> It seems like they found a way to solve nonlinear equations in constant time via an approximation, then turned that into a neural net.

You say that like it isn’t a big deal. Finding an analytical solution to optimising the parameters of a non-linear equation is remarkable.