@fasterthanlime Suggesting proprietary Nvidia drivers as an alternative to Mesa because of LLM assisted contributions is quite an irony. I'm still very critical of applying LLMs to larger parts of a codebase but whoever wrote this is lacking any reflection.