4 ms·
Downvote all you want, but the moment you ask for a new type of chip that's both cheap and useful for current ML work, you are literally asking for the impossib
by TheRealPomax 3y ago
Downvote all you want, but the moment you ask for a new type of chip that's both cheap and useful for current ML work, you are literally asking for the impossible.
- pama 3y agoI didn’t downvote, FWIW. I am also not asking for a chip. I wonder if it’s possible to solve a specialized problem with this type of computing, maybe something that right now is not addressed well, for example applications in quantum chemistry. Which are the startups you had in mind?