• I have a 2018-era MacMini with 64Gb of RAM + negligible GPU, running a 35 billion parameter model and I’m getting 4 tokens per second out of it; tell me again how we aren’t going to be running AIs locally at home?

  2026/03/07