I have more access to information now than the most powerful people in the world did 40 years ago. I can learn about quantum field theory, about which pop star is allegedly fucking which other pop star, etc.

If I don't care about the law I can read any of 25 million books or 100 million scientific papers all available on Anna's Archive for free in seconds.

“It won’t be any more wickedly cool than the internet” - saying something won’t be any more wickedly cool than the most profound and impactful pieces of infrastructure human civilization has erected is a pretty high compliment.

Those millions/billions of people matter a great deal.

For what it’s worth I never use midjourney or dalle or any of the commercial closed systems that steal from artists but I know I can’t stop the masses from going there and inputting “give me pretty picture in style x”

And I also agree that we shouldn’t build systems that alienate people from that accumulated equity.

I'm sure OpenAI's models can shit out an approximation of a new Terry Pratchett or Douglas Adams novel, but nobody with any level of literary appreciation would give a damn unless fraud was committed to trick readers into buying it. It's not the author's work, and there's no human message behind it.

Also, I'm sure that you can eventually just prompt the model with the message you want to put into the story, if you can't already do that.

But that wouldn't be helpful. It would've been memorable, because novels aren't just about the message.

AFAICT current text generation is something approaching bad mimicry at best and downright abysmal in general. I think you still need a very skilled author and meaty brain with a story to tell to make use of an LLM for storytelling. Sure it’s a useful tool that will make authors more effective but we are far from the point where you tell the LLM “write a story set in Pratchetts Discworld” and something acceptable or even entertaining will be spit out - if such a thing can even be achieved.

Given all that, the purpose of LLMs should be to create tailor made content to everyone's tastes. However, it seems the hardcore guardrails put into GPT4 and Claude prevent it from generating anything enjoyable. It seems, even the plot of the average Star Wars movie is too spicy for modern LLM sensibilities, never mind something like Stephen King.

There is somewhat famous digital artist from Russia - Alexey Andreev. Google it, he has very distinctive style of realistic technique and surrealistic situations, like landing big manta ray on the deck of aircraft carrier. Or you can see his old works in his 5-years-not-updates LJ [1].

Now he uses generative AI as one of his tools. As Photoshop, as different (unrealistic!) brushes in Photoshop, as other digital tools. His style is still 100% recognizable and his works don't become worse or more "generic". Is he still artist? I think so.

Where will you draw the line?

[1] - https://alexandreev.livejournal.com/

Was it irony? :)

"A person who creates art (such as painting, sculpture, music, or writing) using conscious skill and creative imagination"

Am I no longer a software engineer?

When everybody is an artist, then nobody will be one.

I think population tends to value "looks pretty", and it's other artists, connoisseurs, and art critics who value origin and process. Exit Through the Gift Shop sums this up nicely

According to Marx, value is only created with human labour. This is not just a Marxist theory, it is an observation.

There may be lots of over-priced junk that makes you want to question this idea. But let's not nit-pick on that.

In two years time people will not see any value in AI art, quite correctly because there is not much human labour in creating it.

And yet it's completely and absolutely wrong. Value is created by the subjective utility offered to the consumer, irrespective of what inputs created the thing conveying that utility.

Yes, I'm aware. This is precisely why I'm stating the prior comment to be "absolutely wrong". Marginal utility is a substantially valid model, LTV is not.

> In fact, there are also other value theories in economy. It's a mostly philosophical choice, and like other philosophical choices, it's not possible to accuse one of them of being wrong.

Sure it is. These aren't theories in a normative sense, they're models of causality for manifest phenomena. They're closer to scientific theories than they are to philosophical axioms. LTV simply doesn't bear out with observation.

Yeah? Well, you know, that's just like uh, your opinion, man

But you pay that for any piece of art though? You appreciate it because you like what it looks like. The utility of it is in how good it looks, it's not how much effort was put into it.

If you need a ditch you're not going to value the ditch more if the worker dug it by hand instead of using an excavator. You value it based on the utility it provides you.

Again, i’m not saying buying a mass made AI art will be wrong. Just personally speaking, it will never evoke any feelings other than “looks neat” for me. So its inherent “art value” is close to 0 as I can guess its history is basically someone put in a prompt and sent it to print (which I can do myself on my phone too!). It’s the same as looking at cool building pics on my phone (0 art value) versus actually seeing them in person (non-0), mostly because the feelings I get from it. That being said, if it makes others happy, it’s not my place to judge.

And in the process, they will obviate the need for Nightshade and similar tools.

AI models ingesting AI generated content does the work of destroying the models all by itself. Have a look at "Model Collapse" in relation to generative AI.