Accidental Tech Podcast696: It Seems Petty, But I Endorse It
https://atp.fm/696
Lots of details from macOS Golden Gate, Apple's AI Tech Talk, road trips, and failed Bobs.
Luke Ratta@atpfm So, the M2 Ultras aren’t supported for the biggest models locally, but they’re also what’s running even larger models in private cloud compute? I’m not sure how that makes sense.
John Siracusa@lukeratta I think maybe they running only the AFM 1 and 2 models, and all AFM 3 models are on Nvidia in Google’s datacenter.
@siracusa Ah, that makes sense. Thank you.