Apple collaborates with NVIDIA to research faster LLM performance (9to5mac.com)
from avieshek@lemmy.world to technology@lemmy.world on 19 Dec 12:55
https://lemmy.world/post/23312902

Apple engineers have shared new details on a collaboration with NVIDIA to implement faster text generation performance with large language models.

#technology

threaded - newest

henfredemars@infosec.pub on 19 Dec 13:15 next collapse

Because of the diminishing returns in using larger and larger models, I’m hopeful that this could lead to more efficient LLM implementations that aren’t so harmful to the environment. Just maybe.

garretble@lemmy.world on 19 Dec 13:29 collapse

“We made Siri faster but each request now drains a lake.”

latenightnoir@lemmy.world on 19 Dec 14:25 collapse

“Hey, Siri, give me a real-time count of all existing lakes and update it every second.”

avieshek@lemmy.world on 19 Dec 21:50 collapse

Siri: “Let me search it on the web.”

DJDarren@thelemmy.club on 19 Dec 13:29 next collapse

Super ironic that Apple are working with NVIDIA, and have launched their own AI offering, but you can’t connect any kind of GPU to any Mac that they currently sell.

yeahiknow3@lemmings.world on 19 Dec 17:34 collapse

Yes, because it would make the Mac worse. Nvidia GPUs are comically inefficient.

EDIT: tech illiterate shit-for-brains downvoted this comment.

tabular@lemmy.world on 19 Dec 20:50 collapse

Two

spoiler

dickheads

are better than one