I think it’s wild that people here choose to compare open-source AI models that run on a single GPU to the behemoths at OpenAI and use the exact same arguments against both despite both not being the same.