[ux-app]

>could AI generated music ever replace Rancid or Junior Kimbrough or Fela Kuti?

AI is ok, but it will never ....

until it does, and then the goalposts will move.

the only logical endgame that I can see is AI replacing all human endeavour (creative, technical, physical, mental).

There will eventually be philosophers trying to make sense of the profound understandings coming from machines.

[davkan]

Ever is a strong word. I shouldn’t rule out a future like that. But I don’t see the through line from our current ai to one that has entirely supplanted all human creation.

That consideration seems more along the lines of worrying about the eventual need to escape earth than a future on a closer horizon worth worrying about.

I’m more concerned about how every facet of our children’s lives will become inundated with shoddy ai being used to extract maximum profits at the cost of any humanness, and the death of all genuine communication on the internet.

[zarzavat]

AI has, in a few short years, gone from hardly being able to string words together, to writing coherent grammatical sentences, to being more proficient than an untrained human in many cases. ChatGPT is way better at writing poems than me, for example. Its style transfer capabilities are out of this world.

Thinking that the progression is going to slow down is just wishful thinking.

[raxxorraxor]

It is more visible with image synthesis. Sure, the style can be freely switched in a few seconds, fitting compositions from trained concepts is very impressive.

But there still is no real creativity. No emerging concepts aside from complete accidents that cannot be replicated again. The same is true for the other direction with models like Clip could create an interpretation of generated images. It is impressive, but there are still clear limitations. You cannot expect linear growth here, it could be that the current AI approaches are wrong, we hit a plateau and need fully new approaches for significant improvements. What we now have is insane amount of data and more powerful hardware, it could be that we have years of iterative and slow improvement while people fine tune their models.

I think LLM have the same problem overall, it is just more difficult to notice.