Happened to me with drawing. I want to use this to make trees / plants. Looks like you can use this for some models but not for fully customized renders with irl buildings or something built from architectural drawings.
short answer: no long answer: the way AI training works (fine tuning weights of a model based on millions of seen data samples) goes against the very goal of what you're trying to achieve in most games. Generative AI predicts tokens (text, image, vertices, etc.) based on what it considers to be the most likely option given the context. This means, AI most often goes for the "safe" choice, you can see this with text gen as well. But in 3d modelling, at least if you want to make good models, you need to be highly specific, especially when it comes to topology. The topology you want to have will depend on the style of your game, the technical requirements, polygon budgets, the structure of your rig, and many other things, most of which may not be known at the time you create your model. This means, your modeling is an iterative process, where you have to make many small changes depending on what your art director envisions. AI is not good at making EXACTLY what you want. It is good at making something somewhat resembling what you want. This may be a good base for someone like a concept artist when its about throwing ideas around and seeing what sticks, but it's absolutely terrible for making a good finished product. retopology algorithms and tools have been around for decades. Why do you think most studios still retopologize manually?
Silly to think it would replace modeling in all cases, but I don't think that is being suggested. Likely it will be used by many people who don't currently know how to model. What the outputs lack in refinement and optimization they make up for in speed of creation. I often notice how there are very few unique items in open world games. So many things need to be made to create the illusion of expanse that there isn't time or budget for novelty. It seems like these models would be great for in game possessions and decor. E.g. a store full of toys, or a workshop full of tools. A theme park ride with hundreds of objects not found elsewhere. That's the current appeal of generative 3D for me.
@@artemusprine i agree about the part of people who are new to the art. there is a very low barrier to entry. but when it comes to using this in a professional environment, there just too many issues. and the reason why assets are being reused doesnt actually have much to do with there not being enough time to model all those little assets. It has to do with memory: Every single object in the game needs materials and textures, and every texture that has to be loaded into memory slows down your game. If you reuse assets, you have to load a texture in only once but it's being shared by hundreds of objects which at least creates the illusion of a fuller world even though you used only a handful of assets. So even if the quality was there - which it currently isnt - populating scenes with thousands of unique assets just won't be a thing as long as our hardware limitations are where they're at currently (and keep in mind, games should be able to run on hardware that is 1-4 years old.)
@@thomasmann4536 Ideally you'd generate them on the fly procedurally and store the parameters... Then use things like elevators to buy you the time needed to generate them.
I don’t care if AI can create everything! I enjoy creating things myself. It is fun to see my 3D characters that I created come to life. I don't think AI can satisfy my artistic desires to draw, paint, and model. while AI can be helpful to artists, it is unfortunate that many people and companies might use it to replace artists.
Same but it's hard to focus on improving in 3D when there might not be many opportunities behind it. I'll continue because I enjoy it, but I know I can't go all in on this because it's not entirely in my hands.
THIS! IF people like creating things, people are going to make things. people don't stop making things because other people have made things. It's not how things work. people didn't stop drawing or painting because digital painting was faster and easier.
I was a 2D artist, I do concept arts and illustrations.. then I decided to be a 3D artist because I was afraid AI was going to replace illustrators and then here we are.... AI for 3D. WTF
Learn how to use AI for your needs. Imagine this, drawing geometric shapes by hand required talent and skill. Then softwares came and made it possible for any Tom,Dick,Harry to do the same. Now softwares have increased the possibilities for art due to the ease of doing previously hard tasks. AI too will replace some or many tasks but it can create new demands if we can learn to use it to make even better pieces of art. It's easier said than done, I know. But it's something we have to adapt to.
Ok lets be honest. When Photoshop introduced generative fill. It wiped out 90% of Fiverr editors who used to manually remove unwanted stuff from photos. Now they are back to cleaning toilets. Instead of a company hiring 10 designers. They might only hire one. They no longer need 10 of them.
@@RenderRides I've seen big companies. Like Old spice using generated images. All they have to do is slap their logo and call it a day. People need to stop coping. I wish it wasn't like this.
I would be ok with this if it was their in-house modelers doing the work, but I would like to know where the bulk of their models and textures come from. I wouldn't doubt alot of these models were scraped from somewhere where they were sold. Even if they were free, it harms artists as a collective to use these things. The artists did not consent to having their art collectively lumped into generative models. I recognize these tools will unlikely replace 3D modeling jobs soon because there are many limitations, but the point still stands.
I understand your concerns. At the end of the video, I mentioned how the industry will eventually establish stricter moral obligations for AI tools, as a collective sense of right and wrong use emerges, along with limitations on information access for AI models. I chose to make a video about this tool over others because I found some of their sources (ObjaverseXL, ShapeNet, and 3D-FRONT) and did some research into their background. This led me to believe their claims are legitimate. Transparency and ethical sourcing are critical, and I'm glad to see these aspects being addressed.
@@RenderRides We can only hope, this wont be the only company to do this in the future though, and others will likely spring up with more deceitful motives. My fault for not checking, great video.
@@RenderRidesit’s not concerns, it’s facts, and is happening now. Nothing wrong in acknowledging bad fortune. Eventually no one will need anyone’s else work in many fields.
not really since you can make your own 3d movies(or whatever project) in half the time. there will still be 3d modeling knowledge needed to keep the AIs on the right track of what you want
Conversely, when u think abt it, just imagine how much this is going to help 3d artists who focus more on animation or maybe like game development etc. It will be a life saver for all those who r not really that well versed with 3d modelling but need & require 3d assets for their work whatsoever
That’s a pipe dream, AI will just raise the bar for quality so it won’t make 3D artists life easier, it will just make everyone require even more work to get ahead of all AI crap There really wont be as much of an advancement as you may think
@@magma5267 now that I believe will matter from person to person, ones who'll be able to make use of it will see it as a boon, those who won't be able to they'll call it a curse🤷🏻
@@rogerotakufernandez5995 i agree, i'd like to use it as a tool aswell but even if AI gets good, i think proffesional artists will still be the best users of it, not beginner 3D Artists. as i said it won't level the playing field as much as people think, Nothing in life ever does its the law of balance in nature.
I recently tried out one of these "fancy" texts to a 3D AI modeler. Asked for a very simple task, and it was only capable of doing a shitty base mesh. So I ended up doing the modeling by myself as usual. Complete waste of time and money right now. You can have fun trying out these apps, but the only thing they are good for now is: tricking investors and stealing their money or making a video about "Will this AI repleace XY ?!" and gaining views and ad revenue due to AI hype. 🤑
That's honestly extremely concerning in a way. I feel like in a not so near future, companies won't even be needing people anymore, and even freelancer jobs will be hell because of AI.
I've only briefly tested it, so I can't make an objective comparison yet, sadly. Both Luma Genie and Rodin AI seem to have their unique strengths. Have you tried them?
Haven't really tested it with rigging, but I'm sure it will be doable for simple stuff. Problem with characters, even if the topology is technically quads, the edge flow matters a lot.
@@timatoma8076 prob a bit of both, tbh. but as with every other industry that AI is merrily undermining at the moment, i think when the dust settles, AI will do basic shit. elevator music, bad press releases, simple 3d models. stuff that requires actual real creativity will remain the domain of humans. of course, that's still gonna put a load of people out of work, and those people are going to get precisely no compensation-instead, assholes like Sam Altman are going to make a fortune. yay capitalism!
I tried rodin with the hopes that it was as good as they all said it was, unfortunately it isnt. Most of the 2s generations i liked were too much for the little AI to handle 😅
@@RenderRides I think it also really depends upon what you're trying to generate. I was trying to generate a Disney style character, but every 2d image that it generated which I liked it massively screwed up the 3d model generation and the texture generation, when I fed it my own image it was worse
My theory is that's because those things are not appealing to THE WIDE audience it doesn't get as much intention. The last bit of work that will be left for us 3d artists are gonna be manual retopo and uv layout... There is your dream job.. eventually cleaning ai art like you would clean a scan
AI will never replace 3D artists. AI will only make it easier for 3D artists, thereby company reducing salaries and making 3D modeling more accessible to developing countries, consequently reducing the overall price.
Yeah it does take trial and error to get things right. There's gonna be a new update where they allow you to input point clouds and voxels as guides to control the generation- would you want to see a video on that?
I dont like AI being used like this in the creative space, the fact that people have to work to bring there ideas to life is really what made me fall in love with 3D and CGI in the first place and this go's against that and it is somewhat demotivating as somebody trying to learn 3D modeling especially because this is probably the most passionate ive ever been in something Yes there are many good uses for AI like AI assistants, stuff outside of the creative field, hell maybe even retopology and uv unwarping but for creative parts like this it just isnt it and sholdnt be used
![ละลาย (LALALYE) - DAOU PITTAYA ต้าห์อู๋ พิทยา [OFFICIAL MV]](http://i.ytimg.com/vi/wo6r9TgausI/mqdefault.jpg)
Me : *starts learning 3D modelling*
AI : Excuse you
Guess I’ll go f*ck myself
Happened to me with drawing. I want to use this to make trees / plants. Looks like you can use this for some models but not for fully customized renders with irl buildings or something built from architectural drawings.
😂😂😂😂
It was much easier to create a tool for automatic topology than to replace absolutely everything
I couldnt agree more
I think AI's most practical usage would be to replace as many technical hurdles as possible while leaving the creativity up to us.
Still waiting patiently :(
@@RenderRides and yet, you still screw up 3d artists like me by posting this video
It's the same, the model is nothing but topology, flat faces meshed together forming an shape
short answer: no
long answer:
the way AI training works (fine tuning weights of a model based on millions of seen data samples) goes against the very goal of what you're trying to achieve in most games. Generative AI predicts tokens (text, image, vertices, etc.) based on what it considers to be the most likely option given the context. This means, AI most often goes for the "safe" choice, you can see this with text gen as well. But in 3d modelling, at least if you want to make good models, you need to be highly specific, especially when it comes to topology. The topology you want to have will depend on the style of your game, the technical requirements, polygon budgets, the structure of your rig, and many other things, most of which may not be known at the time you create your model. This means, your modeling is an iterative process, where you have to make many small changes depending on what your art director envisions. AI is not good at making EXACTLY what you want. It is good at making something somewhat resembling what you want. This may be a good base for someone like a concept artist when its about throwing ideas around and seeing what sticks, but it's absolutely terrible for making a good finished product.
retopology algorithms and tools have been around for decades. Why do you think most studios still retopologize manually?
Silly to think it would replace modeling in all cases, but I don't think that is being suggested. Likely it will be used by many people who don't currently know how to model. What the outputs lack in refinement and optimization they make up for in speed of creation. I often notice how there are very few unique items in open world games. So many things need to be made to create the illusion of expanse that there isn't time or budget for novelty. It seems like these models would be great for in game possessions and decor. E.g. a store full of toys, or a workshop full of tools. A theme park ride with hundreds of objects not found elsewhere. That's the current appeal of generative 3D for me.
@@artemusprine i agree about the part of people who are new to the art. there is a very low barrier to entry.
but when it comes to using this in a professional environment, there just too many issues. and the reason why assets are being reused doesnt actually have much to do with there not being enough time to model all those little assets. It has to do with memory: Every single object in the game needs materials and textures, and every texture that has to be loaded into memory slows down your game. If you reuse assets, you have to load a texture in only once but it's being shared by hundreds of objects which at least creates the illusion of a fuller world even though you used only a handful of assets. So even if the quality was there - which it currently isnt - populating scenes with thousands of unique assets just won't be a thing as long as our hardware limitations are where they're at currently (and keep in mind, games should be able to run on hardware that is 1-4 years old.)
you train a model to be printed in a determined resin printer.
@@thomasmann4536 Ideally you'd generate them on the fly procedurally and store the parameters...
Then use things like elevators to buy you the time needed to generate them.
@@thomasmann4536 With the Nanites from Unreal Engine object quantity e polycount is not a problem anymore
I don’t care if AI can create everything! I enjoy creating things myself. It is fun to see my 3D characters that I created come to life. I don't think AI can satisfy my artistic desires to draw, paint, and model. while AI can be helpful to artists, it is unfortunate that many people and companies might use it to replace artists.
Well said!
Same but it's hard to focus on improving in 3D when there might not be many opportunities behind it. I'll continue because I enjoy it, but I know I can't go all in on this because it's not entirely in my hands.
Can we do AI rigging instead
Sounds awesome, I hope someone creates something like that
Mixamo is pretty fast for rigging humans.
@@AlphaProtoaccurig by reallusion is way better. Free to download
Also mootion have brought out rigging on their 3dmodel ai's.
@@RenderRides does AI do decently at renders or UV mapping yet? I'm a decent sculptor, but absolute trash when it comes to UV and rendering 😂
THIS! IF people like creating things, people are going to make things. people don't stop making things because other people have made things. It's not how things work. people didn't stop drawing or painting because digital painting was faster and easier.
Honestly id use this just for creating the base mesh since thats the only thing I dont want to do on 3D modeling
That's a lot of money for something that does so little ngl
(You can find thousands of free base meshes online -_-
download it
I've been testing these kinds of ai models for at least a year now. They are starting to get a lot more useable finally.
I was a 2D artist, I do concept arts and illustrations.. then I decided to be a 3D artist because I was afraid AI was going to replace illustrators and then here we are.... AI for 3D.
WTF
*You can run but you can't hide*
@@RenderRides y u do that to me bro
broooooo...you just like me fr...i was also 2d artists and chanjed to 3d bcs of AI@@von.3d
Learn how to use AI for your needs. Imagine this, drawing geometric shapes by hand required talent and skill. Then softwares came and made it possible for any Tom,Dick,Harry to do the same. Now softwares have increased the possibilities for art due to the ease of doing previously hard tasks. AI too will replace some or many tasks but it can create new demands if we can learn to use it to make even better pieces of art.
It's easier said than done, I know. But it's something we have to adapt to.
@@AS-ym2bp i currently am using AI for other uses. You're right. We have to adapt. It's just scary to think few years from now how much AI will grow.
Ok lets be honest. When Photoshop introduced generative fill. It wiped out 90% of Fiverr editors who used to manually remove unwanted stuff from photos. Now they are back to cleaning toilets. Instead of a company hiring 10 designers. They might only hire one. They no longer need 10 of them.
The employability bar is going to be higher for sure
@@RenderRides I've seen big companies. Like Old spice using generated images. All they have to do is slap their logo and call it a day. People need to stop coping. I wish it wasn't like this.
Pair this up with Mixamo and you can have ready to animate characters from drawings.
That'd be such a crazy workflow for individuals with a little more refinement
Using AI to make scanning IRL objects better sounds awesome! Or perhaps it could do it from several images
Cool idea!
I would be ok with this if it was their in-house modelers doing the work, but I would like to know where the bulk of their models and textures come from. I wouldn't doubt alot of these models were scraped from somewhere where they were sold. Even if they were free, it harms artists as a collective to use these things. The artists did not consent to having their art collectively lumped into generative models. I recognize these tools will unlikely replace 3D modeling jobs soon because there are many limitations, but the point still stands.
I understand your concerns. At the end of the video, I mentioned how the industry will eventually establish stricter moral obligations for AI tools, as a collective sense of right and wrong use emerges, along with limitations on information access for AI models. I chose to make a video about this tool over others because I found some of their sources (ObjaverseXL, ShapeNet, and 3D-FRONT) and did some research into their background. This led me to believe their claims are legitimate. Transparency and ethical sourcing are critical, and I'm glad to see these aspects being addressed.
@@RenderRides We can only hope, this wont be the only company to do this in the future though, and others will likely spring up with more deceitful motives. My fault for not checking, great video.
One problem I see is that most human artist learn from looking at other peoples work.
@@RenderRidesit’s not concerns, it’s facts, and is happening now. Nothing wrong in acknowledging bad fortune. Eventually no one will need anyone’s else work in many fields.
scary and awesome at the same time!
yeah, it really has the potential to be a game-changer for individual artists in the future
Calculator did not replace Mathematicians.
It replaced many of them
It lessened them by a lot
@@cooleregg5577 can calculator do things that mathematicians do
@@REALdragon_fifteen soon yes
@@cooleregg5577 ik, computer still do , but they are there to assist humans
That would save a lot of time. But before I begin with AI, I want to Master 3D modeling.
idk about the future of 3d modelling, but sure the assett markets will be going spammed by this
already are. saw a guy who uploaded 32000 models on sketchfab.
Thank you for sharing! Wow , absolutely awesome AI . I just tried it and its very usable for low poly games (dont get too close though)
Exactly! And I'm sure it will only get better with further research.
There goes my future 😅
Don't worry, we've always got plan B (rickshaw driver)
@@RenderRides 😆
@@RenderRides 😂😂😂 Right 👍
not really since you can make your own 3d movies(or whatever project) in half the time. there will still be 3d modeling knowledge needed to keep the AIs on the right track of what you want
Conversely, when u think abt it, just imagine how much this is going to help 3d artists who focus more on animation or maybe like game development etc. It will be a life saver for all those who r not really that well versed with 3d modelling but need & require 3d assets for their work whatsoever
That's a really good example!
That’s a pipe dream, AI will just raise the bar for quality so it won’t make 3D artists life easier, it will just make everyone require even more work to get ahead of all AI crap
There really wont be as much of an advancement as you may think
@@magma5267 now that I believe will matter from person to person, ones who'll be able to make use of it will see it as a boon, those who won't be able to they'll call it a curse🤷🏻
@@rogerotakufernandez5995 i agree, i'd like to use it as a tool aswell but even if AI gets good, i think proffesional artists will still be the best users of it, not beginner 3D Artists. as i said it won't level the playing field as much as people think, Nothing in life ever does its the law of balance in nature.
I knew 2 sites already this is new for me
stop it. get some help
What help?
Now the question is if there's an ai that can turn 3d models into 2d art
Yes
All I care for is AI getting rid of retopo and UV-maps. Thats the only AI tool Id pay for.
I'll be on the lookout for them!
Ngl, that would be extremely helpful
I recently tried out one of these "fancy" texts to a 3D AI modeler. Asked for a very simple task, and it was only capable of doing a shitty base mesh. So I ended up doing the modeling by myself as usual. Complete waste of time and money right now. You can have fun trying out these apps, but the only thing they are good for now is: tricking investors and stealing their money or making a video about "Will this AI repleace XY ?!" and gaining views and ad revenue due to AI hype. 🤑
@@superhero.monster Why are you upset over a question I asked in the video title?
That's honestly extremely concerning in a way. I feel like in a not so near future, companies won't even be needing people anymore, and even freelancer jobs will be hell because of AI.
What are your thoughts on Luma Genie?
I've only briefly tested it, so I can't make an objective comparison yet, sadly. Both Luma Genie and Rodin AI seem to have their unique strengths. Have you tried them?
@@RenderRides yeah and it's quite good for props but not for characters
AI took muh job!!!!!
How does it perform when rigging?
Haven't really tested it with rigging, but I'm sure it will be doable for simple stuff. Problem with characters, even if the topology is technically quads, the edge flow matters a lot.
"Not YET" give it 3-5 years.
no
next question
@@evilotis01 that's the kind of confidence we need more in the community 💪
Is it confidence or denial?
@@timatoma8076 prob a bit of both, tbh. but as with every other industry that AI is merrily undermining at the moment, i think when the dust settles, AI will do basic shit. elevator music, bad press releases, simple 3d models. stuff that requires actual real creativity will remain the domain of humans. of course, that's still gonna put a load of people out of work, and those people are going to get precisely no compensation-instead, assholes like Sam Altman are going to make a fortune. yay capitalism!
You Explained it smooth as butter
Haha glad you liked it man!
a robot for 3d painting
FELLOW ARTISTS, RUN YOUR RENDERS THROUGH NIGHTSHADE BEFORE POSTING YOUR WORK!
👍👍👍👍
I tried rodin with the hopes that it was as good as they all said it was, unfortunately it isnt. Most of the 2s generations i liked were too much for the little AI to handle 😅
Definitely takes some trial and error, and I've seen that changing the auto generated prompt and advanced settings really helps.
@@RenderRides I think it also really depends upon what you're trying to generate. I was trying to generate a Disney style character, but every 2d image that it generated which I liked it massively screwed up the 3d model generation and the texture generation, when I fed it my own image it was worse
Why do people use ai for the wrong things. You could use it for topology or weight painting rather than completely replacing modeling.
I'm sure there's going to be niche AIs for that stuff soon
My theory is that's because those things are not appealing to THE WIDE audience it doesn't get as much intention. The last bit of work that will be left for us 3d artists are gonna be manual retopo and uv layout... There is your dream job.. eventually cleaning ai art like you would clean a scan
You don't even need AI for topology or weight painting, there have been tools for both things for years
Sure :))) . And my name is einstein . Ai will never be the same as humans 3d arts . Peace .
Would love to hear more of your thoughts on why that is
I might use it. But I don’t need it.
As with most new tools :)
AI will never replace 3D artists. AI will only make it easier for 3D artists, thereby company reducing salaries and making 3D modeling more accessible to developing countries, consequently reducing the overall price.
Good bye my dream job.
Nah, we still need artists :)
ohhhhhh, im fucked
I know I've started with my prayers 🙏
Nah it doesn't generate detailed things, I still have time to learn how to modeling, AI is too weak now.
You should mention that this is sponsored.
It was mentioned near the beginning of the video around 0:16 and also tagged as 'containing paid promotion' while uploading.
I hate this with everything I have.
Finally this mf uploaded smth
😭 i own up to it, but check out my last upload (a lot of people didn't get notified)
mai rou ya hasu....... karu mai kya karu?
has do bhai, khud pe (been there 😔🙏)
I subscribed. I dont eant to fuse with Joe Biden.
oh
bro really just commented 'oh' 😭
@@RenderRidesbetter than those jay sree ram comments 😂
yeah i'm not sure how that was relevant to the video either 😭😭
I guess I am getting fired soon 😢
Learn how to use these tool better then anyone. This is the beginning you can do it
NO
Fair enough
Tbh it kinda sucks with making simple things so I don't think it will
Yeah it does take trial and error to get things right. There's gonna be a new update where they allow you to input point clouds and voxels as guides to control the generation- would you want to see a video on that?
@@RenderRidessure
Jai shree Ram
🙏
Butlerian Jihad.
Fuck AI
Its replacing everything
Use it as a tool and a tool only. not the artist.
I dont like AI being used like this in the creative space, the fact that people have to work to bring there ideas to life is really what made me fall in love with 3D and CGI in the first place and this go's against that and it is somewhat demotivating as somebody trying to learn 3D modeling especially because this is probably the most passionate ive ever been in something Yes there are many good uses for AI like AI assistants, stuff outside of the creative field, hell maybe even retopology and uv unwarping but for creative parts like this it just isnt it and sholdnt be used
i hate my life
Skill issues.