4 ms·
As long as the inference can be done locally on their chips I don't at all think it's a big deal to train models on Nvidia/other hardware. Are all the iCloud s
by dcchambers 3y ago
As long as the inference can be done locally on their chips I don't at all think it's a big deal to train models on Nvidia/other hardware.
Are all the iCloud servers running on Apple silicon? I assumed they were running on standard rack mounted hardware.
- tambourine_man 3y agoI think Apple considers cloud infrastructure a necessary evil and a commodity. AI isn’t, yet at least, and I don’t think they can afford to treat it as such.