I remembered a recent tweet from @AuthorJMac: “You know what the biggest problem with pushing all-things-AI is? Wrong direction. I want AI to do my laundry and dishes so that I can do art and writing, not for AI to do my art and writing so that I can do my laundry and dishes."
Tbf a world where tech is cheaper than simply doing the laundry ourselves or hiring a minimum wage person makes no sense, at that point we'd basically have high tech robot servarts and we'd be using them for that?
Your final model looks awesome; new sub. I'm a digital designer, I have used AI before to generate assets for some of my work, but it always looks cheap / monotonous if you have a decent eye; and sometimes it feels impossible to get what one has in mind, even if it's something simple. Personally, I think everything that is made with AI should not have any copyright and should be always tagged as an AI generated asset, so everyone can use it for free and companies don't have any ownership over "their creations" (so it discourages to replace actual artists, designers, etc).
As a game developer, I am happy because I will be able to generate 3D assets for my games quickly. However, as a 3D generalist, I am concerned that this could take away jobs from talented artists in the future.
could you make a dedicated video on my VFX is so expensive, i understand its complicated which is why im so interested, im sure the video would be quite long
also what about ai and scripts. gpts can write scripts, but thats pretty new, could u make a script where u just paint the absolutely necessary topolgy lines with the paintbrush, and then blender automatically fills those lines in how u want them and the rest is left to automatic topology that fits into those vertices? this would be around the eyes mouth across the clavical near the hip bone, joints like ankles elbows around the fingers. as long as ur painted lines get these bits right automatically generating the rest from them would land a pretty good topology especially if u already remeshed and told it to leave the lines alone that arnt painted and just find a way to connect them. or is that to much to hope for with a script in blender?
So the biggest time save comes from the concept stage. Using ai to iterate through concepts is obviously a lot faster than having to do the concept work on your own. When it comes to actually modeling though, the ai model only saves you about 30mins to an hour of blocking at best.
if you were making game characters that were anime cel shaded it might save some time because colouring that in blender is as simple as selecting faces, choosing your colour map, picking a colour and clicking the model so u can paint it fast enough and then add an anime shader and bam simple anime. a little bit harder if u want to add non pixelated outlines and cellshading details but probably not as much as having to do texttures for an entire model and since making anime style 3d models has only really been possible for the last 2 to 4 years that im aware of and is different to the normal workflow the market is completely undersaturated compared to realism
hyperhumans rodin ai is really amazing, can u tell us if u think that is better than meshy? it would be hard pressed not to be. it does quads, allows for remeshes, has a sculpt tool u can use after confirming your model so u can sculpt in the web ui. it does textures and stuff really well but the uv maps are funny because its ai is the one downside for characters that need to be animated
true, i started learning drawing and digital art in 2020 to become a concept artist/matte painter and ho boy i literally don't know what to expect from the future i am struggling every day to stay somewhat positive
@@PlatedGuy What I mean is that the process, is depressing. This guy doesn't even realize he is leaving the machine the good part and he is doing the boring/autonomous part. It's contrary to being an artist. Is being someone that doesn't even think of what he is doing.
Why do you say it's not good for hollywood production? Looks good to me, but would also like to know what it is still missing that big prods need? Thx!
So I'm still satisfied with the overall look, but I think a lot of the model could use some cleaner shape design. The textures are very rough and could definitely use more work. Skin shader is a little off, especially for close ups. The clothes model is not great, the details are very loosely blocked in and the shirt isn't properly tucked under the shawl. More importantly though, for actual studio work, we'd need to properly topologize the face for deformations, add an actual hole and inner model for the mouth, cornea and caruncle model for the eyes...etc..
@@crafthive I'm just watching a movie from couple of years ago Occupation:Rainfall. A modern, sci-fi, mid budget shlock. There are aliens in it, similar to the one you've made. Would you mind having a look at those aliens in that film and share with me your thoughts ( no need to watch whole film, just a screenshot, thought if you'd like to watch it it is on tubi platform, free)? I'm very curious what you think about them, because to my eyes, quality is very similar to your work. thx
@@selenophile5256 There will probably be a lot less 3D modeling jobs specifically.. but VFX is more than modeling. And you're still going to need some 3D modelers to modify the outputs.
Well the thing is, you have to remember, right now we have the worst version of AI. It is only getting better and better. So you will have to spend less and less time refining its output. Also I have just watched another video from you, about the state of the VFX industry. Just imagine, instead of doing 300 revisions of pixel f*ing, you just create a scene/shot and then the middle management can torment the AI to move stuff and changes hues.
As a VFX artist myself, I appreciate the work put into this to create valuable and relevant content! I do sense that your opinion is naturally biased though at 2:40. It's definitely something you could do in 10-15 min in zbrush right now instead of using Meshy or TripoAI, but I'm phadomed by how artists across the world don't see the incredible value of iteration this technology brings. 2D is already at a good production state (in only 1-2 years of development) as you demonstrated cycling through dozen of alien design ideas in less than 10 minutes. So give it another 1-2 years and 3Dgen will likely generate a full PBR asset in minutes from text better quality than any artist who spent 10 years practicing their craft and in any style, and that's a GOOD thing. I just think we could live in a 10x better designed world (maximizing design decisions) if we embraced generative intelligence rather than being precious about IP as if artists out of everyone own anything anyway. Not to say that each genAI can be personalised to each artist's unique creative output through fine-tuning. Artists are not only most often hired by a client who owns the work, but even if not, all they produce is an organic remix of what they've experienced and invested time into acquiring the skill to realise. Claiming that whatever artists make is 'theirs' is a foolish mentality adopted for survival in a world reined by capitalism. Not trying to wrestle with anyone, just speaking my mind hoping to engage in an open minded discussion.
And what will happen those 10 year skill honed artists are out of a job worldwide ? And how empty would feel to let the machine come up with original ideas while you do the boring parts ?
@@selenophile5256 It's just a natural job displacement due to tech advancements. Whoever doesn't adapt and reinvent themselves will naturally run out of a job, like it happened with the digital, industrial, agricultural and other revolutions before. We can't just keep the world stagnating just so people keep doing the same thing the same way. To not go too far on the left side though, is important considering that absolute progress is not always good and calculating the implications that it has on society. I think is unhealthy to divide machine and human creativity as if we don't already use the internet for references and concepts, photobashing, kitbashing etc in our creative process. The boring parts like topology, uvs, etc are only still part of the job because companies pipelines and workflows are 10-20 years old, but they can already be automated. Not to say that the whole 3d industry can be made redundant in VFX and other industries when AI video models can generate in minutes the same shots as what producers pay hundreds of thousands or millions to studios with thousands of people for months if not years of work. So traditional 3D may transition out of mainstream production and potentially stay preserved as a high value craft like traditional arts.
As a 3D artist myself, I found this video very biased, to be honest. I assume there is a significant amount of copium involved. Don't look at AI as a competitor, look at it as a tool. Then you'd probably be more honest in your reviews.
I remembered a recent tweet from @AuthorJMac: “You know what the biggest problem with pushing all-things-AI is? Wrong direction.
I want AI to do my laundry and dishes so that I can do art and writing, not for AI to do my art and writing so that I can do my laundry and dishes."
Tbf a world where tech is cheaper than simply doing the laundry ourselves or hiring a minimum wage person makes no sense, at that point we'd basically have high tech robot servarts and we'd be using them for that?
By the time an a.i. generated image is edited enough to be usable, you might as well have done it from scratch anyway.
Your final model looks awesome; new sub.
I'm a digital designer, I have used AI before to generate assets for some of my work, but it always looks cheap / monotonous if you have a decent eye; and sometimes it feels impossible to get what one has in mind, even if it's something simple. Personally, I think everything that is made with AI should not have any copyright and should be always tagged as an AI generated asset, so everyone can use it for free and companies don't have any ownership over "their creations" (so it discourages to replace actual artists, designers, etc).
Yes that's correct, AI generated stuff are garbage, but the problem is... people does not care, they more and more ok with garbage.
nobody's taking anybody's job, the film and games industry can't afford to hire people and ai is a convenient scapegoat for their failures.
Is it even a tablet if you don't reinstall the drivers 3 times a week??? :D
XD
As a game developer, I am happy because I will be able to generate 3D assets for my games quickly. However, as a 3D generalist, I am concerned that this could take away jobs from talented artists in the future.
This is a really cool and well done video, would love to see more like this!
AI doesn't match up things :) your work is beautiful as you always
could you make a dedicated video on my VFX is so expensive, i understand its complicated which is why im so interested, im sure the video would be quite long
also what about ai and scripts. gpts can write scripts, but thats pretty new, could u make a script where u just paint the absolutely necessary topolgy lines with the paintbrush, and then blender automatically fills those lines in how u want them and the rest is left to automatic topology that fits into those vertices? this would be around the eyes mouth across the clavical near the hip bone, joints like ankles elbows around the fingers. as long as ur painted lines get these bits right automatically generating the rest from them would land a pretty good topology especially if u already remeshed and told it to leave the lines alone that arnt painted and just find a way to connect them. or is that to much to hope for with a script in blender?
how much time did you save doing it this way over from scratch? Your final product looks good enough for a video game.
So the biggest time save comes from the concept stage. Using ai to iterate through concepts is obviously a lot faster than having to do the concept work on your own.
When it comes to actually modeling though, the ai model only saves you about 30mins to an hour of blocking at best.
if you were making game characters that were anime cel shaded it might save some time because colouring that in blender is as simple as selecting faces, choosing your colour map, picking a colour and clicking the model so u can paint it fast enough and then add an anime shader and bam simple anime. a little bit harder if u want to add non pixelated outlines and cellshading details but probably not as much as having to do texttures for an entire model and since making anime style 3d models has only really been possible for the last 2 to 4 years that im aware of and is different to the normal workflow the market is completely undersaturated compared to realism
AI should rather be giving us perfect topo and UVs :'(
hyperhumans rodin ai is really amazing, can u tell us if u think that is better than meshy? it would be hard pressed not to be. it does quads, allows for remeshes, has a sculpt tool u can use after confirming your model so u can sculpt in the web ui. it does textures and stuff really well but the uv maps are funny because its ai is the one downside for characters that need to be animated
that's depressing
true, i started learning drawing and digital art in 2020 to become a concept artist/matte painter and ho boy
i literally don't know what to expect from the future i am struggling every day to stay somewhat positive
@@PlatedGuy What I mean is that the process, is depressing. This guy doesn't even realize he is leaving the machine the good part and he is doing the boring/autonomous part. It's contrary to being an artist. Is being someone that doesn't even think of what he is doing.
Why do you say it's not good for hollywood production? Looks good to me, but would also like to know what it is still missing that big prods need? Thx!
So I'm still satisfied with the overall look, but I think a lot of the model could use some cleaner shape design. The textures are very rough and could definitely use more work. Skin shader is a little off, especially for close ups. The clothes model is not great, the details are very loosely blocked in and the shirt isn't properly tucked under the shawl. More importantly though, for actual studio work, we'd need to properly topologize the face for deformations, add an actual hole and inner model for the mouth, cornea and caruncle model for the eyes...etc..
@@crafthive So basically you've created an extra ;)
@@crafthive I'm just watching a movie from couple of years ago Occupation:Rainfall. A modern, sci-fi, mid budget shlock. There are aliens in it, similar to the one you've made. Would you mind having a look at those aliens in that film and share with me your thoughts ( no need to watch whole film, just a screenshot, thought if you'd like to watch it it is on tubi platform, free)? I'm very curious what you think about them, because to my eyes, quality is very similar to your work. thx
what if there will be sora like quality for 3d model?
It's definitely headed there. I'm sure we'll be seeing that sooner than we think, and that will have a much larger impact.
@@crafthivedo you think the majority of VFX artists would survive ?
@@selenophile5256 There will probably be a lot less 3D modeling jobs specifically.. but VFX is more than modeling.
And you're still going to need some 3D modelers to modify the outputs.
clever title
Great job👏
Well the thing is, you have to remember, right now we have the worst version of AI. It is only getting better and better. So you will have to spend less and less time refining its output. Also I have just watched another video from you, about the state of the VFX industry. Just imagine, instead of doing 300 revisions of pixel f*ing, you just create a scene/shot and then the middle management can torment the AI to move stuff and changes hues.
As a VFX artist myself, I appreciate the work put into this to create valuable and relevant content!
I do sense that your opinion is naturally biased though at 2:40.
It's definitely something you could do in 10-15 min in zbrush right now instead of using Meshy or TripoAI, but I'm phadomed by how artists across the world don't see the incredible value of iteration this technology brings.
2D is already at a good production state (in only 1-2 years of development) as you demonstrated cycling through dozen of alien design ideas in less than 10 minutes.
So give it another 1-2 years and 3Dgen will likely generate a full PBR asset in minutes from text better quality than any artist who spent 10 years practicing their craft and in any style, and that's a GOOD thing.
I just think we could live in a 10x better designed world (maximizing design decisions) if we embraced generative intelligence rather than being precious about IP as if artists out of everyone own anything anyway. Not to say that each genAI can be personalised to each artist's unique creative output through fine-tuning.
Artists are not only most often hired by a client who owns the work, but even if not, all they produce is an organic remix of what they've experienced and invested time into acquiring the skill to realise. Claiming that whatever artists make is 'theirs' is a foolish mentality adopted for survival in a world reined by capitalism.
Not trying to wrestle with anyone, just speaking my mind hoping to engage in an open minded discussion.
And what will happen those 10 year skill honed artists are out of a job worldwide ?
And how empty would feel to let the machine come up with original ideas while you do the boring parts ?
@@selenophile5256 It's just a natural job displacement due to tech advancements.
Whoever doesn't adapt and reinvent themselves will naturally run out of a job, like it happened with the digital, industrial, agricultural and other revolutions before.
We can't just keep the world stagnating just so people keep doing the same thing the same way.
To not go too far on the left side though, is important considering that absolute progress is not always good and calculating the implications that it has on society.
I think is unhealthy to divide machine and human creativity as if we don't already use the internet for references and concepts, photobashing, kitbashing etc in our creative process.
The boring parts like topology, uvs, etc are only still part of the job because companies pipelines and workflows are 10-20 years old, but they can already be automated.
Not to say that the whole 3d industry can be made redundant in VFX and other industries when AI video models can generate in minutes the same shots as what producers pay hundreds of thousands or millions to studios with thousands of people for months if not years of work.
So traditional 3D may transition out of mainstream production and potentially stay preserved as a high value craft like traditional arts.
As a 3D artist myself, I found this video very biased, to be honest. I assume there is a significant amount of copium involved.
Don't look at AI as a competitor, look at it as a tool. Then you'd probably be more honest in your reviews.