I assume he's talking about an agent model from his description, in which case it is exactly trying to derive the macro economy from the micro decisions. This

I assume he's talking about an agent model from his description, in which case it is exactly trying to derive the macro economy from the micro decisions. This is a very classical approach. The idea that macro economy follows different models than the micro is somewhat modern. For example this caused a huge crisis in Keynesian economics in the 1970s where inflation models and the Phillips curve were drastically different than predictions and was a factor in the development of New Keynesian theories. The examples you give aren't about individual micro prediction not agreeing with the macro, but that's the case you can come across in various economic theories where is there is a literal disconnect the two and correct micro equations disagree with the macro equations.