The moment in AI development where basic 3D models (cat, Ferrari Testarossa, torii gate etc) can be created with similar prompt-based process instead of 2D images will be the true milestone.
It can happen if people who made image-based AI work on it. Teach AI how geometry/meshes are put togther and all that, and give it lots of examples. Can't say it is easy or I understand how it works totally, but I'm pretty sure it will happen sooner than we know it.
Could you use this render add on to create 3d backgrounds in Blender, from different angles, and use whats rendered as background plates for animation?
I can't seem to do what you did with the prompt for the Collie. I've tried a few base models but nothing is generated except a few deformations on surface textures. Does a model have to have certain prerequisites to work?
I would need AI to help speed up my workflow or achieve things I can’t do myself. Feels like there is still an extremely long way to go, and we’ll run into copyright issues / vulture companies / scams / payments before we can use AI to actually model, rig and animate anything.
And this is why ai is a problem. People will be doing things that the ai is doing for them and then having the nerve to claim to have made it themselves. If you cant do something having ai do it for you is the same mindset of the people that will replace you with ai once it can do it all without you which outside of consumer access ai is already capable of. When you start having ai make your work for you i hope you dont go around claiming to be a creator.
So it takes your camera view as a latent image and generates img2img? Imagine Paying for credits just to do it inside Blender instead of using Stable Diffusion
I'm actually getting mixed information on this one - the sales page says you can, the FAQ on his GitHub says you can't - I'm assuming you actually can, but I'm not 100% sure how
Interesting development. But I wouldn’t call this “free”, with even an exclamation mark. It ‘ll still cost you money. “Free!” insinuates always and completely free. You’re better than using such a clickbait title, Justin.
Год назад+1
I tried to use this program directly and it ceases to be free immediatly after a few tries. It's not usable, you dont even have time to evaluate if it's worth it to put any money there. They are just riding the wave to get as much money as they can before people realize they are not there yet.
One day AI will give you exactly what you want, the tech will be expensive but with extensive trial and error you will be able to negotiation through to the perfect result, or just use a human instead:)
@@kodiakbluefox6643 I do make stuff but that doesn't even matter. Art is about creating and creating is bringing things from imagination to reality. Some people put more value in the process than others, that's just always how it's been. AI art and AI assisted art is just making that difference more noticeable
“Puts a filter over it” is not even remotely close to an accurate description of how this technology works. And as for creating something artistic, it’s really up to the artist to use the tools in a creative way which makes a legitimate artistic statement. Pressing a button and calling it a day may not be very artistic, but there are far more interesting ways to use this technology.
A.I will take away everything we have made. We will be rendered useless. Then guess what happens ? (hint : see what the guys at the WEF are thinking about useless human beings)
And once they release the ai that can do it all without you which already exists they just havent given it to consumers yet, artists will no longer be needed when only prompt writers will be needed to make anything. They will replace and entire post production house of workers with just a handful that know how to prompt ai properly. The worse part is that because its cool new tech artists are the ones "training" the system that will make them obsolete soon.
Please understand that supporting AI is supporting copyright theft on an industrial scale never seen before. All the AI data used to train software like this is stored in massive data sets like Laion 5B, and that data is stolen work scrapped from artists websites without their permission. Aside from that, we need to understand that any work made in this way doesn't belong to 'the prompter'. The US copyright office have clearly stated that any work made by AI is non copyrightable. From a humanistic point of view, were is the self respect as an artist in not creating the work?
@@ZexMaxwell Stable diffusion will always be pulling in other elements into its image creation for you from other sources. AI can not make images based on a few hundred personally feed images alone. Sadly it still doesn't tackle the fact that millions of people will be creating images on the back of stolen data as they have no images of their own to feed into stable diffusions data sets. AI simple can not operate without the Billions of stolen work of creatives and individuals. Lets hope Getty image's copyright law suit against Stable Diffusion goes through.
@@TobiJenkins no. Stable diffusion has the option to pull your own images that you made. Now if you are talking about training data so it could identify objects as objects. Yes that too is within fair use. The world has been training that since the invention of captcha. It also can make images from about 100 sample images. Corridor crew tested that with great success.
@@ZexMaxwell It's not fair use. It is literally why large companies like Getty images are suing Stable Diffusion. The turning point came when the images were taken under the guise of research (which would be fair use) but the second they went commercial with the product (which they did) it became unfair use. We are not talking about transformative use of art work which is deemed okay and copyrightable, the legal issue here is the scrapping of and holding of works and personal data under the false pretences of research. It is a holy immoral practice.
Hi everyone! Let me know what you think about this add-on in the comments below! :)
You didn’t mention that this can be run with a local installation of Stable Diffusion instead of paying for a cloud subscription.
You can definitely do that if you can get it up and running
We all actually need AI to do the retopology and unwrapping in one click with predicted result
This would be really nice
Right? And some one-click rigging would be nice. No more weight painting!
The moment in AI development where basic 3D models (cat, Ferrari Testarossa, torii gate etc) can be created with similar prompt-based process instead of 2D images will be the true milestone.
Totally, but that would be a big jump I think from this image based stuff that's happening right now - I'm excited to see what it looks like though :)
It can happen if people who made image-based AI work on it. Teach AI how geometry/meshes are put togther and all that, and give it lots of examples. Can't say it is easy or I understand how it works totally, but I'm pretty sure it will happen sooner than we know it.
It’s obvious that this will happen one day.
that exists already, look it up
@@voidnimbusthat exists in Unreal Engine
Thanks for sharing... I'm ready to use this tool and more...
Best intro is yours my friend.
at 2:52 you said you would just start by using your "bony file" i completely got lost there. i dont have any assets in my blender editor.
after you render the image with ai how do you apply the image to the 3d model?
Could you use this render add on to create 3d backgrounds in Blender, from different angles, and use whats rendered as background plates for animation?
I can't seem to do what you did with the prompt for the Collie. I've tried a few base models but nothing is generated except a few deformations on surface textures. Does a model have to have certain prerequisites to work?
Odd...i was able to use it before, but doesn't seem to work with my blender anymore. T.T
Thanks. Very helpful.
Oh man time to make an analog horror series using this
why they over used the word free and unlimited. after you used up all your free credits. then it bloats your hard drive.
Can wait for Blender Labs to release its own Blender ai renderer, not an add-on.
Thom from Blender said they are working on an ai renderer
Same, this will be a real game changer for us Blender users.
Is this, a year later, still the one?
00:00 HHHH WHAt's up guuuUUUYYYS
So it's not actually free?
Only for a few renders then you have to pay? Seems like a misleading title.
I would need AI to help speed up my workflow or achieve things I can’t do myself. Feels like there is still an extremely long way to go, and we’ll run into copyright issues / vulture companies / scams / payments before we can use AI to actually model, rig and animate anything.
And this is why ai is a problem. People will be doing things that the ai is doing for them and then having the nerve to claim to have made it themselves. If you cant do something having ai do it for you is the same mindset of the people that will replace you with ai once it can do it all without you which outside of consumer access ai is already capable of. When you start having ai make your work for you i hope you dont go around claiming to be a creator.
So it takes your camera view as a latent image and generates img2img? Imagine Paying for credits just to do it inside Blender instead of using Stable Diffusion
can't you just trace the image by your self ?
It would be cool if it actually made environments rather than images. You can simply use any AI image generator tool and import the image.
That's basically what this tool is doing - it renders it first, then uses AI to work with the image to create the final result
@The CG Essentials it's generating an image, not a playable map or film set, etc..
Which would be very impressive.
@@SynthToshi Oh I see - yeah totally would love to see that
Very nice! I wish I could use this with my own local Stable Diffusion thou.
I'm actually getting mixed information on this one - the sales page says you can, the FAQ on his GitHub says you can't - I'm assuming you actually can, but I'm not 100% sure how
it uses credits... sooo ur titel is missleading.. cheers
WoW! good!
:)
I'd like AI to generate hair and fur with manual control by the artist.
Interesting development. But I wouldn’t call this “free”, with even an exclamation mark. It ‘ll still cost you money. “Free!” insinuates always and completely free. You’re better than using such a clickbait title, Justin.
I tried to use this program directly and it ceases to be free immediatly after a few tries. It's not usable, you dont even have time to evaluate if it's worth it to put any money there.
They are just riding the wave to get as much money as they can before people realize they are not there yet.
nice video
:)
Spline 3d will kill blender?
I find that unlikely
One day AI will give you exactly what you want, the tech will be expensive but with extensive trial and error you will be able to negotiation through to the perfect result, or just use a human instead:)
Well that's the thing - the trial and error is extensive, and the reality is that a lot of the time just doing it yourself is easier :)
🤩🤩
:)
And how can you boast that you have created something artistic if the AI simply puts a filter over it? Something like that can't be legal.
Grow up
@@stayjacob Let me guess. You're not an artist, are you? Otherwise I can't explain your unfounded answer any other way...
@@kodiakbluefox6643 I do make stuff but that doesn't even matter. Art is about creating and creating is bringing things from imagination to reality. Some people put more value in the process than others, that's just always how it's been. AI art and AI assisted art is just making that difference more noticeable
Not a filter, your applying samples of other people's art work over yours without their permission. What can go wrong with that?
“Puts a filter over it” is not even remotely close to an accurate description of how this technology works. And as for creating something artistic, it’s really up to the artist to use the tools in a creative way which makes a legitimate artistic statement. Pressing a button and calling it a day may not be very artistic, but there are far more interesting ways to use this technology.
Pay credits? Interesting. Not interested.
Nah, gonna pass on this one, AI, at least for me, fucking sucks, it takes away the whole “artist” point.
I can see how you'd feel this way
A.I will take away everything we have made. We will be rendered useless.
Then guess what happens ?
(hint : see what the guys at the WEF are thinking about useless human beings)
And once they release the ai that can do it all without you which already exists they just havent given it to consumers yet, artists will no longer be needed when only prompt writers will be needed to make anything. They will replace and entire post production house of workers with just a handful that know how to prompt ai properly. The worse part is that because its cool new tech artists are the ones "training" the system that will make them obsolete soon.
Please understand that supporting AI is supporting copyright theft on an industrial scale never seen before. All the AI data used to train software like this is stored in massive data sets like Laion 5B, and that data is stolen work scrapped from artists websites without their permission. Aside from that, we need to understand that any work made in this way doesn't belong to 'the prompter'. The US copyright office have clearly stated that any work made by AI is non copyrightable. From a humanistic point of view, were is the self respect as an artist in not creating the work?
Using your own training data with stable diffusion will get around the law problem.
@@ZexMaxwell Stable diffusion will always be pulling in other elements into its image creation for you from other sources. AI can not make images based on a few hundred personally feed images alone. Sadly it still doesn't tackle the fact that millions of people will be creating images on the back of stolen data as they have no images of their own to feed into stable diffusions data sets. AI simple can not operate without the Billions of stolen work of creatives and individuals. Lets hope Getty image's copyright law suit against Stable Diffusion goes through.
@@TobiJenkins no. Stable diffusion has the option to pull your own images that you made. Now if you are talking about training data so it could identify objects as objects. Yes that too is within fair use. The world has been training that since the invention of captcha. It also can make images from about 100 sample images. Corridor crew tested that with great success.
@@ZexMaxwell It's not fair use. It is literally why large companies like Getty images are suing Stable Diffusion. The turning point came when the images were taken under the guise of research (which would be fair use) but the second they went commercial with the product (which they did) it became unfair use. We are not talking about transformative use of art work which is deemed okay and copyrightable, the legal issue here is the scrapping of and holding of works and personal data under the false pretences of research. It is a holy immoral practice.
@@TobiJenkins I'm not talking about the research or the previous data set. I'm taking about creating your own.