3 ms·
I remember that. That was when they realized the deep learning was largely unnecessary, and they could just use their massive compute resources to brute force
by yuffffley 4mo ago
I remember that.
That was when they realized the deep learning was largely unnecessary, and they could just use their massive compute resources to brute force the problem space.
Proving that we would greatly benefit from using our compute resources for science rather than showing ads, and then we just kept showing ads.
- dekhn 4mo agoAlphaFold is based on deep learning and it's not brute force.
- TeMPOraL 4mo agoYou could argue that training SOTA LLMs is pre-bruteforcing every problem everywhere all at once.
- tmule 4mo agoWhat brute force? Any citations?
- herunan 4mo agohttps://en.wikipedia.org/wiki/Infinite_monkey_theorem https://en.wikipedia.org/wiki/Infinite_monkey_theorem
- mapmeld 4mo agoIf AlphaFold really is brute force on known protein problem space, would it then be usable as a model for novel proteins?