No no no use it more, make sure to use up as much tokens as possible. They do inference at a loss
> They do inference at a loss
They don't, inference is cheap, especially for agents because of cache hits. The API prices are just inflated.
Ive got a 'Claw interfacing with OpenAI and generating garbage questions and responses. I have an 8k context on mine.
Deletion with OpenAI isnt really deletion. So I'll waste their resources AND train on low quality slop on my side.
My work degrades theirs.
This makes no sense, their value in the marketplace is in usage and inflated promise, not actual revenues