Hacker News Reader

Comment by ljm | Hacker News Reader

Literally nobody will give a fuck about a 24 episode series that exists because you spent a few seconds writing a prompt.

AI doesn’t increase the value of content, it makes it meaningless by destroying scarcity.

Tea. Earl Grey. Hot.

View on HN

Mainly just pulling from how my friends use it now, for the fun of it. We have sent each other silly generated videos of stuff we did or talked about. Kind of like how we used up compute, storage, and bandwidth to send memes and funny videos. If it were the 80s we would spent proportional cost to share a cat picture perhaps, but look at what we did with the bleeding edge even then. Turned the computers into game machines, shared things, wasted time. This is just an extension of that idea. Everything cutting edge ends up getting used for silly things.

I think they are talking about making their own series, not selling something, which will be valuable to exactly one person (or in this case a group chat). Which is awesome.

A whole series would involve writing thousands of prompts, not just a few seconds. But since you bring it up, one thing I think we can eventually expect is an avalanche of fan created continuations of cancelled shows, like an 8th season of Star Trek: The Next Generation, or a decent 8th season of Game of Thrones. Instead of writing a tie-in novel, aspiring genre authors will write screenplays and turn them into AI movies.

Which is the worse thing that cn happen to these shows. Most of them are already going downhill plotwise after a couple of seasons.

> Which is the worse thing that cn happen to these shows. Most of them are already going downhill plotwise after a couple of seasons.

Firefly is the exception. Luckily they managed to wrap it up in Serenity.

I'm not at all sure you can make a blanket statement like this and expect it to be always true. This is classic Disruptive Technology stuff, and I'll spare everyone the tedious explanation of how the typical trajectory will go.

Yes there will tons of garbage, but of course there will be the 1% of the 1% who spend the time and efforts to, perhaps, create something of equal and perhaps better quality than what has come before.

> AI doesn’t increase the value of content, it makes it meaningless by destroying scarcity.

Most art is in some way or form derivative of another work or a combination of somesorts. I enjoy derivative works. To me, AI already has value. I don't seek endless gratification through originality.

Like the hot take but it is needlessly negative because it doesn’t go far enough.

You could make the same argument about musical instruments, or being able to record and playback music, dj tools, etc.

I think what you get is the power law distribution for tons of content. Some of the stuff is still mega valuable, but distribution just gets more and more important and it’s harder and harder to break through. This is what the “democratization” of any previously difficulty-gated endeavor does.

More niches will be created, more fragmentation in tastes, stuff like that. Not just completely valueless content.

Incumbents and platform providers get to win through it all though, because humans still want to fill their time.

The difference is that the jury is still out if these things are truly generating something novel. Or are we just making it pick from the spaces in between an average of all the images on the internet.

The latter would mean everything quickly looks and feels the same.

I also believe some barrier to entry is needed for good art, people are inherently lazy and AI lets us get away with "just good enough" - see llm assisted coding as an example

> The difference is that the jury is still out if these things are truly generating something novel. Or are we just making it pick from the spaces in between an average of all the images on the internet.

Novel in what sense? Most creative endeavor is not completely novel -- it's usually a combination of pre-existing styles, taste, patterns, and other inputs.

> The latter would mean everything quickly looks and feels the same.

I think human preferences will be a good guard rail against this. The reward will be greatly reduced for creating mostly the same things, just like real artists in real life.

> I also believe some barrier to entry is needed for good art, people are inherently lazy and AI lets us get away with "just good enough" - see llm assisted coding as an example

That's a pretty deep philosophical question -- "good art" is also a hell of a rabbit hole to go down!

Guess we'll find out :)

Ironically I think it might, by serving as a synthetic generator to learn to better understand artistic values for those who need it. We wouldn't know what is ever so slightly wrong about replicator Earl Grey, without trying it for a while.

A lot of people right now claim that they cannot tell apart AI output from human art, while many of them seem to grow rather agitated and stressed after repeated exposure. I bet they're going to be forced into exclusively making art manually, or viewing exclusively human art at some point, and through that ways, AIs could increase value of human made data.

I find it funny that wishes for AI arts always seem to be more anime and end to Marvel slop. They want human slop go away thanks to AI slop? I'm not meaning to call out someone for contradiction or inconsistency - as I do understand the sentiment. But it gives me chuckles.

someone's slop is someone else's gold or something.

everyone likes their own slop, hopefully

Reply on news.ycombinator.com