logoalt Hacker News

make3today at 3:18 AM2 repliesview on HN

NVidia has reasonable incentives to keep things open indefinitely, it wants people to use it's GPUs


Replies

noosphrtoday at 3:20 AM

Issue is that llama.cpp is the best way to run models on hardware that isn't nvidias.

show 1 reply
swozeytoday at 4:06 AM

It'd be really nice if I could use my egpu 4090 on my macbook pro..