Pewdiepie's "Odysseus" local AI model - thoughts?

Local models are definitely hardware-dependent, but it comes down to what you actually want to do with them. Even a mid-range setup handles a lot — something like Qwen3.5 9B runs smoothly on a Mac mini with 16GB RAM for most everyday tasks. As your needs grow, so does the hardware requirement. Personally, I run my local document RAG app Informity AI with Qwen3.6 35B on an M5 MacBook Pro with 48GB RAM and it’s rock solid.