News for hackers. Stuff that matters.

A comment on Running a 180B parameter LLM on a single Apple M2 Ultra. Parent comment

Re:Running a 180B parameter LLM on a single Apple M2 Ultra by ramesh31 on Thursday September 07, 2023 @12:13PM (#37421196)
>I think people will be surprised that consumers ultimately end up benefitting far more from LLMs than the providers. There's not going to be much moat or differentiation to defend margins... more of a race to the bottom on pricing

Should be pointed out that this didn't just happen out of thin air. These open models still cost millions of dollars to create. Meta let the genie out of the bottle, but it won't be free forever.

[ Parent | on HN ]
Re:Running a 180B parameter LLM on a single Apple M2 Ultra by logicchains on Thursday September 07, 2023 @12:22PM (#37421344)
>These open models still cost millions of dollars to create. Meta let the genie out of the bottle, but it won't be free forever.

This particular model was funded by the UAE government. If they could do it, it should be similarly possible for a western government to create and release one as a public good.

[ Parent | on HN ]