Comment by yalogin
If deepseek and qwen are capable why does one need to use OpenAI models? Are their models really that much better? If not the only they bring is the hosting service. N that scenario how long do they have this advantage before aws or Microsoft takes over?
> If deepseek and qwen are capable why ...
Of course, you can go further and run qwen locally. Or even train your own nanogpt. Why not if it's capable, right? And this 'if' is a big question.