A B-School friend and I had an exchange regarding this piece recently by Economist Paul Ormerod. The opinion tackles themes like individuals acting as rational actors and in isolation, policy creation, and of course network effects. Of importance to me, as of late, has been automation, and the possibility of creating policy that is agile enough to adapt in real time to changes in the market. I allude to some of this. If you'd like to discuss further please reach out.
It's really a great read... below is my response to a portion of the piece.
"We discussed parts of this pretty extensively before...why create policies that assume rational actors.
One challenge that I come across when creating rules for clusters (in data, machine learning mechanisms) is that those clusters are very porous. So while one or many actors may be influenced by a few variables (time, shopping trend, weather) to act on option A today, tomorrow there could be exposure to another variable (shipping rates, news alerts, coup in MENA) leading for Option B. How do you create economic policy (or rules around data) with an amount of flexibility that will still yield some predictability? Also, option B leads to something else. Always.
This is where my heads been at lately. I think the notion of a collective/constant value in clusters of data or large groups of people is an important one to build on. If you can spot that one character trait you're at least building on reality and not what path you hope rational actor will take.
Machine learning with oversight gets close to a place that I think resembles agile policy. Can you imagine?"