Comment by _DeadFred_ | Hacker News Reader

Comment by _DeadFred_ | Hacker News Reader

Parent Full thread

Your meticulous prompt is using the work of thousands of experts, and generating a mashup of what they did/their work/their commitment/their livelihood.

Their placement of books. Their aesthetic. The collection of cool things to put into a scene to make it interesting. The lighting. Not yours. Not from you/not from the AI. None of it is yours/you/new/from the AI. It's ALL based underneath on someone else's work, someone else's life, someone else's heart and soul, and you are just taking it and saying 'look what I made'. The equivalent of a 4 year old being potty trained saying 'look I made a poop'. We celebrate it as a first step, not as the friggen end goal. The end goal is you making something uniquely you, based on your life experience, not on Bob the prop guys and Betty the set designer whose work/style you stole and didn't even have the decency to reference/thank.

And your prompt won't ever change dramatically, because there isn't going to be much new truly creative seedcorn for AI to digest. Entertainment will literally go into limbo/Groundhog Day, just the same generative, derivative things/asthetics from the same AI dataset.

And that's exactly how your brain work. What you call "creativity" is nothing more than exactly that: mixing ideas and thoughts you were exposed to. We're all building on others' work. The only difference is that computers do it on a much larger scale. But it's the very same process.

This is completely absurd and reductive point of view, which I always assume is a cop out. Just because it's called "machine learning" doesn't mean it actually has anything to do with how human learning or human brain works, and it's certainly not "exactly how" or "very same". There's much more going on on in human creative process, aside from mere "mixing": personal experience, understanding of the creative process, technique and style development, subtext, hidden ideas and nuances, etc. Computers are very good at mixing and combining, but this is not even close to what goes into actual creative process. I hate this argument

> personal experience, understanding of the creative process, technique and style development, subtext, hidden ideas and nuances

All of these are just human being exposed more to life and learning new skills, in other words -- having more data. LLM already learns those skills and encounters endless experience of people in its training data.

> I hate this argument

That's very subjective. You don't know how the brain works.

You seem to not fully understand the quote. LLMs learn patterns/noises from an existing output, these are not skills nor endless experience learned. It's like saying you learn how to make a cake by learning how it should look like not how it's composed. LLMs mock the studio ghibli style images, didn't invent the style or learned the endless experience the studio accumulated over the years. In fact it's a mocking of the images and it just looks horrible

It's not just more data, it's deeper understanding of the fundamentals, of the idea and of the tools used, as well as the process of creation itself. It's what makes studying art interesting: why did author chose to do this and that, what's their style, what was the process, etc. For LLM the answers will universally be "because it was in the prompt and there was appropriate training data" and "the author prompted the model until the model returned something tolerable". You may argue that not all art has or needs depth, or that not all people are interested in it, but that doesn't mean that we should fill our cultures with empty boring slop.

> That's very subjective

I was expressing my opinion of this argument which absolutely is subjective

> You don't know how the brain works.

Neither does grandparent comment's author, didn't stop them from making much bolder claims.

But you aren't being creative here. Just using the 'average' of tons of actually creative peoples work to create an 'average' computer predicted scene. The opposite of art. Warhol already did it and did it better.

If I see a painting, I see an interpretation that makes me think through someone else's interpretation.

If I see a photograph, I don't analyze as much, but I see a time and place. What is the photographer trying to get me to see?

If I see AI, I see a machine dithered averaging that is/means/represents/construes nothing but a computer predicted average. I might as well generate a UUID, I would get more novelty. No backstory, because items in the scene just happened to be averaged in. No style, just a machine dithered blend. It represents nothing no matter the prompt you use because the majority is still just machine averaged/dithered non-meaning. Not placed with intention, focused with real vision, no obvious exclusions with intention. Just exactly what software thinks is the most average for the scene it had described to it. The better AI gets, the more average it becomes, and the less people will care about 'perfectly average' images.

It won't even work for ads for long. Ads will become wild/novel/distinct/wacky/violations of AI rules/processes/techniques to escape and belittle AI. To mock AI. Technically perfect images will soon be considered worthless AI trash. If for no other reason than artists will only be rewarded for moving in directions AI can't going forward. The second Google/OpenAI reach their goal, the goal posts will move because no one wants procedural/perfectly average slop.

artist have a style,you can see a work of art and know who made it, with these AI images its all random all over the place no direction, they can call them self's artists but i will never see them as that

Reply on news.ycombinator.com