I think there is a good question in here somewhere: Is it better for startups to monte carlo (and/or) genetic algorithm style variations on a theme to find the

I think there is a good question in here somewhere: Is it better for startups to monte carlo (and/or) genetic algorithm style variations on a theme to find the best way, or is it better try find something truly different to create and push the state of the art into a new area? And of course, that brings up the meta-question: did I just propose a false dilemma, are the two things really different?