A practical guide to AI image generation, from diffusion to agents
danielball.com
danielball.com
I actually experimented with creating a set of skills around rendering early AGI style EGA graphical scenes, and it worked surprisingly well.
So rather than storing full bitmaps, you instead store sequences of simple drawing instructions designed to recreate some specific scene - like you would an SVG.
As a result you end up sidestepping a lot of the problems with larger generative models, which can have issues with grid alignment, palette choices, and high-frequency noise.
I’m super excited to see more LLM-driven procedural graphics stuff in the future!
It works okay but you'd probably have to do a little bit of coaxing to get it to come out exactly right.
>As a result you end up sidestepping a lot of the problems with larger generative models, which can have issues with grid alignment, palette choices, and high-frequency noise.
All of these can be avoided with simple prompts and don’t require all this ceremony.
Convo then the artifact from the convo (sorry, apparently no way to share the pics but they are just captured from the artifact).
Convo: https://claude.ai/share/78cd63c8-ffdd-4569-8de2-ac8ab5c5cfe0
The difference i see is the shading. Could easily tell it to do flat fills instead.
Regardless it has none of the supposed problems of AI generated pixel art.
Learn to draw.
Then ask the model to restyle the image.
You can get wonderful photographs with the exact composition and scene you want for all sorts of use cases.
Second hack.
Model the scene you want from low resolution geometry (boxes, spheres etc) for the composition. Then use that to ground the generated image (ie give it as reference) and ask the model to re-imagine it in the style you want.
Succesfull artists don’t necessarily develop the pieces beyond conceptualization and editorial control.
See for instance Damien Hirsch. Or the academic painters with their own workshops.
AI can handle large amounts of routine tasks. Like converting an image to a 3D mesh. I don’t frankly see the difference between that and Damien Hirsch giving his craftsmen a task to deliver a piece.
If you talk to modern art critica you will end up with argumentation where the concept of the piece is what matters not the execution or it’s quality.
If original concept is all that matters very hard to see how AI generated steps could not be part of that pipeline.
Nobody is claiming getting random stuff from generic non-specific prompting counts as art.
But what if you do an original painting or drawing; then have AI convert it to 3D mesh and print it - and pass it off as art? That is a non-obvious situation.
But that's not really how art market values the merchandise.
Conceptualization suffices for the market after sufficient prestige and brand value exists. I guess the Damien Hirsch reference was not that obvious for this venue so -
https://www.businessinsider.com/why-damien-hirst-is-controve...
Note: I'm not saying you should measure the value of art with money. I'm saying those people _who measure the value of art in monetary terms_ may value branding and conceptualization more than manual skill.
I'd like to ask you a question. I'm asking in good faith. Here are a few random examples of images people create with AI:
https://civitai.com/images/146193427
https://civitai.com/images/17108902
https://civitai.com/images/63175621
https://civitai.com/images/10703900
https://civitai.com/images/28673784
https://civitai.com/images/59645082
https://civitai.com/images/97017376
https://civitai.com/images/63656891
https://civitai.com/images/99196859
Fundamentally, what makes art, well, art?
Is it the physical act of drawing it manually? Surely not, because generative art[1] (the original non-AI kind) is a widely recognized thing.
Or maybe it's the amount of effort that's spent creating it? Surely not, because paintings as trivial as a simple black or red square are considered art[2].
Or maybe just the very act of self-expression makes something into a piece of art?
Here's Wikipedia's definition:
> Art is a diverse range of human activity centered around works, crafted as an expression of creativity or imagination, which are expected to provide an audience with a worthwhile experience—generally, an experience of emotion, thinking, or beauty.
So is the thesis that, no matter how it's used, and no matter how much effort is spent, an image created with the help of generative AI can never satisfy this definition?
They are random noise pastiches.
Art requires authorship.
I mean that's my definition :)
The author needs to have sufficient level of artistic skill and talent to be able to produce pieces with artistic merit. Even if they just prompt it - which they likely wont because prompt based outputs alone generally suck.
My definition for sufficient level of skill where AI tools become useful as a form of expression with true artistic merit, would be that you must be able to generate the same image with traditional media without AI:s help
If it turns out there was deep authorship in these works on some level (other than just running through random prompts) then the situation becomes more nuanced.
You should be able to tweak each individual detail with exact authorship. And _understand_ what you are doing.
So if an established painter, photographer or cinematographer uses prompt based tools, there the tools become part of expressive workflow that likely has higher artistic standard.
You can't really develop this sensitivity with prompting alone (generally).
_General_ artistic value is a DIFFERENT thing from doing something a person personally enjoys, though!
It's totally fine to enjoy prompting images.
People enjoy doing puzzles, building legos, practicing piano, playing computer games etc.
Ie engaged in activites which they enjoyu without external valuation mechanism involved.
These things are called _hobbies_. And it's awesome to have hobbies whatever they are.
I don't think the author is posing as a professional. These models are like having a 3D printer at home. Very handy in a pinch, but what they make doesn't replace the real thing.
The problem isn't that they look shitty to pixel artists or experts, they look shitty for anyone with a bit of observation.
It's junk art the same way Mc Donald's is junk food.
I think we could do a bit better here with the explanation of what looks bad to you..
- the magenta background is still visible along the edges - the color palette is all over the place - random noise pixels of totally different colors are visible scattered around - the lighting and shading doesn't look that good
There's plenty of real food that is as heavy and as salted, and as such no healthier, than junk food. But it actually tastes good. (I can't exactly put the finger on what makes an Italian pizza better than a US 7/11 one, but it's not about the former being healthier).
McDonalds', over time, will make you less healthy.
AI slop over time will perhaps change aesthetics, but hard to chalk that up to anything other than personal preference.
I just don't get the complaints about "use proper art" when it comes to games. Most of the time the game has some other, far more glaring, visual issues (anti-aliasing, clipping, icons-not-matching-items).
I have a suspicion that when you're looking at (and generating, because you then look at the results) AI generated images and text more you get desensitized and are more likely to say that something looks okay when to other people it won't.
How come pretty much every modern UI is garbage? Are the devs just lazy? Like google.com on mobile does not have parity with the desktop version (there's no "set specific date range option).
In anything you do, there's a point where you've got to say "okay, this is good enough" and ship it. In fact, I would argue that using pixel art itself is "settling" in the majority of cases.
That's the wrong layer to tackle this problem. Or I should say: that's necessary, but not sufficient. Yes, the model should be trained to roughly generate something which looks like pixel art (i.e. with a concrete pixel size, reduced/consistent-ish palette, etc.), but enforcing that should be done deterministically in post-processing.
you either trust and/or verify. the people clearly cannot verify and clearly do trust. vagueposting about ai bad won't break that vacuous trust, on the contrary, due to the incentives involved, induces distrust on the human expert side.
so maybe you wanna actually elaborate?
The worst result would be if these models were any good vs being absolute slop. Then human artists would need to worry. But in fact there is no need to worry at all, as surely people (aside from a few people with bad taste) will clearly see the difference.
― person who has never attempted to program for a specific retro system.
Peep the graphics hardware limitations of the Nintendo Famicom/NES, for example: https://noelberry.ca/posts/nes/index.html
You can see these limitations guiding the designs of actual retro games if you know what to look for.
Example: In CAPCOM's Mega Man (1987), the famous “Yellow Devil” boss fight plays the way it does because only the moving chunks of the Devil's body are composed of Sprite tiles. When one of those chunks reaches the opposite side of the screen, that part of the Devil's body becomes drawn with Background tiles instead: https://www.youtube.com/watch?v=8RbPqqPrL_M
This is also why the Devil, Dragon, and Wily boss fights (anything larger than a regular Robot Master) take place entirely in a black void — the Background layer is exhausted drawing the bosses and there's nothing left over for actual play-area background.
For the tasks where several control nets should be adhered (depth and canny and segmentation) dreaded SDXL is still unmatched. SDXL also the only one handling influence weight for each control net normally, this ability completely vanished on modern models
Control nets in early post-SDXL models like ZIT or Flux kind of work too, but still much worse in quality when combined. I suspect pure diffusion architecture of SDXL is really favorable to mixing control signals with precision, DITs probably can be trained to similar level for mixing signals - but it is much harder and never seen in practice. At least every post-SDXL model I tried is trained with single one in mind, judging from my attempts. This rant is from personal expirience.
This part's a little confusing to me. What does "suffix" mean in this context? (Is it referring to just like the style-instructions boilerplate text you might include with each image?)
And then also: does locking a set seed really have an effect one way or other for consistency? Wouldn't changing the prompt text even if only slightly have just as much a randomization effect as changing the seed, making fixing it or using a random seed both functionally the same? Or is that not how the latest models work?
Just visit Civitai, look at the pages for the base models, filter for images with the ComfyUI workflow in them, then you can just download and drag and drop them onto your local ComfyUI app.
It won't bridge the gap to gpt-image-2.5 though, and it's going to slow down the process further.
With a 4090, Qwen Image 2.1 Turbo takes around 15s per image, depending on the resolution. That's without the prompt enhancer.
At the risk of stating the obvious, you can't really compare local models to proprietary frontier models. For all we know, those models are heavily subsidized and have trillions more parameters than the smaller, modest ones you can run on a single GPU.
I've been running a prompt-adherence benchmark site [1] all the way back since the days of SDXL. If you look at the progress dropdown, I do think the gap is closing.
June of this year saw the release of some of the biggest open-weight bangers (Krea 2, Ideogram 4, and the kind of slept-on Boogu-Image 0.1), though the license terms can be kind of meh.
The best locally hosted model scores 8 out of 15, while the proprietary ones score 12 out of 15 on the text-to-image metric. If you want something that can reasonably decent for local gen while still being relatively small for modest GPUs, take a look at the recently released Moss-Image model [2]. But if you're expecting results at a level of something like GPT-Image-2.5 Sunburst, you're going to be disappointed.
The technical ability of anthropic models continue to steadily improve, but I still find the creative and compositional choices to fall far short of anything I would ever use.
If we want to use a pixel art generator commercially we need 'practical guide' to making a small change to one animation, two months after creating the original character.
We need images generated in psd layers, with actual text using a font.
Sticking to the brand palette religiously.
These thing may reach your standard consistent at some point, maybe. It's not important since before and after this point, they will require "messing around" to produce something useful but they can produce something if you need "good enough" but if you demand actually good, they won't arrive ever.
Generating grid is cheap but inconsistency creeps up very quickly.
AI-generated posters don’t have to be horrible, 990 comments https://news.ycombinator.com/item?id=49764791