The models can't replicate the input that's given to them so there *really* isn't any real ownership issues in that regard (in spirit perhaps, but that's for the courts to decide (still)). At this time it's as capable of copyright infringement as a person is, but at the end of the day it'll be derivative none the less. As for the blender/3D modelling space, many larger companies have already incorporated the tech into their workflows and have tools for generating e.g. the uv maps for their models or creating variations for existing assets whether that's static assets like trees or dynamic characters. Afaik some game studios use it for map generation as well, but again, it's just part of the workflow rather than *the* workflow. That's kind of where the current iteration of AI is though. It's very powerful if you have the technical knowledge and supporting tech for applying it. E.g. you won't be creating prompt chains to automate office work if you don't know every step of the office work or generating a masterpiece drawing if you don't know what a masterpiece drawing is made up of. Which also makes it a bit scary because you can basically supercharge existing solutions to a very scary state of pseudo-autonomy without having to worry about the AI stopping to say "hey, I don't think we should kill all those kids" and instead it'll just process the request. As for image generation, if you're interested in further exploring that then you'll have the nicest workflow with automatic1111's webui, generating images with txt2img, amending details using img2img with inpainting and then upscaling to desired resolution. Lot's of different methods to get even fancier (controlnet, pix2pix), but that's a solid base to start from.
@@nicolebalian a1111 can be quite overwhelming in the beginning, but the community around it is very helpful and there is lot's of good tutorials for it. Another option is aqualxx's stable-ui which basically gives you access to an easier version of a1111 and as it uses stable horde you won't have to worry about not having good enough GPU. Btw, I think it could be worth a try to use chatgpt to turn the script of this "long" video into a plethora of shorts as I think the topic and the format would probably suit that demographic quite well.
It was joyful and inspiring to watch you generating images of endless possibilities.
😘😘
I am glad we had this lovely chat 😀😘😂
Thank you for joining😆
Love the content 🔥
Thank you Dan❤️❤️
very helpful video
The models can't replicate the input that's given to them so there *really* isn't any real ownership issues in that regard (in spirit perhaps, but that's for the courts to decide (still)). At this time it's as capable of copyright infringement as a person is, but at the end of the day it'll be derivative none the less.
As for the blender/3D modelling space, many larger companies have already incorporated the tech into their workflows and have tools for generating e.g. the uv maps for their models or creating variations for existing assets whether that's static assets like trees or dynamic characters. Afaik some game studios use it for map generation as well, but again, it's just part of the workflow rather than *the* workflow.
That's kind of where the current iteration of AI is though. It's very powerful if you have the technical knowledge and supporting tech for applying it. E.g. you won't be creating prompt chains to automate office work if you don't know every step of the office work or generating a masterpiece drawing if you don't know what a masterpiece drawing is made up of.
Which also makes it a bit scary because you can basically supercharge existing solutions to a very scary state of pseudo-autonomy without having to worry about the AI stopping to say "hey, I don't think we should kill all those kids" and instead it'll just process the request.
As for image generation, if you're interested in further exploring that then you'll have the nicest workflow with automatic1111's webui, generating images with txt2img, amending details using img2img with inpainting and then upscaling to desired resolution. Lot's of different methods to get even fancier (controlnet, pix2pix), but that's a solid base to start from.
Thank you for your insight and outlook 🙌 I’ll check out automatic1111!
@@nicolebalian a1111 can be quite overwhelming in the beginning, but the community around it is very helpful and there is lot's of good tutorials for it. Another option is aqualxx's stable-ui which basically gives you access to an easier version of a1111 and as it uses stable horde you won't have to worry about not having good enough GPU.
Btw, I think it could be worth a try to use chatgpt to turn the script of this "long" video into a plethora of shorts as I think the topic and the format would probably suit that demographic quite well.
i like aiart