pull down to refresh
So then why do anything, like, why generate thoughts about the economy and post them in ~econ?
Because reality constrains you: there's a finite, limited set of ideas that are true or useful. Compare chess: there are infinite potential moves, but only a few make sense and of those some (or usually one) is best. There's a reality feedback when you make mistakes.
Literature, especially (science) fiction doesn't. It's just whatever, anything goes
Is he delusional, and if so, does that even matter?
What I think you mean is, could the LLM eventually help produce 1000x more Ian Bankses, dilute the market of speculative fiction so much that the stories become next to worthless because now we can all become giga-brained, elon-level imaginative and f-forward toward infinite abundance, universal high-incomes to a time when we all just plug ourselves directly into the sun to obtain god-level omnipotence?
There's a false analogy in your argument about chess, i.e. it doesn't exactly track with the question which was why write about ~econ? The reality feedback happens in the real world, and then is studied and written about by economists. Anything true, is only so because it is cogent with reality. Its usefulness is only derived by how true it is in a subjective manner.
Similarly, in fiction, the unreality of the subjects must also be true. The symbols and language for approaching them in great literature are more abstracted from reality than in the study of economics, but nonetheless they are true. If it were otherwise, they would not resonate with readers.
The argument, as I understand, is that the value of literature will trend to zero because machines will make more, better stories than humans. But I think they will only make better stories for robots. Imagine I said that because I am the superior intelligence, I can write/tell a better story for mice than Sir William Squeakespeare, the preeminent mouse storyteller. How could I tell a story for a mouse when I have never been a mouse?
But hasn't the mind always been infinite?
Yes, but here's the thing: unlike the mind, an instance of an LLM isn't infinite. The only infinity in LLMs is that you can create infinite LLMs, but each instance that you "chat" with, is finite, and not only that, it is static!
Thus, you can add a source of near-infinity, for example a random number generator of arbitrary size and feed that into your LLM (this is one of the main things that causes incoherent hallucinations in LLMs) or rely on "natural" randomness from GPU circuit timing, but no matter what you do, a human fiction (or non-fiction) writer will evolve with their story, whereas an LLM cannot evolve, nor can it retain that evolution. Only the next training round, if it incorporates the traces right, can cause that evolution.
But hasn't the mind always been infinite? Homer might have also opined that any pleb could imagine up a story like this, and then he would have not created. In fact, Homer might have been such a pleb.
So then why do anything, like, why generate thoughts about the economy and post them in ~econ? I would prefer to read your thoughts even if they are assisted by the LLM than pure and unabashed slop.
Why secure your wealth with SHA256 encryption? Surely quantum will crack that too, one day?
The point I am trying to make is, if there is no threshold on human accomplishment, and it is in fact infinite, then perhaps you are right, and business case for monetising good fiction might actually be shot. I don't think that is most writers' raison d'etre, and I do believe there is a cosmic balance that upholds work that has integrity, that somehow makes it worth its while.
Why do you climb mountains? Is it not possible to fly to the peak with a helicopter and enjoy the same view? You probably get something out of the struggle, which I think storytellers do as well. I can accept that the majority of fiction might be cooked now, like the majority of any creative content.
I'll level with you, I don't have a lot of experience in publishing, so I haven't worked out the economics of it. I probably should.