[ttt3ts]

The problem with LLMs is the big ones don't run on consumer hardware. If 12GB or 24GB ram would run them you would see just as much interest.

The open source models will keep up as hardware allows IMO.