excreationist
Married mouth-breather
There is a thing called "Deepfakes" which involves a face in a video being replaced very realistically with another using neural networks.
https://electronics.howstuffworks.com/future-tech/deepfake-videos-scary-good.htm
Already using the neural networks of our brains, we generate simulations of random scenarios. I think that in the future there could be neural networks that are trained with almost every aspect of reality that can generate simulations and fill in all the details.
In the same way that writers can generate descriptions of realistic conversations and behaviours, a highly trained neural network (or a network of them) could generate characters that we could interact with.
This shows a video zooming in on a scene from a neural network that keeps adding detail based on patterns it has learnt (animals, etc)
https://www.youtube.com/watch?v=dbQh1I_uvjo
The neural network I'm talking about would add detail in a way that mirrors normal reality. It would be able to handle it in 3D. It would be called The Dream MachineTM
Thoughts?
https://electronics.howstuffworks.com/future-tech/deepfake-videos-scary-good.htm
Already using the neural networks of our brains, we generate simulations of random scenarios. I think that in the future there could be neural networks that are trained with almost every aspect of reality that can generate simulations and fill in all the details.
In the same way that writers can generate descriptions of realistic conversations and behaviours, a highly trained neural network (or a network of them) could generate characters that we could interact with.
This shows a video zooming in on a scene from a neural network that keeps adding detail based on patterns it has learnt (animals, etc)
https://www.youtube.com/watch?v=dbQh1I_uvjo
The neural network I'm talking about would add detail in a way that mirrors normal reality. It would be able to handle it in 3D. It would be called The Dream MachineTM
Thoughts?
Last edited: