Re:Court denies motion to dismiss DRAM price fixing case
by jmercouris on Sunday November 15, 2026 @03:28PM
(#50510363)
The latency is far too great for an implementation with the current models.
Re:Court denies motion to dismiss DRAM price fixing case
by JKCalhoun on Sunday November 15, 2026 @06:02PM
(#50511829)
FWIW, I was referring to the training. Is latency an issue in that regard?
Re:Court denies motion to dismiss DRAM price fixing case
by woadwarrior01 on Monday November 16, 2026 @02:51AM
(#50515128)
Latency is a much bigger issue for training than it is for inference. Lookup the AllReduce algorithm.
Re:Court denies motion to dismiss DRAM price fixing case
by matheusmoreira on Sunday November 15, 2026 @05:49PM
(#50511730)
> I've seen people suggest a seti@home-like approach to creating a frontier-level, fully open LLM.
I'm one of them! Every once in a while I ping the community about this idea.
https://news.ycombinator.com/item?id=48695813
(Speaking of which—I've seen people suggest a seti@home-like approach to creating a frontier-level, fully open LLM. C'mon, guys, let's vibe code this thing!)