Sounds like a money making strategy. Also, given how expensive all this shit is if inference costs _more_? That’s not cheap to me.

But again the original argument was that they can run forever because inference is cheap, not cheap enough if you’re losing money on it.

Even if the current subsidy is 50%, gpt would be cheap for many applications at twice the price. It will determine adaption, but it wouldn’t prevent me having a personal assistant (and I’m not a 1%er, so that’s a big change)