If you spend more money training the model and offering it as a service (with all the costs that that entails) than you earn back directly from that model's usage, it can only be profitable if you use voodoo economics to fudge it.
Luckily we live in a time period where voodoo economics is the norm, though eventually it will all come crashing down.
You’re right, but that’s not what’s happening. Every major model trained at Anthropic and oAI have been profitable. Inference margins are on the order of 80%.
That’s true, but OpenAI and its proponents say each model is individually profitable so if R&D ever stops then the company as a whole will be profitable.
The problem with this argument is that if R&D ever stops, OpenAI will not be differentiated (because everyone else will be able to catch up), so their pricing power will disappear, and they won't be able to charge much more than the inference costs.
You're missing that they're pricing the value of models progressing them towards AGI, and their own use of that model for research and development. You can argue the first one, and the second is probably not fully spun up yet (though you can see it's building steam fast), but it's not total fantasy economics, it's just highly optimistic because investors aren't going to buy the story from people who're hedging.
LLMs by themselves are not going to lead to AGI, agreed. However, there's solid reasons to believe that LLMs as orchestrators of other models could get us there. LLMs have the potential to be very good at turning natural language problems into "games" that a model like MuZero can learn to solve in a superhuman way.
If you've been around, imgur, basically. All the image hosting solutions before it (imageshack) and all the file hosting solutions before them. Yahoo.
"Crashing" in this context doesn't mean something goes completely away, just that its userbase dwindles to 1-5-10% of what it once was and it's no longer part of the zeitgeist (again, Yahoo).
the only way I can see that happening here (userbase dwindles to 1-5-10% of what it once was) is if another better service came about but I don't think that is what @bayarearefugee is talking about when s/he said 'come crashing down' ?
Luckily we live in a time period where voodoo economics is the norm, though eventually it will all come crashing down.