At this point I think the burden of proof should be on ML people to explain why that shouldn't be my goal. Every person in that industry has stated that the technology is going to displace 30% of jobs within the decade. Why should anybody support something which is certain to absolutely destroy society while simultaneously having no positive payoff?

So it's not really LLMs you want to see fail, it's the businesses people are building around them. I can sympathize with that for sure but to me it seems obvious that there's many potentially beneficial applications of the technology that don't involve taking jobs.

Right, though I don't see much distinction given that the cost of training the models is so high. The existence of the technology is equivalent to the existence of these businesses, is it not?

Suppose the technology advances no further than it is today, and I'll even grant up to an order of magnitude improvement. I'm not interested in speculating over so-called AGI though. What would you point to as an overwhelmingly positive justification of its existence? I won't argue against it. I'm just curious.

AlphaFold 2. It's not an LLM but it's not that different technologically. EDIT: that last sentence is actually a bit unfair to AlphaFold 2 as I consider it to be much more impressive technology than LLMs, but it comes from a common lineage.