MS is switching its billing system from a flat subscription rate to a token-usage system that has the potential to bill users at a significantly higher rate.
MS is switching its billing system from a flat subscription rate to a token-usage system that has the potential to bill users at a significantly higher rate.
Of course it has “the potential to”, that’s the whole fucking point of this change.
The compute hardware required to give you that obsequious-chatbot-on-demand is phenomenal. AI as a whole has been burning billions upon billions of dollars getting to these lofty heights where it’s alleged that everyone’s job is threatened, any day now.
And now’s the time to get to the other side of the equation, the dreaded return on investment, where people have to pay what it actually costs. Plus a modest amount for that actual return of course, that’s the thing that all those investors actually want.
Whether Microsoft has actually managed to get enough people hooked on the particular brand of crack they’re selling is questionable. In any case it’ll be fun to watch at least.
Not sure it’s going to be all that modest. A lot of the data centres have been financed using debt, plus all the GPUs inside are depreciating hard, plus the rate needs to cover the cost of training and workforce, plus they won’t be running at 100% capacity all the time so they’ll need a bit extra to cover low-utilisation periods, plus this has been sold as the next big thing and the saviour of the world economy, so modest returns simply will not do.
Not going to be so much “inference costs plus ten percent” as “starting at inference costs times ten”. Personally, can’t wait for the bubble to burst and the whole thing to collapse, but I’ve got my schadenfreude ready for when the big bills start to arrive.