[0x00cl]

Maybe the need/want data.

[impure]

OpenAI and most AI companies do not train on data submitted to a paid API.

[dortlick]

Why don't they?

[WhereIsTheTruth]

They also do not train using copyrighted material /s

[anhner]

If you believe that, I have a bridge I can sell you...

[dr_dshiv]

And it’s a massive distillation of the mother model, so the costs of inference are likely low.