I was just reading the first chapter of Software Design for Flexibility and a lot of the things the authors talked about reminded me so much of (what I understand) about Machine Learning (NNs in particular). With the important distinction that they don’t conform to the commonly cited advantage of FP that it is “easy to reason about.” Am I way off? I’ve heard of evolutionary software design but don’t know much about it. It seems the chapter is a call for “anti-fragility” in software design and the closest thing that I can think of is things like GANs etc that arrive at their function through an immense number of iterations.