I think the future of creating stories looks like this: You still draw a manga, but only the storyboard. You make it look "decent" enough, just capturing the proper characters, poses, emotion, and whatever necessary details of the scene. Again, only fast clean sketches. Then, once your story is sketched out, we will be able to use AI to analyze the image and create full detailed manga pages based on our outlines. So it's like having an assistant but even crazier because instead of just shading / doing backgrounds for you, it would even be able to make your drawings better. Now of course, the better we are at drawing, the better results we should get from the AI's because the closer we can translate our imagination to reality, but still this would be insane right?
I don't have the fear to fail. I tried to use AI for writing, and it already is pretty useful. Sometimes I think, if a scene could be written by AI, it might just be skipped altogether. Instead of letting an AI write, I'd rather write only the points I would give the AI as minimal as possible. Instead of telling the AI to write a filler scene, where the main character travels from A to B and meets person C on the way, I would just write a few sentences like "Then he left A to reach his next destination B. On his way he got to know C, (maybe some details), and now he's already there". But sometimes I need to write something, so as a first draft or for ideas, AI might be useful. But for art, AI doesn't work good enough. But it's difficult to get exactly what I want, especially if what I want doesn't exist a hundred times already. And if I want the same person wearing different clothes or in different poses,, it's pretty difficult. Looks like you found some tool for that. I didn't have success with midjourney yet. But where I found AI useful is text to speech. I have a lot of dialogs in my story, so I wanted it to be spoken by AI, and even if it doesn't have emotions and different intonations, I'm pretty happy with it so far.
Good for you for not having fear of failure. If AI art isn't good enough for your use case today, maybe someday it will. Every week there is a new update. Chances are, we will get there. In my opinion, we should just do what we can with the tools available, but drawing by hand is always an option.
@@evan-ezquer I'm still figuring out what's the best approach to get my characters. I tried drawing, modeling, AI and vector graphics. Each has different strengths and weaknesses. And some mix is always possible, too.
@@porky1118 I think going SD route has the most control over the characters. But the problem is you'd have to create a LoRA model for each character, which makes it a time-consuming process. With MJ, you just generate 1 character, and then you can use that as a reference for various poses. If you don't have a lot of characters, maybe using SD will suit you better. Anyway. It's great to have someone like you who's also doing the heavy lifting of figuring things out. I'd like to formally invite you to my community so we continually discuss these things and swap ideas.
@@evan-ezquer SD means self drawn I guess. I probbaly have hundreds of characters. Most of my stories are short stories of around 10000 words, and they have around 5 main characters. And I usually want them in different clothes at least. I'm not good enough at drawing. I can draw good, but it takes a lot of time. Especially when I want clean lines, this might get a little difficult. So maybe if I stick with a more sketchy style, That's why I tried VG, so I get at least clean outlines and can easily change the style (like changing the outline style or removing them altogether) or the proportions. Maybe I have to try the charactre reference thing again. I saw the feature, but I wasn't happy with it. But I generated the first variant of the character with another AI, so the variants looked pretty different. For example I always try to generate a feminine young boy. But often I just get a girl instead. If I add terms like "curvy" instead of feminine, I often got oversized breasts 😅 Yeah, I guess I should join the discord.
AI bias sucks. When I try to generate a "regular character", average guy and all, and AI always wants to give me guys with big guns and sixpacks lol. What's VG? Anyway, we can talk more on Discord.
I agree with you whole heartedly. As a programmer working in AI, and someone who's been trying to finish a story for 13+ years, this approach has helped me finally finish some of the storylines I need planned out to finish writing. Like you said, AI can't write a very good book, but it is great for giving you suggestions and ideas of direction, and countless times I don't even use them, but the prompt me to think of something better.
🌟 Join my community: discord.gg/RE2SrNhxQV
✍ Sign up to my newsletter: neoblush.ck.page/storyteller-group
I think the future of creating stories looks like this: You still draw a manga, but only the storyboard. You make it look "decent" enough, just capturing the proper characters, poses, emotion, and whatever necessary details of the scene. Again, only fast clean sketches. Then, once your story is sketched out, we will be able to use AI to analyze the image and create full detailed manga pages based on our outlines. So it's like having an assistant but even crazier because instead of just shading / doing backgrounds for you, it would even be able to make your drawings better.
Now of course, the better we are at drawing, the better results we should get from the AI's because the closer we can translate our imagination to reality, but still this would be insane right?
💯 I think this is where things are headed too!
well that was informative, i think i will give it a try
Thank you for this video!
I'm also looking to create comics with AI.
I'm waiting for your next videos.
Glad to have you partner.
Just subscribed. I'm fascinated by this subject and interested in seeing how younger creators are adapting to this new technology. Great content.
Welcome aboard! Grateful to have you.
I don't have the fear to fail.
I tried to use AI for writing, and it already is pretty useful.
Sometimes I think, if a scene could be written by AI, it might just be skipped altogether. Instead of letting an AI write, I'd rather write only the points I would give the AI as minimal as possible.
Instead of telling the AI to write a filler scene, where the main character travels from A to B and meets person C on the way, I would just write a few sentences like "Then he left A to reach his next destination B. On his way he got to know C, (maybe some details), and now he's already there".
But sometimes I need to write something, so as a first draft or for ideas, AI might be useful.
But for art, AI doesn't work good enough. But it's difficult to get exactly what I want, especially if what I want doesn't exist a hundred times already.
And if I want the same person wearing different clothes or in different poses,, it's pretty difficult. Looks like you found some tool for that. I didn't have success with midjourney yet.
But where I found AI useful is text to speech. I have a lot of dialogs in my story, so I wanted it to be spoken by AI, and even if it doesn't have emotions and different intonations, I'm pretty happy with it so far.
Good for you for not having fear of failure. If AI art isn't good enough for your use case today, maybe someday it will. Every week there is a new update. Chances are, we will get there. In my opinion, we should just do what we can with the tools available, but drawing by hand is always an option.
@@evan-ezquer I'm still figuring out what's the best approach to get my characters.
I tried drawing, modeling, AI and vector graphics. Each has different strengths and weaknesses. And some mix is always possible, too.
@@porky1118 I think going SD route has the most control over the characters. But the problem is you'd have to create a LoRA model for each character, which makes it a time-consuming process. With MJ, you just generate 1 character, and then you can use that as a reference for various poses. If you don't have a lot of characters, maybe using SD will suit you better. Anyway. It's great to have someone like you who's also doing the heavy lifting of figuring things out. I'd like to formally invite you to my community so we continually discuss these things and swap ideas.
@@evan-ezquer SD means self drawn I guess.
I probbaly have hundreds of characters.
Most of my stories are short stories of around 10000 words, and they have around 5 main characters.
And I usually want them in different clothes at least. I'm not good enough at drawing. I can draw good, but it takes a lot of time.
Especially when I want clean lines, this might get a little difficult. So maybe if I stick with a more sketchy style,
That's why I tried VG, so I get at least clean outlines and can easily change the style (like changing the outline style or removing them altogether) or the proportions.
Maybe I have to try the charactre reference thing again. I saw the feature, but I wasn't happy with it. But I generated the first variant of the character with another AI, so the variants looked pretty different.
For example I always try to generate a feminine young boy. But often I just get a girl instead. If I add terms like "curvy" instead of feminine, I often got oversized breasts 😅
Yeah, I guess I should join the discord.
AI bias sucks. When I try to generate a "regular character", average guy and all, and AI always wants to give me guys with big guns and sixpacks lol. What's VG? Anyway, we can talk more on Discord.
Keep up the good work looking forward to the manga comic course
Thank you for your support! Means a lot. Saw you in the group too. I'm excited to see your creations soon :)
I agree with you whole heartedly. As a programmer working in AI, and someone who's been trying to finish a story for 13+ years, this approach has helped me finally finish some of the storylines I need planned out to finish writing. Like you said, AI can't write a very good book, but it is great for giving you suggestions and ideas of direction, and countless times I don't even use them, but the prompt me to think of something better.
Spot on! Most of the time you don't copy the outputs, but they make you think of something better.
Man I have watched all of your videos in a binge unknowingly lmao
I hope you learned something 😉