Using Control Theory to Train RNNs

I think the control theory, one is a good example, especially with RNNs. If it's a known dynamical system, for instance, right, then you can very easily just build that dynamical system rather than training some people to emerge. Yeah. No need to try and train the RNN from scratch to do like back props or something like it,. Just on a neural and like a deep learning set. Like, you want to back top through that 10,000 time steps like that's never going to work. But if you, you know, took a different approach to it and you're like, how could we get that signal back and you use something else? Like