Great essay Mel. So what does a Ph.D. in Agent-based Modeling entail these days? Back in the early 80s when colleagues and I at Mich were calling for use of complexity theory to provide an artificial reality check for economists, we did not think they would merely replace Homo-economicus with Machina-economicus. Mainstream economists merely uses agent-based models to co-opt it and confirm their dogma. Truly diverse (e.g., recipricans, altruists, etc.) demands what I call Homo-complexicus, and generalist methods, as you suggest. Moreover, I contend that it is not the individuals that should remain the unit of analysis, rather it should be the their interactions in co-evolution with their institutions.

This might also interest you: https://medium.com/@daneke/artificial-insanity-update-machine-learning-and-democracy-burning-e98d4e672e7a