Hacker News new | ask | show | jobs
by truelson 240 days ago
I think this points to future optimizing will be focused on making inference really inexpensive.