30001st poster gets a cookie (cookie thread (Part 7))

just for the record I’m pro-ish AI art (in the sense that I think it should be ~allowed not the sense that I want to maximize it) but Strongly Anti Sam Altman (for other reasons)

4 Likes

imagine your last name being Altman

couldn’t be me

2 Likes

OMG I LOVE YOUR AVI SO MUCH

SHES JUST LIKE YOU. SAME VIBES. LAID BACK. BOOZE. ALCOHOL. GUN. SHOOT SHOOT. DANGEROUS. MYSTERY. AAAAAAAAAAAAAAAAAAAAAAA

1 Like

This assumes demand exceeds supply. If they’re already producing the amount of cars they think they’ll sell in a given time frame, they would literally just be cutting jobs or pay.

2 Likes

tutuu keep this up and her brain will fry from having to do the same task over and over

1 Like

oooh is that why everyone likes me here instead of killing me?

1 Like

hahahahaha

it would be a whole lot of work yes

1 Like

Where ai images becomes commercially viable, I don’t see commissioning artists bankrupting companies wanting to use the technology

imo the technicalities will become that an entire sub-industry is born where companies form specifically to hire artists and deal with training AI models to sell to clients

2 Likes

shame i like anko too much

1 Like

I mean yes
if supply manages to outweigh demand then that’s a net positive

2 Likes

I don’t think it will
but if it did its not a bad thing

1 Like

Unless you’re sourcing your labour in an unethical way, you’re better off just commissioning a piece. Unless the model already exists, in which case… the model already exists.

The amount of training data you need for a model is very high. Training an ai off of a smaller quantity of data is tldr using a partially trained ai.

There’s a reason it takes 10gb vram to run the whole ass model.

sure
I will reiterate I do not think commissioning of artists to make pieces independent of ai imagery will die even when ai imagery becomes commercially viable as well

I’m not sure you’re grasping the necessary size and diversity of data required to train and improve LLMs. And the issue sort of is that the art produced by AI is, by definition, worth less than art produced by a human. It being cheaper is literally the only reason corporations would bother using it. If you want to then say that this could turn into other kinds of media being produced faster like games, movies, TV shows, book, etc. well that gets devalued too because the market gets saturated. There have to be more books published in a year than I’m going to read in my lifetime. If every jackass can spit out a novel’s worth of text in an afternoon, there’s going to be just mountains of them that are roughly the same quality and equally worth practically nothing.

1 Like

supply of art already substantially outstrips demand significantly*, it would surprise me a lot if AI image generation made that less true

* or, like, you can play word games about literal meaning here. I mean something along the lines of “there are a lot of people who would like to be paid to make art, way more than the market for people who want to buy art can support at anything even slightly resembling a wage the artists could live on, with a few exceptions at the very top of the range arts careers in full generality tend to be poorly-paid and very competitive”

3 Likes

I’m not suggesting that these hypothetical artist AI commissioning companies would be small

1 Like

Yeah. How does a person get good enough to be as good as our best artists now if they can’t survive on making art when they’re mediocre? Like when people ask why we don’t have artists churning out hyper realistically marble sculptures anymore, it’s because rich fuckers aren’t finding talented artists when they’re still young and inexperienced, and paying them to be their personal art goblin for the rest of their life.

2 Likes

How

also because who the fuck in 2024 cares about having a hyper-realistic marble statue

1 Like