Gene’s running MiniMax (MiniMax-M2, the open-weight model from Chinese AI lab MiniMax) as his local LLM, and says it’s the fastest option out there if your machine has 32GB of memory to work with — “that is the model that everybody points you at” at that size, per Gene. Other models may offer advantages at larger memory sizes, but MiniMax-M2 is specifically tuned for the 32GB sweet spot, whether you’re running it on a Mac Studio or a PC with an older Nvidia card. It’s free to download and run locally via tools like LM Studio or Ollama.
Model download: MiniMax-M2 on Hugging Face.
[Just Two Good Old Boys #182 32:51]
