I think the parent is referring to the fact that LLMs and generative models are, for the time being, “creating new markets” and “creating new problems”, but not really solving any of the hard problems we already have. I see it this way: you have this incredible piece of technology. You can use it to take the scraps away from a bunch of starving artists, and convert that survival income in wealth for some entrepreneur. Or you can use it to cure cancer or aging or make fossil-free airplanes. At the moment, the first application seems to be the one moving the market, and that’s really sad.

On the side of hope, if somebody were to demonstrate how to use a bunch of GPUs to make the e-coli bacteria live for a little longer using an approach that has a semblance of generality, I guarantee that a lot of old people would be willing to part with their fortunes in exchange for a sliver of hope for them or future generations.

You have a deterministic AI program to show me?