Hi friends, mod here.
I want to take a moment to address the issue of low-quality AI-generated content that occasionally gets posted in this community.
While /c/comicstrips hasn’t had a specific rule about this until now, it’s become clear that some AI-generated comics being shared fall well below the standard of quality we aim to maintain. These low-effort posts not only disrupt the overall experience for readers, but they also devalue the hard work of the illustrators and artists who contribute original content here.
Starting today, a new rule is in place:
AI-generated comics are no longer allowed in this community.
Thanks for understanding and helping keep /c/comicstrips a high-quality space for everyone who loves comics.
So if you’re ok with low quality, why then you object about AI comics quality which is mostly significantly higher than “stick figures”?
There’s a difference between “low quality” and “simple”. Cyanide and Happiness, to use their example was simple, not low quality.
XKCD also a great example
yes. Yes. it is.
Because I believe stick figures are of a higher quality than AI. An AI generated comic shows me that you cared so little about it that you couldn’t even draw stick figures. And this has nothing to do with effort, but artistic intent. It shows that whoever did it had no desire to inject themselves into the comic and would rather have it look generic and bland. Fidelity does not equal quality.
Because generative AI is a lazy shortcut for the untalented or dispassionate. If you give half a shit, put in minimum effort and draw a damn stick figure in MS Paint.
Because AI “art” being posted around isn’t even consistent throughout the panel, and is often devoid of effort and context so even when it’s not that interesting, it’s not even pleasant to look at.
Comic artist like C&H and Oatmeal, while having terrible art, at least are funny and with context. There’s human in it, and they can call the style their “own”, even if it’s stick figure it’s stick figure with distinction. AI art just make you think “is this made from AI?”
You want people to appreciate AI comic? Sure, do the touch up first and make everything consistent.
Oatmeal is actually pretty well drawn. Not a good example. Xkcd would have been a better one.
Fair point. XKCD, Casually Explained, and MinuteScience all are stick figure but they still have some distinctive way to do their thing.
Absolutely! AI comics just lack soul.
because AI is lazy
You don’t even have to draw the stick figures yourself.
Nice library. But does using this solve anything? From the post and other comments, one of the main issues with AI generated comics is that the author do not present himself in the images, and by using this tool, the images are also automatically generated by users description. There is more control over the final layout and there is needed more creativity to create it. But in the end I see the same issues about “being lazy” and having no author personality in the images. PS: I’m not saying this tool should be banned, just want to point out some similarities of this tool to AI tools.
If you hate the mouse like I do, yes. Think of codecomic as comparable to OpenSCAD only where OpenSCAD is for making 3d models, codecomic is for making webcomics. (And of course 3d models could be made in FreeCAD, but OpenSCAD makes for a different paradigm.)
Full disclosure, I’m the author of codecomic. (I’m the AntiMS who owns that repository on Gitlab.) Just to explain why I wrote it in the first place, I was GMing a D&D game at the time. And I have my own system for organizing my GM notes (which is a whole other code project that I haven’t Open Sourced yet) that involves writing Markdown and running a build process that generates a static HTML “website” locally which I can look at and navigate at the table.
A picture (in this case, a webcomic or storyboard) is worth 1000 words and my brain can process it quicker while I’m at the table and thinking on my feet. (I don’t want to be at the table and be like “hold on everybody while I read these three paragraphs of notes about the next 30 seconds of what is likely to go down in-game.”) Interpreting a webcomic is just something my brain does faster.
But if I have to slog and toil at the Gimp user interface to crank out tons of webcomics… I’ll probably just rage-quit GMing rather than endure that.
Plus organizing all the images so they appear correctly in my notes would be a pain. And making tweaks and revisions to comics I’d made previously would be a nightmare.
So I wrote codecomic to speed things up. And the coolest part. I wrote an extension for my markdown-based notes system to let me embed codecomic source code blocks directly in my markdown source files. Running the build for my notes system would produce HTML like usual, but with the generated comics directly in the HTML documents.
(And then scheduling conflicts happened and the campaign fizzled out. Sigh. Such is the life of the forever GM. Lol.)
I definitely didn’t want to have to fight with something like Stable Diffusion for this use case. If I wanted to tweak an existing comic and move one stick figure 20 pixels to the right or modify the text in a “speech bubble”, I didn’t want to have to regenerate the image with roughly the same keywords and pray to my graphics card that it would output exactly what I have in mind. I just want to modify my codecomic source code to change the X coordinate of that stick figure by 30 pixels.
Codecomic is quite limited and opinionated in its current functionality. I’ve definitely got things in mind regarding how I might extend it in the future to give users more creative freedom. (Mind you, I probably wouldn’t get back to working on it any time soon unless folks start expressing interest in it. Keep in mind the codecomic repository has exactly one (one!) commit in it.) It doesn’t support backgrounds currently. I’d be interested (if others expressed interest in it) in potentially adding support for other, more sophisticated ways to represent characters than just stick figures. More ways to draw stuff. Other ways of drawing lines that have more “character”. Etc etc etc.
You definitely have a point when it comes to the current state of codecomic. But whereas something like Stable Diffusion can’t be extended to give users more creative freedom, I don’t think the same could be said about the approach codecomic takes. Codecomic could definitely be evolved more to allow the user to put more of themselves into the comics while still reducing some of the overhead and “fiddliness” of using something line Gimp for making web comics.
One more point I’ll make. I think the approach that codecomic takes would allow for “incrementally” adding more of “yourself” into comics in a way that Stable Diffusion couldn’t. Spending more time designing your Stable Diffusion prompts/keywords/inputs/whatever could give you “better” (and I put “better” here in very large double quotes) results. But only to a point. After not too long, further tuning your input is going to give you increasingly diminishing returns. With codecomic’s approach (or at least the vision for codecomic), there’s not really a limit to how much creativity one could pour into a given webcomic.
Also, thanks for your input! I’ll definitely at least make some notes and let that influence the direction I take codecomic in the future. (If I do continue working on it, that is. No promises.)
Edit: Some screenshots from my GM notes to show how I’m using codecomic:
Warning: Spoilers for the published module Forgotten Relics from the book Eberron: Rising From The Last War.
Thanks for your replay, it was very insightful. I have to say that I do not have anything against your tool. I think it’s wonderful that you created it and allowed anyone to use it. Also, I do not create comics, I just came across this post, so I don’t really care about the rules in this community. I just hate that so many people are refusing any AI tool because it is AI. I also don’t like how everyone is pushing AI into their products, but in my opinion there are places, where these tools are really useful. That said, I understand, that people want to ban AI generated comics. It only seemed interesting to me that as an alternative, you suggested using another tool for generating comics by computer (although there are some similarities, I know that AI tools are very different from codecomic). So lwt me repeat, codecomic sound awesome and I only wrote my comment to start a discussion.
I have one question, would it be okay to create codecomic source code with LLM?
And one final note on AI tools. Some of them are really complex. I don’t use them much, but some of them allow user to edit only parts of the image with a prompt, or create an image based on user sketch, etc.
Well, we can agree to disagree. Lol! Seriously, I honestly don’t have much charity for what is being called “generative AI” lately. Wouldn’t be caught dead using it. (Hell, I’m a mod of the “FuckAI” community here on Lemmy. Which isn’t actually saying that much because there are a lot of mods in that community. Heh. Anyway.)
But! To each their own.
Thanks! I’m glad to see some interest in it!
Well, as someone who is vehimently anti-AI, I don’t really think it’s “okay” to use LLMs or Stable Diffusion or whatever at all. But I wouldn’t say it’s any less ok to use an LLM to generate codecomic source code than to generate… I dunno… C++ source code or OpenSCAD source code or whatever. (Last I heard, the chat bots were so abysmally bad at OpenSCAD as to be completely infeasible to shoehorn into that use case. Not sure if anything’s changed about that. Similarly, I’d think with codecomic being as obscure as it is with very few example code snippets out there to train on, I’d be really surprised if any LLMs out there would be even a fraction as good any time soon at codecomic even as they’ve ever been at OpenSCAD.)
(This is probably a good point to say IANAL and none of what I say is legal advice.)
Beyond that, I purposefully licensed codecomic under the GPLv3 on the premise that I’m not ok with anyone else distributing any portion of codecomic (or anything I Open Source, really) without ensuring the recipient has the corresponding source code. (Also, codecomic’s source code has a “terser” binary form that can save space, so there is a rough equivalent to the difference between “compiled” and “source” code when speaking of codecomic code. Plus, codecomic can be used as a Go library instead of using the codecomic DSL and Go is a compiled language.) Given that basically the only codecomic source code publicly available right now is in my repository and covered by the GPLv3, any codecomic source code any LLM would be trained on any time soon would be licensed under the GPLv3. And any codecomic source code an LLM output would pretty much have to be so closely copied directly from that that I’d say it would ethically be well within my rights to demand attribution and compliance with the “source code provision” of the GPLv3. (Honestly, I’m not likely to spend money on lawyers to force the issue really. I might send a cease and desist or something if I found out that was being done, though. But you didn’t ask whether one could get away with it. Just whether it’s “okay”.)
I know in some jurisdictions, courts have basically ruled that the copyright owners of the training data don’t own the copyright on the output of LLMs trained on their work. (I think some courts have ruled that the human running/using the LLM doesn’t own the copyright on the LLM’s output either because to be covered by copyright requires specifically a human to exercise some amount of creativity in the creation of the work. One might be able to argue that “prompt engineering” qualifies as “creative”. I hope that argument doesn’t fly, but who knows what the courts will eventually decide.) But in those cases, the LLMs would be trained on data from such a diverse set of sources, under such a diverse set of different licenses, and from such a diverse set of authors that it can’t really be tracked down “which works” are being copied. But in the case of codecomic, unless/until wider-ish adoption happens, I’m thinking it would be a harder sell to try to claim that LLM-generated codecomic source code isn’t a derivative work of codecomic itself.
But again, if lots of people started publishing codecomic source code under licenses that don’t have a “source code provision” or anything roughly equivalent, I don’t think it would be any worse to use an LLM to generate codecomic source code than to generate source code in any other particular language. Hopefully that answers the question. Heh.
Yeah, I wouldn’t say any of those use cases really make LLMs any more similar to codecomic, though. There’s still a big difference between the two approaches to producing an image or webcomic or whatever.
And how does it is differ from AI?
It differs from AI in that it’s completely unintelligent and doesn’t try or pretend to be intelligent or creative in any way. It leaves all the intelligence and creativity up to the user. It involves no “training” on large quantities of scraped data. It won’t do anything it isn’t explicitly told to do. The exact placement and pose of every stick figure, the precise layout and size of the individual frames, the exact content of every chunk of text is all explicitly and precisely specified by the user of codecomic. (In a source code file.) Also, a given source code file will only ever produce exactly the same webcomic whereas generally with generative AI, the exact same input can be used to generate a bunch of candidate images from which the user must select the “best.”
With something like Stable Diffusion, it does rely on lots of training data and the user’s only input into the content of the “generated” output is to throw a word-salad of keywords at it and tell it to “discern roughly something that fits these keywords”. The user doesn’t specify the exact location of anything in the resulting image. The user doesn’t have control over what exact text appears in the resulting image (and typically AI can’t even do text that’s sensical). At best the user can “influence” what’s output by tuning the keywords and hope with their fingers crossed that the Stable Diffusion model does roughly what the user has in mind.
Something tells me you didn’t even read what is there.