from Hacker News

after witnessing the rise of deep learning as automatic feature/ pattern recognition over classic machine learning techniques, an intuition arose that the more you automate at each level you get better result therefore i immediately focused on neuro-evolution.

I spent some weeks reading on neuro-evolution publication with the same desire to automate at every level then some fairly good logic arose 'could Genetics be good from the onset of evolution' Didn't they get better at searching through out the solution possibilities over each trial(generation)/time.