Frameworks, not prophets
Read across, not up
A generation of readers found their views on AI and work from a single popular voice. That is risky. This course deliberately drew on Keynes, Autor, Piketty, Frey-Osborne, Brynjolfsson, Acemoglu, Bresnahan-Trajtenberg, Fisher, David, Diamandis, Van Parijs, Darity-Hamilton, Shapiro, the ILO, the OECD, the IMF, and the NBER. The practice to carry forward is: if one voice is doing most of the work in your head, add three more. Find where they disagree.
Steelman, then decide
For any claim you are tempted to believe, write down the strongest version of the opposing view. Then decide. This is the habit most separates good analysts from bad ones. Applied to this course: the strongest objection to everything you have just read is that past automation waves also looked catastrophic at the time and did not become so. That objection is real and serious. Any conclusion you reach has to survive it.
Think in time horizons
A good analyst keeps three horizons distinct. Next five years: what the evidence now shows and plausibly extends. Ten to twenty years: scenarios with real uncertainty. Fifty years and beyond: imaginative work where nobody knows. Mixing them produces confident bad predictions. Separating them produces honest ones. Most post-labour claims live in the middle horizon; they deserve to be held loosely.
A credit, and a handoff
David Shapiro's work has done a lot to bring post-labour economics to a wider audience and shape its popular framing. This course drew on his writing alongside many others. The point of a course like this is not to replace his voice or anyone else's — it is to give you enough footing to read him and them, disagree where you should, agree where you should, and then build your own position. That is the actual point of education. Good luck.
Name the three writers or researchers on automation whose views you trust most. Where do they disagree with each other? If you can't name at least one disagreement, you have a prophet problem.