Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

The absolute time depends on hardware, optimizations, exact model, etc; it's not a very meaningful number to quantify the reinforcement technique they've developed, but it is very useful to estimate their training hardware and other proprietary information.


It's not about the literally quantity/value, it's about the order of growth of output vs input. Hardware and optimizations don't really change that.


Exactly, that's why the absolute computation time doesn't matter, only relative growth, which is exactly what they show.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: