Hacker News new | past | comments | ask | show | jobs | submit login

One wonders what happens now that Stable Diffusion (etc.) is available.

It seems SD's advantage is more in newness/originality, I'm unsure it's actually cheaper right now - one needs to factor in the time/compute taken to get the good enough result. I presume Vogel is currently much more skilled in searching PD images than in running SD.




SD lacks humanity. Jeff’s art looks “bad” but bad looks human. His games are immediately recognizable, and that familiarity is friendly, comforting, and inviting. All tile art SD games are going to look the same and have a slight uncanny valley feel to them which will be off putting.


I'm not be able to tell pixellabs ai art apart from professional painted art.

Of course the best result is when the tool is combined with a good artist.

I'm 100% sure the price for game art will drop significantly, if it already hasn't.


Not sure I’d bet on it. With the right prompts (describing interesting combinations of styles and themes verbally) you can get SD to create some pretty interesting art. Haven’t tried models that are specific to tile art and pixel art though it would be interesting. But, you’re right in that of course there has to be a human involved when everything comes together in a game — that, I don’t think can or should be replaced.


There are also legal uncertainties around that. Someone would likely figure it out it Jeff himself wouldn't just blog about it. My impression is that some companies don't want AI made content on their platform. Which could mean an abrupt stop to one of the revenue streams, Jeff sounds like he depends on a lot.


Until recently valve was banning ai from their game store, so you're not wrong.


I’d rather buy Jeff’s games than ones made with AI art.


If you figure out the right prompts and the right model it feels like you could get a pipeline going with SD. You still might want to hire someone to turn those into actual assets, but now someone can create maybe 5 assets in an hour instead of one in two hours. If you get fancy you might be able to use things like inpainting to create tiles that blend nicely.


If you're going to care about metrics like "asset creation per hour," there already exist countless ready-made assets online, many even free to use. If you care about quality over quantity, then the value of an artist spending two hours on a single asset is obvious.

There is no aspect of asset creation where using AI to generate assets results in a better quality game versus human artists. Using artists to "improve" AI generated content is counteproductive because of the lack of direct control over what the model generates, something you don't need to deal with if a human is creating the art to begin with.

There are already tools available that let you create tiles that blend nicely. You don't need AI for any of this.

AI would be useful automating tedious and time consuming tasks. I've seen many artists say they could use an AI tool for automatically unwrapping UVs, generating spritesheets, JSON files, etc.. but the only thing AI ever seems to be used for is automating the creative process itself, and the results are at best technically competent (which is a result of simulating the work of human artists in the dataset) but never innovative or creatively compelling.


This is what irritates me about the current generation of AI. It is designed to give the appearance of "making humans unnecessary", when people mostly need help with drudgery and busywork.


> It seems SD's advantage is more in newness/originality

What? Originality? Where? Every time someone enthusiastic shows me a SD "creation" there are at least a few elements I've seen before. Sometimes I can point exactly where.


It's not that SD is so original (one could get creative with the prompts), it's that PD art/stock tends to the same bland look - perhaps if someone is looking at the right places it isn't? But I wouldn't know where.


SD is the same way, though. Things look superficially interesting and technically competent (from a distance, fine and intricate details are always a problem) but because the result is always an amalgamation, everything tends to look the same, poses are generic, faces inexpressive, the more you look the more you notice the broad similarities.

Arguably (and IMHO) the most interesting part of SD is when the models break down and you get surreal horrors beyond comprehension. And that isn't even creative, it's just fun to look at.

And of course none of that is even touching on the ethical problems of SD and other models being trained on copyrighted material, and primarily being used by corporations to push working artists out of jobs, which is not a problem with stock photos (until all stock photos, themselves, are AI generated which seems inevitable.)


Oh, stock photos will soon be "AI" generated ... by "AI"s trained on "AI" generated images. Infinite loop, baby!


PD art is less likely to pull from well known sources like GoT or the cover for Yellow Submarine, though.




Consider applying for YC's Fall 2025 batch! Applications are open till Aug 4

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: