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

Respect to Jensen, one of the OGs of the valley and a good dude. But LLMs (eventually) running on iphone hardware will crater this run


Not sure how much inference on the edge will impact things unless you think we’ll hit “peak training” in the near future. I would safely wager that most H100 nodes will be used for training rather than inference.




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

Search: