I made a comic book using AI generated art in Midjourney

I used AI generated art to make a comic book. Here's what that was like.
Midjourney
www.midjourney.com
Dall E2
openai.com/dall-e-2/
Stable Diffusion
stability.ai/
The full Goats comic is below
www.webtoons.com/en/challenge...

[@TheMaddypax]

I watched this video, used your tips, took an old un-produced screenplay I had lying around, and made my own graphic novel using MidJourney... in 3 days. I'm blown away. And also filled with existential dread about what the future looks like. In a few years will we just type a set of prompts into A.I. to make us our own indIvidual movie with actors that we like, living or dead? Quite possibly. I'm honestly freaked out. But thank you for posting this!

[@moteworks]

Where can we read this novel

[@undrsonr5316]

AI will end the stream of bad artists. Just that. Take a Dave Mckean or a Simon Bisley comics or a Robert Crumb biographical comic, amazing artists with very idiosyncratic styles. Personality. AI just get the past and recycle... and EVEN if AI gains full consciousness, self-consciousness the AI art will reflect AI emotions, fears, desires...If AI get into this level of expertise - check the AI images, lacks knowledge of perspective, the hands are almost always deformed, etc, because it does not THINK the FORM, it just create from noise and a database - for now, maybe it can change in the future - so until we have REAL self-conscious AI capable of think FORM by FORM itself, AI will not get that far... Maybe it will? sure! If so humans will probably return to humans interacting with humans in art form or get bored and kill the species! :) but one thing out of all this hype: human consciousness using tools developed by human consciousness created artificial intelligence... so we have a limit, the emulation of consciousness developed by ourselves will probably reach a limit also - to believe a mind can create a mind beyond its own limitation is like religion, not logic, like faith in a transcendence by limitation. Transhumanists do so. Time will tell. and remember, the FUTURE IS NOW, tomorrow we can not be here anymore. so worries are over rated. ;)

[@dinupetrecristian150]

@@undrsonr5316 maybe adding another AI to check the finished generated image hands and make sure they look normal would be easy to implement ... And that AI to be trained only on a datasst with hands

[@undrsonr5316]

@@dinupetrecristian150 it’s not that simple. The AI would have to understand form in a relationship with the space it occupies. Goes beyond the level of thinking neural network can do and the way it process image from noise. Not an easy task for AI. Like self driving cars, the AI can’t think form in relation to each context. That’s why there is not a safe self driving car yet… IF AI can go to the level of real self awareness maybe this will be possible…. Again, this is all cool stuff, but there is an enthusiasm in excess that blinds people to the limitations of the technology, neural network, in it’s core. Btw, check the imagens, the same problem with hands you can see in noses, eyes lots of time too, because it needs to think form into the space under optic rules,etc… not just use a database of millions of images and getting the ones with similarities and “mashup” into new ones… it’s beyond neural network… never say never, some day some researcher may come up with a new machine learning process… when and if that happens then a REAL AI and a real revolution will come!.

[@satana8157]

The problem is that, with movies that can be finished in a day, who's going to watch all of these?

[@carlosalbertolealrodriguez5529]

Now more than ever, we must study art history, to understand what kind of art we are producing with these tools.

[@drugartwall]

you are not producing any art with mid journey ... if you studying art history before, you will know that , mate :)

[@yannberte8291]

The most important thing in an image is the balance - the colors, the ideas, the themes, the graphic styles, the light, the masses... in short everything that constitutes an image. All the artists whatever the eras, the countries... do a balanced work! If you can figure that out, you don't need to be a living encyclopedia - who can brag about that ??? Studying the balance of a work, of ten... is what is fundamental. Balance is peace, harmony, rest of the soul, the possibility of living with the world and not in the world and basically it is what we are all looking for in our existence and a fortiori unconsciously when you are in front of a work of art - which makes this balance something universal. Inwardly we are looking for this particular little moment of respite and a work echoes this search. Consider this manner, attitude... and you'll feel much better for it. Look at a young woman in the street for example; she is going to carry a handbag in correspondence with the color of her shoes, her lipstick...: it is a question of a point and a counterpoint like in music, in poetry...! She seeks balance without necessarily putting the right word on it. I did the ESAG (Ecole Supérieure d'Arts Graphiques) in Paris 37 years ago now (oops...) Believe me I know what I'm talking about on this particular subject and don't listen to the contemptuous beardless clowns who imagine that they know art by collecting completely external knowledge ; it's bullshit ! 🍻😉

[@itsallfunandgames723]

Knowing artists and art styles is going to be invaluable for the best operators out there creating AI art, and it won't surprise me if a certain sub-section of young people become experts on painters they way they currently know all their favorite bands, rappers, and pop singers.

[@PapayaPositive]

Today Concept Artists get to feel what I felt years ago when 3d Asset Libraries appeared. As a game artist, I became a drag-drop-edit machine which was detrimental to my creative skills and made the job drag. This is the era when the artist can come up with a final product alone or with a small team. The really scary part is that the gap between creating and being skillful is slowly disappearing, to the point where "entertainment" will shift from something you consume, to something you "create". Soon, everyone will be a "creator". Creating what? Content. What is content? What we used to call "art". Yet one more thing that points to a UBI social-credit technocratic future...

[@hyperbolicfilms]

I've had to make this giant shift in my brain over the past 20 years. When I started, I wanted to model every object and background on screen in my comic and animation work. I wanted to make all the titles and graphics. Now just to get something finished, I rely more and more on stock models or images or graphics. And in a way its good because it allows me to do more, but there's less of me in the final product. So much of what I do is curating the final product.

[@MeatCatCheesyBlaster]

If you do enough mushrooms you might start to wonder how much of YOU is really YOU at all and maybe what AI is teaching us is that we are all more like We or IT

[@LieorDie24]

Like i said in an upper post i use this tool too but i dont feel myself like a creator okay maybe a little bit i had the idra of how the picture should look but thats not the same i take photos too with my photo and i made a pretty cool shot of a firefly it took me a long time to find the right iso stuff and pose to get it sharp and look good yesterday i made the same picture with the Ai in 30 seconds on doing my photo i was pretty proud with this Ai tool i wasnt and maybe people photographers art designers and so on will loose their jobs only because of such Ai tools thats why it makes me sad too.

[@remo1wodmnetwork9605]

You were dealing with the Genie of the Lamp. You have to be VERY specific with how you word your wish.

[@CrniWuk]

My words exactly. I think people have, no offense meant, an almost childlike naivity about the use of Ai and what it will mean to our society and the job market. The potential as a technology, is incredible. The things that it might do in a few years? Amazing. But, I believe it will have very negative impacts on our society as a whole. On one hand, because of abuse. This will happen. It's innetivable. Scammers, political manipulation, terrorists you name it. And the other side, devalue of intellectual labour. Ai will make it possible for 1 or 2 people to do the work of 10 or even 100 people. It will not "replace" human workers but it will make existing work more efficient. Exponentially! And if people belive, so what? Then I just have to get better! Well, good luck competing with a machine then. It's not a "gid gud" problem. Supply and demand. If supply increases but demand doesn't what happens is that you have to lower your price. Income shrinks. Even for those that that still have a job. And we can't all become indiedevelopers or freelance creators. Regardless how much skill we have. There are only so many consumers out there. I think a lot of workers, not just artists, will find them self in a position where weavers have been 200 years ago when the steam engine automtised their jobs and 1 worker was capable of doing the job of 100 people. It was great from a technological point. Not from a social one.

[@hyperbolicfilms]

It's absolutely a systemic problem that is going to get worse. I really did think the truck drivers would be affected long before the artists, but here we are. There's a great article from some time back about the Luddites by Cory Doctorow that I think speaks to some of your points, and how they weren't anti-technology, but anti-what the technology was doing to workers. locusmag.com/2022/01/cory-doctorow-science-fiction-is-a-luddite-literature/

[@danitaylor92]

Humans always loose their job when technology is improved. Artists with solid and complete skillset won't fear IA for a long time, but they will be caught up at some time for sure, sadly not because of how much technology will evolve. I think the deterioration of the artistic culture (visual, literal, musical etc) is the real demise and people are more used to consume content rather than appreciating art, because art needs education. We shouldn't blame anything on IArts and its devs, it's 100% fine and somehow entertaining. Our culture is not, our culture reached a critical level and is getting worse. In comments below i read someone saying IArt is good because comics artists can use it to create backgrounds, focusing on doing better character art, to be fast as mangaka are, so readers don't have to wait and will enjoy comics more. That's an uneducated take by someone who doesn't know how comics are done, in any region of the world lol (but I admit can be quite right on the time issue). No way a good quality comic can have their background done like that (tracing 3D models is still more effective and it still requires knowledge of perspective and 3d building). What i fear is that people like this, totally uneducated, will become (i fear they already are) the majority of us and will decide what's the art quality standard, elevating the IA repetitive patterns over the human genius. That's what I'm scared of and that's what I can see killing the artists' job, not IArt. That's also what's killing the art of singing, it's not to blame it on autotune and who created it. It's just our culture deteriorating and lowering the standard. Anyway, comics businness industry looked like it was fucked even before IArt (and not because of PC as lot of people like to say) so we shouldn't really bother at all, humanity is doomed, no matter what external reasons we'll blame for our failure

[@RM_VFX]

You also can generate the character with a nondescript background, then generate the background separately, and merge them together in photoshop pretty easily. It does require some art ablility to do so.

[@Jmo__tv]

It’s such a difficult topic with ai potentially running artists out of positions… maybe in the future human / ai art will just serve as different genres

[@lop0159]

Thats what i hope will happen. That ai art is just another genre like oil painting or digital art

[@AscendantStoic]

Specificity and consistency are a still a huge issue with A.I art ... which is where artists could come in and integrate the A.I as just another tool into their arsenal of tools like they did with Photoshop/Krita/Clip Studio and Blender. Image2Image tech gives artists the chance to be specific and consistent because you can draw sketches with poses and specific camera angles and details very quickly (in 5-10 mins), then let the AI artistically render it ... then the art here becomes a collaboration between the artist and the AI and it will help improve productivity significantly (and with Stable Diffusion which is open source it's available for everyone for free .. aside from the one time payment to get a good PC video card with lots of VRAM).

[@iamthe80s49]

Ah yes, all those artists in all those positions. So many positions.

[@Jmo__tv]

@@iamthe80s49 animators,logo designers, web designers... etc

[@lotsaspaghettimamaluigi]

The gap is closing fast, even the hands look good now. Soon ai art will be indistinguishable from hand drawn art

[@xbon1]

actions are much better done in stable diffusion, and absolutely nothing is banned, no banned words, everything is allowed if you run it locally on your own GPU.

[@hyperbolicfilms]

I am going to explore it in the future. When it was integrated into MJ as “beta” this past week, I was getting some great action, but not as artistic results. But I think that’s a matter of crafting my prompts better.

[@morizanova585]

I wonder if you ever get consistent style outside of anime in Stable Diffusion . My Interior and Exterior scenes seem mismatched in term of shape design or characteristic ( although I`m using the same artist or lighting style ) That will breaking contuinity in my opinion . Thats why I think real artist still needed . At least to make seamless from draft idea to be final product

[@Tarazed609]

As an artist, I'm embracing this new method of creation. More than anything, it speeds along the process of reaching the type of art I want by giving me a base to work on, or new inspiration I would never had thought of. Because it's so quick, I can go further within days whereas it would have taken me months before, which is a real boost to moral as well.

[@reidmoto]

I think that’s a very wise approach vs. focusing on fear of loss IMHO. There will be major shifts in how and what type of art will be created in the very near future. As a songwriter, I thought of the massive amounts of new music that would flood and clutter the market when DAWs like ProTools started hitting their stride. But I quickly changed my attitude to the mindset you shared and benefited greatly because of it.

[@TheGreschler]

Great video and I really enjoyed the storybook. I'm trying to do a kid's story but, like you, having trouble keeping a consistent design between panels. I'll try some of your tips. You do a great job of covering all the big issues.

[@hyperbolicfilms]

Good luck! It feels like Midjourney has changed again in the last two weeks and is even less consistent than before. I am having trouble starting part 2 of Goats because I can't get images that feel consistent with the first portion.

[@eloureirotubeyou]

Just impressive your comic! Wonderful result!

[@hyperbolicfilms]

Thank you, I appreciate it!

[@alexiades]

It would be great if you could share some of the prompts you used in some sequences so that everybody can learn how to start their own comics book story... Thx

[@SuperAlphaKirby]

I'm happy for all the commenters who were able to make their own graphic novels into dreams come true, but i won't lie as an artist i kinda feel bad that no one will hire me to make their graphic novels...

[@hyperbolicfilms]

It does suck, but hiring an artist for a comic or graphic novel has always been prohibitively expensive for most writers. If you're talking 22 pages, that's anywhere from $1000 to $8000 to get the art done. My hope is that I can maybe make enough money on some of the AI generated comics to hire an artist to do my dream project, because that one is something I'd like done with no compromises. And with AI there are a lot of compromises.

[@thebarbaryghostsf]

Thanks for this! Been trying to figure out how to create a comic to accompany a concept album I released with my band this month. This is super helpful, thank again for these tips!

[@Amelia_PC]

Lovely! Honestly, it's the best comic book made by AI I've seen so far. Good job! I work with comics and I was trying to speed up my process with AI (img2img in Stable Diffusion UI V2). The problem is cleanness and consistency. It still doesn't keep these aspects, but it's great for painted abstract comics or cover art (for cover art, if you know how to sketch, it's a killer tool. I've tested and it's amazing how AI follows my sketches).

[@talkingmonkeeyz4000]

Thanks for figuring all that out for us man. *Subscribed*

[@corridor444]

Thanks for sharing your experience

[@WalidDingsdale]

enjoy listening to your reflection, comments and other talks about this ai implications from a professional artist's perspective. keep going

[@SolveForX]

Feels like the best application for this right now is a David Mack approach. Where the images can be used as collage.

[@MikeKleinsteuber]

Nice piece which raises some interesting points. I really don't know how the art industry will handle AI but I suspect we'll just get used to having a new tool at our disposal. We've already had copyright issues with tools like Photoshop but at the end of the day the artist controlling things is still the artist and I don't see why that will change. Keep pushing the boundaries and see where you get. But don't forget to have fun along the way...the journey after all is the best bit...

[@jsequeiramontoya]

Really nice story, great combination with the ai images! Congratulations for the good work!

[@hyperbolicfilms]

Thanks a lot!

[@simplythebest2k]

I’ve seen a few that look like other artist painting. From what I’ve come to learn of AI is that it pulls from a library of photos and art. These AIs basically use the internet as their library. Then it makes a Frankenstein of it. So it’s kind of like tracing and collaging. Now the computer does it in such a way that humans have not that is harder to tell. So it’s not something that is easily spotted. A lot of people think the AI is learning and creating things out of nothing and that is simply not the case at all.

[@jeremycointin1996]

Brilliant! That "actor name" tip is the game changer.!

[@TiagoTiagoT]

For a little bit more control, look into the img2img stuff with Stable Diffusion and similar

[@nathanbanks2354]

Last week I played with the stable-diffusion image to image transfer with text to modify it, and I'm having trouble getting anything I like the look of. This is probably because my vision of what I want things to look like doesn't exist in the training data. I don't have access to dall-e 2, but the Open AI language & code generation is certainly impressive. You managed to get tons of images with a consistent style which is really quite commendable, especially without using any input image at all. Well done!

[@hyperbolicfilms]

I have had 0 luck using image inputs. I've tried and tried in hopes of using it to generate some action sequences, but so far even giving it a high image weight doesn't seem to do much. Very interested in trying image 2 image.

[@librabys]

It's just like calculators. They did not make mathematicians obsolete, they just transformed the profession forever allowing them to skip the most repetitive and basic part of the job and focus on more complex and interesting realisations. They still need to understand what they do and interpret the results with their own skills. They have to write their own thesis. The calculator was just a mere tool in the process. The human is necessary to make the right inputs, and to take the raw data obtained and make something usable out of it.

[@hyperbolicfilms]

I think you are maybe right, the difference being the same equation will give you different results almost every time when it comes to some of these AI art programs. But it is a very excellent tool for anyone who is a story teller. Every wannabe film and TV person can generate a bunch of cool images much closer to their story idea than just ripping images from Pinterest to make a look book.

[@CrniWuk]

Bad analogy I am afraid! Here let me give you a quick example. Who did the calculation before the calculator was a thing? Humans did. Basically they took very large calculations, broke them up into smaller ones and gave them to a number of people - quote often females! - which would then calculate it. That was NOT done by mathematicians. Once computers actually could do that? They became obsolete. Of course the calculator didn't make mathematicians obsolete. Because that was never the intended target of calculators. Being a mathematician is about problem solving. They didn't calculate everything on their own. They had "human" calculators do that. And that was a practise that we can find even as far back as when Newton did his mathematics. I do not want to attack anyone. Seriously. But what frightens me more than the potential of Ai is the naivity. Either because people understimate it or because they don't understand the intention and mechanic behind Ai. Ai is to Creatives right now what the steam engine was to weavers. Before the steam engine and electricity one weaver operated one machine. With the invention and use of the new technology, one weaver could suddenly operate 100 machines or even more. This is what we're looking at now. Ai is not meant to replace workers and human labour. It's meant to make it more efficient. At a rate that we haven't seen before in intellectual work. And just like the Steam Engine devalued labour enormously back then, it will do the same now. And that is NOT a question of skill. How can you charge for a product the same price when you created in one day what took you usually a month? What will happen is an overresaturation of the market while at the same time one person can do the work of 10 or potentially 100 other people. When you can throw out designs in mere hours that would take weeks to make, what will that mean in the end? Companies are already looking now very carefully at the software as a potential mean to drastically lower their creative teams. And once they get that chance this is what they will do. A lot of people believe this might eventually "replace" artists completely. I don't think it will. But it will lead to a situation where you won't be able to generate income anymore trough it. And we have historical presendence when we look at the industrial revolution and what the exponential growth in efficiency did to the labour market and society. And we are not only talking about the crative industry here. All intellectual tasks, can eventually be "automated".

[@ThePlaceForThings]

pretty sweet. loved the comic, wish I would have read it before watching the video! Would have made for a cool “behind the scenes” video. Going to use some of these techniques to make storyboards for my screenplays 🧠 ❤️‍🔥

[@valkopuhelin2581]

"So many people have ideas in their heads that they have never been able to get out in a satisfying way." THIS!

[@milestrombley1466]

Maybe hand draw stretches of a scene you want and generate it into a better style in the app. I gave up drawing a long time ago and became a writer instead. This looks like a big opportunity to bring my story ideas to life now. 😁

[@hyperbolicfilms]

They haven't achieved any kind of consistency between images when inputting a hand drawn image. I believe Midjourney is working on something like that based on things they've said in their office hours events, but could be months before we see something like that.

[@FloKorp86]

Dude! Your comic looks absolutely incredible! I'd be very interested in your process because I want to tell the story of my grandfather in comic form, but never had the chance to bring those ideas to life until now. How did you get the panels to be so consistent and make the characters appear in the same place? Should I be working with image prompts to make places look similar to one another or is that just a matter of figuring out the right text prompts?

[@hyperbolicfilms]

Thanks! For actually laying out the comic, I suggest Comic Life, which is what I used after I generated the images in Midjourney. To get similar backgrounds is really just curating the images that Midjourney generates. I throw out 14 images for everyone image I use.

[@reidmoto]

Wow, thanks for sharing that very specific metric!

[@ogbogukalu7877]

Hey. I think you might want to try Stable Diffusion. Granted, it has a lot of the same weaknesses but has much better workarounds. E.g img2img and inpainting. For example, if you generate a scene you like overall but the characters faces have turned into blobs, they'd be no need to start over. Just mask the face area alone and inpaint and you get a fixed face more often than not. Or perhaps you just can't get an action scene or some other scene going. You can do a rough sketch of what you want and upload that as a starting point. Generally you get much more coherent results

[@hyperbolicfilms]

I was actually just trying to install it locally, more to try img2img to do some virtual makeup. Ran into some kind of python error, but will try again later in the week. My attempts at using img2img on huggingface have not been great, and I wonder if it's just that I don't know what all the sliders do.

[@ogbogukalu1839]

@@hyperbolicfilms Been trying to send a comment with links to help and youtube keeps deleting them so i just sent an email instead. It goes through what ui to use, settings with examples to get good results and so on. Good luck.

[@dibbidydoo4318]

Something like stable diffusion's automatic1111 fork has img2img and inpainting, where you have more control over a scene by drawing a rough sketch and you can remove elements of the scene that you dislike or want to change with inpainting. I personally think Automatic1111's Stable Diffusion has the most features and settings with 20-30 features.

[@adrianoeliezer123]

Amazing! How did you manage to keep the style consistent throughout the work?

[@hyperbolicfilms]

The people were consistent by using actor's names, and the style was consistent by not forcing a style of Midjourney. Just giving a mood to the prompts seemed to give me a consistent painterly look.

[@RoseAbrams]

I have absolutely zero artisitc skills, but Midjourney enabled me to make a graphic novel. My graphic novel would not exist if it wasn't for Midjourney, but now it does. That's why AI won't destroy artists and creativity, instead it will empower creativity. It will enable so many people to create works that the world would otherwise never see. Not everyone can become an artist or hire an artist, and for those who can they will always get a better and more accurate result. AI has just opened another door for non-artists.

[@Tuttlm27]

How to you manage to keep producing the same character over and over to use for the novel?

[@RoseAbrams]

@@Tuttlm27 Using names of actors and other celebrities, instead of my own character names

[@it6647]

Where can I read it?

[@gandolphgandolphini]

Anybody who is fluent in a language and can draw will be safe for a while in terms of finding work. And it's always good to note that machines will always need human operators. They don't have motivations or desires and won't just make stuff without being told to. I like the example in the video where the guy asks for picture of a person running and can't get it. To "run" for a computer simply means to be on. Any pic of your protagonist not dead is a pic of them "running" as far as a computer can tell. And that's just one needle in the needle stack. I asked multiple AIs to generate a picture of a "rock band" and almost exclusively got back pictures of literal stones and geodes. Even after explicitly asking for images of people "playing guitars", I doubt AIs understand what arms are, why they exist, what they do, or how many any given object is supposed to have. It shows in their work, and it's not the kind of information you can just look up. I find AI at is best for drafting multiple versions of a thing and the just collageing the best bits into something usable.

[@williand8]

You're relying all your arguments in the limitations of the first wave of AI we have (which is pretty mind blowing already actually)

[@gandolphgandolphini]

@@williand8 I am. Although I think the language barrier will persist to some degree just about always. How long have English, French, and Arabic coexisted and yet miscommunication still happens about daily between the speakers of these tounges. Even more so with humans and computers. Computers have only begun to speak back so miscommunication with be the rule until it becomes the exception.

[@CrniWuk]

@@gandolphgandolphini The thing is that you have to look at what the intention is behind Ai. Ai is to intellectual tasks what the steam engine was to physical ones. People often misjudge here someting I am afraid. They often believe that it is meant to "replace" jobs entirely. But that's not what it will do. We will find our self in a situation - and not just with creative work it's just where it will happen first - that 1 or 2 people can do the work of 10 or even more. And they will do it faster and more efficiently. Already now we can see that companies and business are very carefully looking at those Ai Image Generators as means to drastically lower their creative teams. It's not quite there yet. But it might be in 1 year. Who knows. I think it will happen rather sooner than later. Particularly now, as many companies have a real motivation to cut costs drastically. Creatives right now are in a situation not unlike weavers found them self in 200 years ago, where the invention of the steam engine created an efficiency and automation like never seen before. The nature of the task made it easy to be automated and suddenly one person could operate 100 machines. The result? Extreme devalue of labour. To an extend where it caused mass poverty and even starvation. And this is what will happen to a lot of artists and creatives out there. When you can do someting that would take weeks if not months in mere hours with typing a few words in a software, what kind of charge can you expect? People will argue that you pay for the experience but I don't think that will be of any solace. Even if you have one artist/creative making a living out of it somehow, what will the others do? Not all weavers lost their income due to the steam engine. But the majority of them did. The amount of content that can be created here, will eventually completely oversaturate the market. The number of designs, illustrations and work that will be created will skyrocket. More than it already is. There is a reason why freelancers are struggling so much even before Ai became a thing. And we're talking about really skilled artists and creatives. There is no way to compete as humans with machines. And honestly, the creative process behind using the Ai? It's not rocket science. Not after I tried the software my self and how easy and intuitive it is to use eve now, how readily people share their prompts and texts out there. And we can also expect that the software will get better over time in actually understanding what we want from it. The technical aspect has been nearly completely eliminated here. And that is what the whole goal of Ai is. To actually automatise inellectual tasks. - That does not mean that the Ai has actually real intelligence by the way just that it can very efficiently mimic our work process! Not this generation of AI worries me. It's the one after it. And the one after that. This let us call it "commercial" Ai, compared to the prototypes we use right now, might be voice activated and actually connected in a more intuitive software where you can make changes on the fly. Imagine Photoshop with an inside AI that you can actually talk to. Not happy with the face? Replace it. Want it all following images to have the same character in a specific pose? Just tell it to the Ai and it will "paint" it for you. And as the situation is right now, you might even have an Ai writing the text for the character and story based on your "idea" or "concept". And this commercial Ai, could become a reality in the next few years already. We really haven't seen anything yet. I am sure about that. The only limitation here, as of right now, is processing power and easy access to large quantities of information and data. That's pretty much it really. We have only begun to see what Ai can do. We are not even near it's full potential I am afraid. And our society is not ready for that impact.

[@gandolphgandolphini]

@@CrniWuk I totally said that. Machines will always need operators. Machines don't make things by themselves.

[@TiagoTiagoT]

Another thing to further enhance control is this "Textual Inversion" trick that lets you teach the AI with just a few examples; I'm not sure what's the hardware requirements to run it locally though, and dunno if there are any online services available.

[@paulspvk6049]

Stable diffusion is open source and can be run on a personal machine, so no restrictions about certain concepts. It also has img2img which allows you to give it a starting image allowing you to direct the composition. It can either be a scribble in paint, or a 3d render of some basic models. Even Dalle Mini/Craiyon generations work as a good start since Craiyon is good at composition but super bad at details ( due to the architecture). It also doesn't have the same default "Midjourney look". So while the default output can look bad, with the proper inputs ( for style etc. ) it can provide much better images for comics.

[@AscendantStoic]

1-I don't think there should be copyright for A.I created material unless it's heavily edited by the creator of the final product and the A.I generated art is just a small element in it. 2-Image2Image technology in Stable Diffusion (or using a "source image" in Midjourney) is the key to getting "specificity" you want .. you can sketch a drawing or a comic panel in 5-10 mins with very basic colors in MSPaint or scan it from a piece of A5 paper then input it into the A.I + prompts + mess with the settings and it will turn your sketch into a finished fully rendered work using the sketch as guidance (and you get to decide how much to depend on the sketch and how much to depend on the prompts or on just pure randomness, SD GUI provides all these options)....this is the tool that will be most practical for artists to integrate into their workflow.

[@hyperbolicfilms]

Have you found Image2Image to be repeatable? I have a hand drawn comic that I'd love to run through it and see if it can essentially upscale my crappy black and white art.

[@AscendantStoic]

@@hyperbolicfilms It will give you a lot more consistency and specificity than working with just a prompt, you still require a prompt but the settings in the GUI allows you to tell the A.I how much of your image to use VS the prompt VS letting it go wild and chaotic (make sure to use "seeds" if you like a specific result but want to slightly alter it, using seeds is another method to keep things more consistent) It will also allow you to do action scenes more easily like the martial arts fights.

[@AscendantStoic]

@@hyperbolicfilms C h e c k R e d d I t --> (image2image functionality) o n S t a b l e D i f f u s i o n p a g e.

[@binyaminbass]

I watched this video for one thing: how do I get consistent looking characters across many prompts. And you delivered! Using famous people. Brilliant idea. I hope that stays legal. Or maybe the AI could learn consistency.

[@kvasac123]

It is not legal if you plan to use it commercially. In fact this is downright bad advice as you need mode release consent form to use someones likeness and when one of these celebs file you a lawsuit it is not going to get cheap on you.

[@nathanielalderson9111]

@@kvasac123 Only if you take a photograph of that person. Also in IP, "celebrities" or "public figures" are exempt from requiring model release forms. As are on duty public servants. This is completely different. This isn't like past. There's really no precedent. This is many faceted nuanced. Difficult. Who owns imagination? Who owns pattern recognition? Who owns the tools we use? Analogy: Do I have to pay a royalty everytime I use a specific knife with "Sandra" enscribed on it, to make my sandwiches for lunches that I sell? These are images that did not exist before, using similar or "recognisable" celebrities, using a mashup process. A smooth montage, as it were. This goes way beyond "change something at least 25% different". Nobody owns the rights to these images. The "creator" might, but technically they didn't create them, because the AI did.

[@Loony.Luna.]

just installed stable diffusion. i don't know if the sun exists anymore.

[@hyperbolicfilms]

!dream "a glorious sun, shining rays of warm light down on Funny Things" -W 704 -C 15.0 -S 32452345234

[@Loony.Luna.]

@@hyperbolicfilms

[@weilam03]

i was thinking of doing this too and the idea of making the chatacters consistent was one of the biggest things also keeping the backgrounds and sceneries consistent would be hard.

[@zakj3769]

Yeah I'm wondering the same thing, I know Scott Detweiler touches these up in post and fixes faces in other Ai programmes, but I wonder how to do close ups, different povs and zoom levels, and emotions and keep them consistent

[@hyperbolicfilms]

I think I was able to get a little better consistency with characters, backgrounds, and having emotion on the character’s faces in the second comic I made. But it was shorter and that helped a lot.

[@TheVitkuz]

Very intresting video. It seems like you strugle a lot to create comic book with unpecice tool. Can you create video where you show your pocess... It will be itresting to see workflow

[@timbacodes8021]

where can we get our hands or a digital look at “The Astounding Adventures of Captain Lightning“ ?

[@hyperbolicfilms]

I tried Archive.org, but a lot of my sites were Flash-based in those days, so they can't be accessed easily any more. I likely have the files on a hard drive or data DVD somewhere, but not sure where they might be.

[@ergohash2517]

i saw that stable diffusion allows for a more precise control by using your own visual references and photobashing. hovewer, midjourney also allows base photo refs ig i am not mistaken

[@morizanova585]

I`ve been trying these idea using Stable Diffusion but having no luck to get consistent styles except put Studio Ghibli as part of prompt . Some styles having more consistent if used as same type of renders but not if used in another type of renders . i.e : Some style maybe great for character but produce strange /mismatch output when used in "background type of prompt"

[@shada0]

I think for this style to work further we would need a different art direction for the word balloons, the flat colors whites just completely clash.

[@CrniWuk]

There are already attempts for upgrading Ai-Image generators in a way so they can actually learn how to paint and work in your style. I am afraid, we have only seen the begining of what's possible. I am fascinated and afraid of the prospect what we might see in 1 or 5 years from now. We are at the dawn of a completely new age.

[@allangoacher]

Very interesting video, thank you - did you use composites of the output from Midjourney (i.e generate the backgrounds and characters separately) or was it characters and scene together straight out for MJ?

[@hyperbolicfilms]

This comic was straight out of Midjourney. I flipped a couple of images before laying them out in Comic Life 3, but otherwise the prompts generally asked for the character, their action, and a background.

[@yaellichaa5169]

Love it!!!

[@darrenkairis9733]

Good for you, brother.

[@LeftBoot]

Good attempt and experiment. I can only imagine how time consuming that must have been. Maybe an uncut video could help train a specific creator AI. Maybe a UI providing multiple chains of synonyms. I enjoy testing new AI with... 'ornate descending staircase into a colourful misty neuronauts universe.' The action of descending appears to confuse. I wonder if one day we will have a specific comic book creator AI. Imagine all communities have their own individualized bots, each generating custom digital material. Bots having 'their own' market. All of our groups exchanging blocks of creations, fuelling future ideas and concepts. It's all about the training. 👍 💡Merry Christmas 🎄

[@michaelt7209]

The only use out of this art generator would be as an alternative for having cover art for your book or album. Anything else like a graphic novel or children's book is going to be a headache. One guy I know of pulled it off, the guys name is Kevin hess and he graphically adapted the book starmaker. Which is an old book with an omniscient or all observing narrator. The book doesn't really have a plot, dialogue, or characters in it. The book is mostly of philosophical musings and world descriptions, which is perfect for art generators that specialize in high concept art of environments or scenery. Also thanks for your tip for using actor names@

[@hyperbolicfilms]

Midjourney v4 is a massive improvement since I made this video, and it is definitely capable of making a children's book. Coherence between images is still hard, but getting better with every version.

[@theleadingedgeenglish1260]

Great video and thank you

[@cobaltplasma]

As a professional artist I think AI-generated images are very, very interesting, more of a useful tool that can help democratize the creative process for folks who don't have as much training or experience and allow them to visualize and realize ideas they've had but couldn't quite put them forth out in the world to their satisfaction. I also think for professionals and professionals-in-training that it's a tool to springboard off of, something to generate ideas and directions and maybe composite things youwouldn't have thought of otherwise, but in the end it will fall to you as an artist to craft and mold what the AI spits out into something usable for your client or for yourself.

[@librabys]

As an artist i use it in the way i would use references, because i can sometimes have pertinent references already mashed up from lots of references, instead of wasting my time on pinterest to find myself all those references.

[@hyperbolicfilms]

It's definitely been inspiring to me to start drawing my hand drawn comic again, and in doing some repaints on my second comic, it got me thinking about how I could paint limbs and backgrounds that fit with the style of the AI generated art, so embracing it can definitely be a good inspriation and learning tool.

[@hyperbolicfilms]

@@librabys Yeah, it seems like a great tool for mood boards and things. It's very helpful as an idea generator.

[@cobaltplasma]

@@hyperbolicfilms I've been experimenting with generating general forms and ideas and then painting over those to make them more bespoke, more of my own creation. I'm glad it's inspiring, I think it should be, and so long as you're happy with what you make with it then in the end that's all that matters :)

[@CrniWuk]

The thing is that you have to look at what the intention is behind Ai. Ai is to intellectual tasks what the steam engine was to physical ones. People often misjudge here someting I am afraid. They often believe that it is meant to "replace" jobs entirely. But that's not what it will do. We will find our self in a situation - and not just with creative work it's just where it will happen first - that 1 or 2 people can do the work of 10 or even more. And they will do it faster and more efficiently. Already now we can see that companies and business are very carefully looking at those Ai Image Generators as means to drastically lower their creative teams. It's not quite there yet. But it might be in 1 year. Who knows. I think it will happen rather sooner than later. Particularly now, as many companies have a real motivation to cut costs drastically. And when it comes to creative work there is absolutely no risk involved either. If it doesn't work perfectly? Who cares. It's not like drug prescriptions - which by the way Ai can already do as well like human doctors, but no one's touching that yet, because errors here can be deadly. Creatives right now are in a situation not unlike weavers found them self in 200 years ago, where the invention of the steam engine created an efficiency and automation like never seen before. The nature of the task made it easy to be automated and suddenly one person could operate 100 machines. The result? Extreme devalue of labour. To an extend where it caused mass poverty and even starvation. And this is what will happen to a lot of artists and creatives out there. When you can do someting that would take weeks if not months in mere hours with typing a few words in a software, what kind of charge can you expect? People will argue that you pay for the experience but I don't think that will be of any solace. Even if you have one artist/creative making a living out of it somehow, what will the others do? Not all weavers lost their income due to the steam engine. But the majority of them did. The amount of content that can be created here, will eventually completely oversaturate the market. The number of designs, illustrations and work that will be created will skyrocket. More than it already is. There is a reason why freelancers are struggling so much even before Ai became a thing. And we're talking about really skilled artists and creatives. There is no way to compete as humans with machines. And honestly, the creative process behind using the Ai? It's not rocket science. Not after I tried the software my self and how easy and intuitive it is to use eve now, how readily people share their prompts and texts out there. And we can also expect that the software will get better over time in actually understanding what we want from it. The technical aspect has been nearly completely eliminated here. And that is what the whole goal of Ai is. To actually automatise inellectual tasks. - That does not mean that the Ai has actually real intelligence by the way just that it can very efficiently mimic our work process! Not this generation of AI worries me. It's the one after it. And the one after that. This let us call it "commercial" Ai, compared to the prototypes we use right now, might be voice activated and actually connected in a more intuitive software where you can make changes on the fly. Imagine Photoshop with an inside AI that you can actually talk to. Not happy with the face? Replace it. Want it all following images to have the same character in a specific pose? Just tell it to the Ai and it will "paint" it for you. And as the situation is right now, you might even have an Ai writing the text for the character and story based on your "idea" or "concept". And this commercial Ai, could become a reality in the next few years already. We really haven't seen anything yet. I am sure about that. The only limitation here, as of right now, is processing power and easy access to large quantities of information and data. That's pretty much it really. We have only begun to see what Ai can do. We are not even near it's full potential I am afraid. And our society is not ready for that impact.

[@YouDoTheShoot]

I raed your comic very good definitely wanted more

[@jimhinkle7245]

I want to do this exact thing. Thanks for your video.

[@fausto104]

The program looks awesome and I was gonna go check it out after your video but the fact that some words are banned is the flaw in this program. Art is suppoesed to push boundaries and make new ones for others to push further. The nprogram is cool but limited because of that I will not be trying out these programs until that flaw is fixed.

[@hyperbolicfilms]

You can get around a lot of it with creative language. I've since made some things with gore and nudity.

[@Candy_McK]

Stable Diffusion is not restricted.

[@fausto104]

@@Candy_McK I'll give it a look. Thanks.

[@StephanNeururer]

Great comic

[@bloodstreamcity]

I've published two comics so far using MidJourney, Lungflower and The Eldridge. I generate images in MidJourney and then tweak and assemble them in Photoshop. The technology still has a long way to go, and it's a challenge, but it's very possible to do.

[@mikkey246]

Amazing

[@elylioney6390]

That ending scared me

[@Cyterous]

Hey :) How did you succeed to consist the same character on different phases? we want to make a comics (manga actually) for a friend's BDay and not sure how to tell the AI to do his face in different situations. Thanks ahead !

[@hyperbolicfilms]

The best way to learn how to get consistency is to go to Campfire NYC's comics page, subscribe to their newsletter, and download Steve Coulson's Making of comic. It's 72 pages of his process with the latest version of Midjourney. There are many commercially available Stable Diffusion releases like Dreambooth that let you load a person in. I haven't used them, but people seem to get great results with uploading images.

[@kyounokuma]

Thanks for sharing. This is pretty cool. I especially like your tips for generating a series of consistent images. But I'm pretty sure Hollywood's gonna come calling if anyone is publishing books using an actor's image without permission. As a matter of fact, I wouldn't be surprised if they start to go after the AI platforms for usage rights violations. I think a big fight is coming.

[@hyperbolicfilms]

They should probably. There's a lot of press the last few days of Greg Rutkowski being upset that his art is being used. The companies should let him remove his art from their databases. Anyone should be allowed to do that. The training model can be trained on all of art history that is out of copyright and people who want their faces as part of the training data.

[@NicVandEmZ]

You should try the same thing with staple diffusion

[@wayneplatt5192]

This would make a cool topic for a course or tutorial...

[@NoCodeFilmmaker]

As a good artist myself I don't know what other artists are complaining about. I think they're just jaded because they feel threatened. To me this has all just been a fun and powerful TOOL to further expand and inspire my creative output. If they want to tear down Midjourney and the content created then they might as well attack Photoshop, Illustrator, etc. Because none of that is raw organic art. It's all creative vision assisted through powerful tools provided through the use of various platforms. I can imagine that similar arguments were posed by traditional filmmakers and analog editors at the onset of digital filmmaking and editing. As artists and storytellers we'll always have a job and place because unlike machines, we are creative storytelling beings and that can never be replaced. The ability to create coherent stories from abstract ideas. So these guys need to give it a rest and adapt or go extinct like the dinosaurs 💯

[@GuenevereSchwien]

To the comment about putting artists out of business. In my ONE day using Midjourney I find that it repeatedly creates the SAME woman's face. If you just watch the feel it really is quite repetitive, so I don't think illustrators will be out of a job.

[@porkpork2169]

midjourney needs a feature that allows people to feed the ai other images to go by along with the text based explanation so it has layers of details

[@elvisdeane44]

Their image-based prompts work much better in the new v4. But maybe too good in that I had a photo of actor A and asked it to make it look like actor B, and the results looked more like actor A. It also seems to keep the source clothing.

[@charleswang734]

I really appreciate to you coz the valuable vedio, AW, could I add u in my MJ friend list as I have a plan to create comic fiction LOL.

[@yusuftemmytope6597]

According to what have been seeing online on how to create an ai generated comic...human artist are still safe...cus i still feel ai can't get me what jorge Jimenez comic pages could look like if i want that style

[@Hencekevin]

i have a question, i have two characters and i want midjourney to make scenes with them but it change all the time the style of the characters, how do you do to mantein the style and the characters in the images?

[@hyperbolicfilms]

It is a very difficult thing to do well. I generally have used the names of actors in my prompts to get some consistency, then describe the clothes, and generate a lot of images. The best research on the subject of consistency is done by John Walter on Medium. He has many articles on the subject, and he's explored more than medium.com/@johnwalter-counsellor

[@VaibhavShewale]

seems like a wonderful option to me

[@billyliar1614]

I think like all technology it will be great used as a tool but if over-relied on as a shortcut will in my view only be detrimental to artistic expression. So much synthesised material out there and it shows. It takes years of dedication to really master an art. I guess it all boils down to existential questions which are above my pay grade, but can a machine really get to a point where it produces an authentic and original artistic response, and if it does what does that say about man vs machine and indeed the human ''soul'' ?

[@WhiteNoiseChannel111]

After you get a character from midjourney, how do you tell midjourney to use that character in different scenes? Say you need it once inside a kitchen and once inside a bedroom

[@RegularRegs]

I don't have the money or art skill to actually create my comic so I'm using midjourney. It's working. And I get to tell my story. Mix of comic life (1 month free trial.. Will buy for $30) , midjourney, gimp, and some gpt4 for tweaking dialogue and even helping keep continuity in my image prompts

[@chudum]

What is the name of the comic font you use. Love it!

[@hyperbolicfilms]

Berylium

[@MeNoOther]

Can we have something like this for cartoonist? Upload a model sheet of the artist’s characters, then type in a script/screenplay, then have the a.i. take the characters from the model sheets and arrange them in a comic book layout or in an animated story. People would be able to make the stories they come up with, without a major studio. That’s a practical use of a.i.

[@hyperbolicfilms]

That sounds like it would be great. I'm curious about the Image2Image function of StableDiffusion, where you draw something rough and it fills it in, whether it can grow into something like you're saying.

[@AscendantStoic]

@@hyperbolicfilms Image2Image is pretty great, I messed with it for the past few days, it has huge potential specially for artists to speed up their process significantly while also providing the consistency and specificity you mentioned (which is lacking in regular A.I art generation).

[@morizanova585]

Not sure if there are tools which already have capabilities like that . But if you want , it possible to train the AI so he can recognize the style or something like that . Search "AI Training " in youtube if you have spare time to understand and give a shot

[@upcdowncleftcrightc9429]

If you make art on midjourmey, do you own the art? Total freedom to use it and profit, etc? Anyone know?

[@Sams.Videos]

6:58 Precision is what every artist is looking for. Unfortunately, I've been wasting a lot of time trying to get there with Mid Journey, Dall-E and Stable Diffusion, but never reaching my vision. Those AI tools are way too random if you want to translate your idea into something tangible.

[@Candy_McK]

As an artist, I can edit everything the AI spits out. I tried MJ and DallE too, and currently, I'm using Stable Diffusion because I can run it on my own PC for free. For me, it's not a stand-alone tool but it boosts my work on backgrounds for my webcomic like hell.

[@Sams.Videos]

@@Candy_McK Interesting. Would you like to share some insights about your workflow? I'm curious an eager to learn some tricks! 🙂

[@Candy_McK]

@@Sams.Videos Did you get my response or has YT deleted it?

[@Sams.Videos]

@@Candy_McK Hi. Guess YT deleted it... I did not get your response... 😕

[@Candy_McK]

@@Sams.Videos Probably because I tried to send you some links... to other YT videos. 🤪 Check out the chanels of 'Bryan Fogaça Rosado Art', 'rossdraws' and Noah Bradly. They made useful videos about merging AI with their own art. I'm simply using 3D for my characters, AI for my backgrounds and textures, and stitch it together. A little color correcting, overpainting, photobashing, whatever it needs - ready to go.🤗

[@mikebrophy]

Two things you said that are spot on... (1) this thing will eat your brain. Hahaha! It totally has. And (2) "it's like having an infinite number of monkey artists." LMAO!!! It is!

[@DemonCrest250]

How do you layout panels using the AI prompt?

[@hyperbolicfilms]

The images were generated in Midjourney, but I did the layout using Comic Life 3.

[@EnricTeller]

Seems like you could create an absurdist comic based on what the bot can create. 😆

[@mikkey246]

I knew he was going to say he should go with the 30$ option lol, but this is interesting

[@tomwd.2825]

Well I found out, if the word dancing is used, you can have an action like scenery, in a way it could produce a running picture...

[@obikenobi1906]

this is hilarious!

[@67cans2]

If one looks through a perspective lens as of music, no one owns a single note.

[@TheKevphil]

_"Duck and cover!"_ Our heroine is buried alive by a host of Donald Ducks.

[@gabrielex]

I think one trick would be to use words that have just one meaning, for example duck can mean several things, better to use something like crouch. AI is not smart, so it needs very precise and incontrovertible input, it can't interpret things based on context that much.

[@hyperbolicfilms]

Great point. I tried a prompt with crouch instead to see, and it still gave mixed results. Limbs get jumbled, even in the newer test versions that have been better for action. But clarity of language is really important.

[@orsenal]

soylent green, anyone?

[@harry486]

Suddenly the early to mid 2000s do seem like a very long time ago....

[@sebmcc4047]

This is amazing. Which art style is this?

[@holdthetruthhostage]

This almost changes everything just Imagine if the AI can absorb Art styles & allow you to create Comics with an art style you want.

[@hyperbolicfilms]

I got some very, very good looking Superman images out of the new test version of Midjourney that is based on Stable Diffusion, but getting him to perform specific actions or having that art style stay consistent wasn't possible. But the image2image feature that improves a drawing might be interesting for getting specific action.

[@holdthetruthhostage]

@@hyperbolicfilms hey you truly helped from seeing your book I was thinking it's only good for realistic story telling for now or making a faceless story

[@DevinTheDude93]

This reminds me how in engineering it use to take rooms full of college educated people to create concepts and prototypes that can now all be replaced by 3d modeling and psychics simulators.

[@davidg3924]

The people running the software are engineers “ college educated people “ all that’s changed there is the toolset, AI to generate a work of “art “ is completely different, it’s removing the human from something that is by definition a human endeavor…

[@CrniWuk]

@@davidg3924 Well. That is and always was the whole point behind Ai - and the computer in general when you think about it. Honestly, what the Ai does now isn't exactly the new thing. They could, in theory do this for years. It's just that two things have over time changed. The processing power and the easy access to large quantities of information and data for the Ai to learn from. The Ai needs in some cases access to millions of information, images, you name it. I mean google started training their Ai like what? 10 or 15 years ago? Back then their only target was to teach it how to recognize a "cat". The principle behind that isn't very different from what we see now. Ai right now is to intellectual work what the steam engine once was to physical labour. It means that it has the potential to multiply the efficiency of what one person can do. 1 or 2 creatives can now do the work of a whole team in a day which might have taken them weeks before. Imagine that. A whole advertising campaign that took a month, done in a day. At the fraction of the cost and labour. And that's just one industry. Pretty much all jobs will be affected by this one way or another. I am afraid we have not seen anything yet.

[@Uratz]

There's a lot of Talented artists that are not doing it because they fear too much.

[@beaverson]

Probably because most artists like the whole process of creating, and don't wanna hand the process over to soulless machines..

[@moali4085]

Out of curiosity, how did you keep your character's appearance constant throughout the different image generations?

[@hyperbolicfilms]

Using actor's names was step 1. But what I've learned now from using Midjourney v4 is that keeping a character consistent is easier when the results of the artwork are hazy and not so clear. You can see in some other people's more recent comic work that sometimes the clothing changes panel to panel, or the hair changes a bit, because the AI doesn't know to be consistent. So as I prepare to make the next issue of Goats, I'm looking at adding some artistic style to the shots that fuddle the results enough to make a person match from shot to shot. Prompt not only an actor's name, but the clothes and hair they have, and add an artistic style that will blur the lines just enough to keep things vague. Watercolors are great for this.

[@yume1243]

I read an article with the creator of midjourney. I believe that acts of violence especially photorealistic depictions are not produced by midjourney by design . Ai generated artwork is not protected by copyright unless it has a human element implemented into it. I would say that your comic would be protected since you added text. And many people do refinements and paint overs so I believe copyright will not be an issue.

[@RM_VFX]

The question of who owns the art is clear in MJ's terms of use, both parties own some rights. The question is, how much is the software really changing some of the source material? I've seen some images based on art I recognize that are almost not transformative at all.

[@hyperbolicfilms]

I really think it would be helpful if there was a way to see what images the AI used as inspiration for the final art it delivers. Opening up the training model so that artists can opt out of having their work used, like a KZbin copyright claim.

[@defrank1870]

Wait until it can do animation.

[@superanimator1604]

Fuck no

[@rogersinger2161]

how did you get the consistency of style???

[@hyperbolicfilms]

Midjourney seems to have a default style. I generally gave a :: sad tag in my prompts, and that was enough to get a consistent look

[@rogersinger2161]

@@hyperbolicfilms thank you. I'll give that a try.

[@Sams.Videos]

As an artist you will always have to make order out of chaos, make sense out of nonsense. That is what AI can not do with those image generators. AI comes up with nonsense, it is the duty of the artist to make sense by ordering all the nonsense into something coherent.

[@zakj3769]

your comic was amazing and the story really worked, it was very dream/ dark comics/murakami-esque, and nice ending, can't believe you don;t have more views! just one question, how did u get the same style in all of them? its hard to be consistant no?

[@hyperbolicfilms]

Thanks! I tried to make the story work with the dreamy art, because there is no way to get around the inconsistencies at the moment. I think having a consistent set of tags after your character, description, action, and location helps keep the same style. I tagged almost everything with :: sad, which seemed to give it a fairly consistent mood and framing. Midjourney V3 does seem to produce a specific look when you don't give it a lot of stylistic direction.

[@Babybear69]

Just did my own then I found this video.

[@goldenplayer4966]

Maybe in the future we get to work watching content on-line. With Soo much content available people don't know what's good, so your going to be paid to watch things and give feedback, as AI can't really watch and say if something is good of not (for now). Not the dream Job but people with knowledge of having watching tons of content on the internet may be useful in the situation, I wouldn't mind to work as someone who is paid to evaluate works.