Hacker News new | ask | show | jobs
by daralthus 466 days ago
inference speed of the models is probably the bottleneck