

every data structure that deals with understanding by trying to store the prior inside the computer is wrong… like how copper wires guide EM waves, computers only need to guide the priors, not store them

prior is a mathematical concept; randomness is a real concept. mix the two and you get psychosis.

// footnotes

understanding always points outwards towards the thing being understood. you need to do I/O in order to use that understanding. a model is always wrong, so avoid having it in the first place.


a “model” in computing is maybe just a fancy way to say [cache]

i don’t consider caching itself as intelligence, although the caching strategy itself can be intelligent

i think understanding needs only coupling (find the thing to interact with) and use (interact with it). caching is not required.

understanding something means that you have access to it however you want. it isn’t cheating to look up a dictionary to use words correctly, but that means that the understanding is held by you plus the dictionary.