Ask HN: How well too LLMs scale?

2 points by pyuser583 ↗ HN
How well do LLMs scale? Can we expect one LLM instance per person in the future?

1 comment

[ 3.4 ms ] story [ 13.8 ms ] thread
Yeah. LLaMA-7B can already run on a phone, and mobile RAM/AI compute is scaling pretty well.

I dunno if it will be "common," as most companies will keep their models and tuning in the cloud. There needs to be a concerted community effort to run stuff locally.