You are kind of glossing over the B2B market where contract pricing is basically just MBA vibes and the fact that people don't really care necessarily about the performance of the language model once it hits a baseline. They care about how it integrates into their lives. Precisely where first mover advantage comes into play. Having to train a language model all over again is it's own sunk cost.

▲

dofm an hour ago | parent | next [-]

OpenAI themselves said, in their revenue projections, that they expect the consumer vs enterprise revenue split to be 50:50, though — see:

https://news.ycombinator.com/item?id=48451053

	▲	hparadiz an hour ago \| parent [-]
		That's actually a really good sign if they are getting that many consumer subs. I was expecting it to be more like 1:4.

▲

shimman an hour ago | parent | prev | next [-]

B2B absolutely cares, especially since ZIRP is unlikely to ever come back in our lifetimes. No sane corporation is going to continue to throw billions down the drain with nothing to show for it.

	▲	hparadiz an hour ago \| parent [-]
		Most people using ChatGPT aren't using it for coding. They are using it for writing emails, working with spreadsheets, doing research, and writing reports. They could not care less about the coding aspect of language models. ZIRP in this context is meaningless. It's just another expense for every law and accounting firm. There's an entire world beyond tech jumping into this stuff right now. Like to give you perspective on this. I called my aunt in Germany who is an almost retired MD and she was the one that brought up ChatGPT and Claude to me.

▲

cyanydeez an hour ago | parent | prev [-]

b2b needs actual ROI and that's no where near. CEOs would be yelling loudly if this were returning them cash instead they're just jettison people to afford the bills.