This makes me think. In 2012 people still think that genetic programming, a numerical optimization technique (and not the most efficient one), magically solves random machine learning problems.
It would be interesting for me and others here to learn from what you know rather than be left with what you wrote.