I wonder if it can run the 4GB model? @nomic_ai gpt4all uses llama.cpp and can run on CPU so it might still work

→ View original post on X — @simonw