Post #2763944

I don’t think it’s actually going to be doing a lot of LLMing. It has a coprocessor for inference type stuff, but only 8GB of RAM on the SOC. So I assume they just stuck this in for marketing.