This is a post for the David Laibson Video. The thing that stood out to me as very interesting is his thoughts about hybridization of the simple Occam’s razor type models with the more complicated machine learning predictive models. Earlier in the interview, he mentioned that it is likely that sometimes these will be working in tandem and sometimes in parallel. This is echoed in the conversation about the combination of the rational-based economics theories being combined with behavioral theories of economics to create a model that we can actually use for predictions. It seems that the best way forward might always be some kind of confluence between these two methods, especially when dealing with something as complicated as economics. This also made me wonder about where the line is between something that is simple enough to rely just on machine learning and something that requires this merging. Early in the interview, Professor Goodman mentions that she is fine with google maps using only machine learning and not using the theory based Padua Rainbow, so that got me thinking about what the factors are that would make us want to incorporate more than just ML.