Creating with prompts as a workflow sucks if you are a visual artist though. Maybe it's motivational if you are more of a writer and you lack the skills to draw, sculpt, paint and compose, light etc. But I think many artists like the process of creating art with their hands and eyes. AI takes a lot of the fun out of the creative process, and I think that is depressing to a lot of artists, even if you could still make a livelihood using AI. Besides, promt based AI art creation is more like the client comissioning the work, and the software become the artists. If you work as an illustrator, why should the client use you as a middle man?
BTW, IMO there's nothing ethical about Adobes approach to AI. Yes, they might legally own the rights to train AI on their stock libraries, but did the artists know about this or read the fine print when they submitted their art to Adobe many many years ago, even before the current forms of AI art was invented? Now their stock images will compete against AI generated art trained on their images, and the artists will see diminishing returns on their royalties, while Adobe will likely be able to keep more of their revenue.
@@JayantSikka Wow, that is a very long reply to an old post! Did you use ChatGPT or AI to write that? Anyways, I think unless a good legal framework around AI and copyright is going to stop it, the future looks bad for artists, writers, musicians, and a lot of the creative industries.
I really love the manual labour element of 3d modelling, I like looking at an objects detail and modelling them down to the seam where the object was welded and I've been this way for 2 decades, going to miss these days.
This is along the lines of how I have been trying to think lately. I definitely see a lot of potential for AI to speed up my workflow and help me to accomplish things that were just way too overwhelming previously. But it's still tough since the changes are happening so fast. I lean towards more towards animation than anything else but seeing what happened to digital art freaked me out. So thanks for the perspective.
26:40 Woohoo it's me :) Thank you Andrew for this really insightful and very smart analysis of the AI future. I 100% agree on your take and am rather excited what lies ahead.
Love this episode and Q&A. Great picked questions with thoughtful answers. While I definitely think polygon replacements like voxels are helpful in the sculpting, modeling, and simulating workflows, I think rendering will stay with triangles for a while longer. During UEs tech demo and presentation on Nanite, they actually discuss several rendering mediums (polygon, voxel, point cloud, etc.) with the pros and cons of each. In the end, they concluded that triangles were still the best option, so that's what they did with Nanite (also, doing anything else would be slow to adopt since so much is based on triangles now).
The problem today is to create meaningful images that turn on emotions on the viewers according to the creator intention. Simply beauty images are going se their value down to northing. Explaining images will become more important to justify what's behind.
I just don't see it, midjourney has been art directing itself, I think it will be so abstracted and made simple that it'll be a chore many companies will just tell the Intern to do. Perhaps like "hey kid can you feed the machine something, anything, everything comes out good", and boom! Something unique.
Keep in mind AI will never replicate or replace the motivation garnered from other artists, true blue inspiration motivating countless people to go forward with their own art, in any form.
The part about rendering with a crappy computer, while your analogy is fair and makes sense, I wholeheartedly understand that person's position. At one point you start getting really slowed down by constantly having to wait for the computer to do it's thing and spit out an image so you can adjust the scene further. And if you do that for long enough, it honestly makes you lose the passion for rendering. Nowadays, I'd recommend looking at used rtx 2060/2070 or even 3060 12gb GPUs for 200-300 euros/usd, since you will get optix which will speed up the render even if rest of the computer isn't the fastest thing. I've spent countless hours rendering big scenes on a ryzen 5 1600x because my gpu wasnt supported, and I couldnt upgrade at the time due to mining boom, which made me just stop rendering all together even though I wanted to get better and pursue it as potential full time job. I am happy with what I managed to produce with that computer, but looking back at the process and the amount of time any little change took, I do not miss it.
While I agree that workflows are being replaced, I think artists should be paranoid because the motivation behind changing workflows is efficiency and profit. Auto-gpt and agents are already showing signs that the higher level process of art directing can be replaced. However I think there actually is an abundance of motivation. People don’t get motivated just by positive things. It can be from fear too. Imagine this chaotic period being where AI just becomes useful enough to make you a fortune and not great enough to be AGI and making work obsolete, it is imperative to learn fast and act fast to have a chance for better tomorrow.
I agree with justin sooroojdeen - while this is efficiency and profit is definitely pro-corporation, there's enough reason to be optimistic and not paranoid - as Andrew says - as long as you're willing to learn, grow, and adapt with the technology. With the photography analogy, it certainly added efficiency and profit. Many people in simpler use cases could just take pictures themselves - perhaps as it should be. But then, people also realized how much skill a _good_ picture requires
Thx for the timestamps. Best and worst aligns with my own experience. And yes, studios holding out due to "old guard" and most common excuse is the rigging and animation systems. That said, working well for games IMO.
This Q&A made me contemplate to get twitter😂 I've noticed that the Poliigon addon doesn't really work plug and play if the Colormanagement is set to AgX. Thought of mentioning this issue.
@@hombacom AI is a copy machine. Somewhat like a human, but with the brute force of a machine. Instead of just doing copies it tries to learn from them, and do better. But it's a machine, and when posed with a new problem like different poses, movement, (hands are almost a entire different systems from the body, they are difficult thing for humans too) or upside down pictures, AI literally has no idea what it is. It's trying to learn but it's a machine. "Is that a human?" "Nah, that's not a face" "Could it be a face upside down" "There is no such thing"
@@xanksauri89 i disagree about ”tries to learn from them and do better”. It just mimic shapes and patterns with randomness on top. Its Interesting that it can do impressive images but still doesnt have a clue what it is in reality, it doesnt visualize 3d in mind. A human maybe struggle with upside down and just simply rotate screen. We can judge result and start over, therefor we control ai tools and make it more Interesting.
He did answer this, in a roundabout way, when talking about how Poliigon watches AI. Currently AI can be good with some prompts but not others, and not reliably. So AI may be a good starting point, artists will still need to know how to model to art direct or tweak whatever the AI prints out. I'd say it's still worth learning the techniques even if you don't apply them regularly. It helps you be confident in making changes if needed.
The way to adapt to AI is to get out of digital art yesterday. Get as far away from that highly rendered CGI style as you can, it's worthless now as people will be (are) deensitized to it due to it being everywhere. Keep your art off the internet and if you have to put it on Glaze or Nightshade it. Sell yourself online instead your opinions your personality, draw/make art in videos dont post finished results as files that AI can take. Thats how you adapt, not by learning to prompt aka be AI's bitch.
I saw a video about generating 100 ebooks to sell it on different plattforms. I think we will soon have a problem with too much of all. Saturation in all corners, even if you want support other creative human beings, it is hard to find out if there is one behind.. What we maybe need is, that ai art and assistance should be marked. I dont pay for things, made by 4 words, trained illegal by real artist input. Sorry, i am human. I like ai but not in every aspect of my life ( even we will be there in few years ).
everyone can photograph but doesn't mean everyone is good or creative.. only a handful will be.. the same with AI art, it is just a different set of skills, expertises and perspectives. AI expands the creative and mental horizon for those who are of that temperament and thinking type but yeah perhaps not great for those who are more technical and have honed their skills for years to draw something to perfection..
I want to earn, I was naive and went for an art school I may start it within few months, I'm done for, yes art is lovely but now what, can't change my course my country don't work that way T_T
Creating with prompts as a workflow sucks if you are a visual artist though. Maybe it's motivational if you are more of a writer and you lack the skills to draw, sculpt, paint and compose, light etc. But I think many artists like the process of creating art with their hands and eyes. AI takes a lot of the fun out of the creative process, and I think that is depressing to a lot of artists, even if you could still make a livelihood using AI.
Besides, promt based AI art creation is more like the client comissioning the work, and the software become the artists. If you work as an illustrator, why should the client use you as a middle man?
BTW, IMO there's nothing ethical about Adobes approach to AI. Yes, they might legally own the rights to train AI on their stock libraries, but did the artists know about this or read the fine print when they submitted their art to Adobe many many years ago, even before the current forms of AI art was invented? Now their stock images will compete against AI generated art trained on their images, and the artists will see diminishing returns on their royalties, while Adobe will likely be able to keep more of their revenue.
@@JayantSikka Wow, that is a very long reply to an old post! Did you use ChatGPT or AI to write that?
Anyways, I think unless a good legal framework around AI and copyright is going to stop it, the future looks bad for artists, writers, musicians, and a lot of the creative industries.
I really love the manual labour element of 3d modelling, I like looking at an objects detail and modelling them down to the seam where the object was welded and I've been this way for 2 decades, going to miss these days.
This is along the lines of how I have been trying to think lately. I definitely see a lot of potential for AI to speed up my workflow and help me to accomplish things that were just way too overwhelming previously. But it's still tough since the changes are happening so fast. I lean towards more towards animation than anything else but seeing what happened to digital art freaked me out. So thanks for the perspective.
26:40 Woohoo it's me :)
Thank you Andrew for this really insightful and very smart analysis of the AI future.
I 100% agree on your take and am rather excited what lies ahead.
Andrew: "Photogrammetry isn't art"
His last video title: "#40: The Art of Photogrammetry"
(I jest, both are great videos)
37:20 The freeway analogy is called induced demand in economics. I dont believe its true on all cases but it’s an interesting concept.
Love this episode and Q&A. Great picked questions with thoughtful answers.
While I definitely think polygon replacements like voxels are helpful in the sculpting, modeling, and simulating workflows, I think rendering will stay with triangles for a while longer.
During UEs tech demo and presentation on Nanite, they actually discuss several rendering mediums (polygon, voxel, point cloud, etc.) with the pros and cons of each. In the end, they concluded that triangles were still the best option, so that's what they did with Nanite (also, doing anything else would be slow to adopt since so much is based on triangles now).
The problem today is to create meaningful images that turn on emotions on the viewers according to the creator intention. Simply beauty images are going se their value down to northing. Explaining images will become more important to justify what's behind.
I just don't see it, midjourney has been art directing itself, I think it will be so abstracted and made simple that it'll be a chore many companies will just tell the Intern to do. Perhaps like "hey kid can you feed the machine something, anything, everything comes out good", and boom! Something unique.
Keep in mind AI will never replicate or replace the motivation garnered from other artists, true blue inspiration motivating countless people to go forward with their own art, in any form.
I call on your bullshit, Try to give that argument when an artist is flipping burgers to pay bills
The part about rendering with a crappy computer, while your analogy is fair and makes sense, I wholeheartedly understand that person's position.
At one point you start getting really slowed down by constantly having to wait for the computer to do it's thing and spit out an image so you can adjust the scene further. And if you do that for long enough, it honestly makes you lose the passion for rendering.
Nowadays, I'd recommend looking at used rtx 2060/2070 or even 3060 12gb GPUs for 200-300 euros/usd, since you will get optix which will speed up the render even if rest of the computer isn't the fastest thing.
I've spent countless hours rendering big scenes on a ryzen 5 1600x because my gpu wasnt supported, and I couldnt upgrade at the time due to mining boom, which made me just stop rendering all together even though I wanted to get better and pursue it as potential full time job.
I am happy with what I managed to produce with that computer, but looking back at the process and the amount of time any little change took, I do not miss it.
Andrew actually said Arts will not be replaced, but the artists will be.. "When everyone can create, there's no point to call someone the artist..."
While I agree that workflows are being replaced, I think artists should be paranoid because the motivation behind changing workflows is efficiency and profit. Auto-gpt and agents are already showing signs that the higher level process of art directing can be replaced. However I think there actually is an abundance of motivation. People don’t get motivated just by positive things. It can be from fear too. Imagine this chaotic period being where AI just becomes useful enough to make you a fortune and not great enough to be AGI and making work obsolete, it is imperative to learn fast and act fast to have a chance for better tomorrow.
I agree with justin sooroojdeen - while this is efficiency and profit is definitely pro-corporation, there's enough reason to be optimistic and not paranoid - as Andrew says - as long as you're willing to learn, grow, and adapt with the technology. With the photography analogy, it certainly added efficiency and profit. Many people in simpler use cases could just take pictures themselves - perhaps as it should be. But then, people also realized how much skill a _good_ picture requires
Thx for the timestamps. Best and worst aligns with my own experience. And yes, studios holding out due to "old guard" and most common excuse is the rigging and animation systems. That said, working well for games IMO.
Great insights, as always! 💡
Thanks, Andrew! ❤
BTW are you using any AI in your workflows at Poliigon? 🤖
This Q&A made me contemplate to get twitter😂
I've noticed that the Poliigon addon doesn't really work plug and play if the Colormanagement is set to AgX. Thought of mentioning this issue.
Thanks Andrew
انا من ليبيا وليس لدينا مصممي اتردي هناك مادا سأفعل حتا اتفوق على الجميع ❤
Concentrate on drawing human hands - you have a few more years left before the AI catches up there!
AI is really bad at acrobatic poses or a human being upside down, it looks totally messed up surprisingly.
@@hombacom AI is a copy machine. Somewhat like a human, but with the brute force of a machine. Instead of just doing copies it tries to learn from them, and do better. But it's a machine, and when posed with a new problem like different poses, movement, (hands are almost a entire different systems from the body, they are difficult thing for humans too) or upside down pictures, AI literally has no idea what it is. It's trying to learn but it's a machine.
"Is that a human?"
"Nah, that's not a face"
"Could it be a face upside down"
"There is no such thing"
@@xanksauri89 i disagree about ”tries to learn from them and do better”. It just mimic shapes and patterns with randomness on top. Its Interesting that it can do impressive images but still doesnt have a clue what it is in reality, it doesnt visualize 3d in mind. A human maybe struggle with upside down and just simply rotate screen. We can judge result and start over, therefor we control ai tools and make it more Interesting.
My god the artifacts in background is hurting my eyes 😖
Render in blender
...so will we still make 3D models or not?
He did answer this, in a roundabout way, when talking about how Poliigon watches AI. Currently AI can be good with some prompts but not others, and not reliably. So AI may be a good starting point, artists will still need to know how to model to art direct or tweak whatever the AI prints out.
I'd say it's still worth learning the techniques even if you don't apply them regularly. It helps you be confident in making changes if needed.
The way to adapt to AI is to get out of digital art yesterday. Get as far away from that highly rendered CGI style as you can, it's worthless now as people will be (are) deensitized to it due to it being everywhere. Keep your art off the internet and if you have to put it on Glaze or Nightshade it. Sell yourself online instead your opinions your personality, draw/make art in videos dont post finished results as files that AI can take. Thats how you adapt, not by learning to prompt aka be AI's bitch.
mew sponsor ad sucks for premium
I saw a video about generating 100 ebooks to sell it on different plattforms. I think we will soon have a problem with too much of all. Saturation in all corners, even if you want support other creative human beings, it is hard to find out if there is one behind.. What we maybe need is, that ai art and assistance should be marked. I dont pay for things, made by 4 words, trained illegal by real artist input. Sorry, i am human. I like ai but not in every aspect of my life ( even we will be there in few years ).
I’m you
everyone can photograph but doesn't mean everyone is good or creative.. only a handful will be.. the same with AI art, it is just a different set of skills, expertises and perspectives. AI expands the creative and mental horizon for those who are of that temperament and thinking type but yeah perhaps not great for those who are more technical and have honed their skills for years to draw something to perfection..
this will be a story of betrayal
If AI is enough of a reason to get you to stop. Then you didn't really want it that badly. You we're just looking for a reason to quit.
I want to earn, I was naive and went for an art school I may start it within few months, I'm done for, yes art is lovely but now what, can't change my course my country don't work that way T_T
I would write a 4 page essay about how that statement is the most stupid and ignorant thing I have heard but I am too worried about my future rn