Unified memory on Apple Silicon. On PC architecture, you have to shuffle around stuff between the normal RAM and the GPU RAM.

Mac mini just happens to be the cheapest offering to get this.

Local LLM is so utterly slow even with multiple $3,000+ modern GPUs operating in the giant context windows openclaw generally works with that I doubt anyone using it is doing so.

Local LLM from my basic messing around is a toy. I really wanted to make it work and was willing to invest 5 figures into it if my basic testing showed promise - but it’s utterly useless for the things I want to eventually bring to “prod” with such a setup. Largely live devops/sysadmin style tasking. I don’t want to mess around hyper-optimizing the LLM efficiency itself.

I’m still learning so perhaps I’m totally off base - happy to be corrected - but even if I was able to get a 50x performance increase at 50% of the LLM capabilities it would be a non-starter due to speed of iteration loops.

With opelclaw burning 20-50M/tokens a day with codex just during “playing around in my lab” stage I can’t see any local LLM short of multiple H200s or something being useful, even as I get more efficient with managing my context.

But the only cheap option is 16GB basic tier Mac Mini. That's not a lot of shared memory. Proces increase bery quickly for expanded memory models.

Sure, but aren't most people running the *Claw projects using cloud inference?