3 ms·
>They said In 20 years, suggesting it's not possible today. It's not exactly far away today. The building blocks are there. They're weak, but they are all fund
by barking_biscuit 4y ago
>They said In 20 years, suggesting it's not possible today.
It's not exactly far away today. The building blocks are there. They're weak, but they are all fundamentally there.
My wife loves the show SWAT and I used Stable Diffusion to insert her into some still shots of the show in place of Chris and she lost her mind. I'm seeing people getting some pretty good results with video-to-video techniques using ControlNet, so my next experiment is going to be to see if I can do it with video, and use Eleven Labs to clone the voice and replace that too.
I'm virtually certain personalized entertainment of this nature is coming, and it's going to be huge.