so this means that neural nets are basically over right? Most research is going to focus on symbolic regression/genetic progamming instead?

pic related. MLP gets btfo by the genetic programming, gradient boosting and symbolic regression methods.