(Sequential) Information Design

on YouTube ↗01h 02m

" summary

Surveys the information‑design toolkit and gives concrete characterizations: distributions over actions and states must satisfy a martingale (prior) constraint and an obedience condition, so feasible outcomes are solutions to a linear program; applies this to auctions (how bad revenue can be under worst information for a first‑price auction) and to a seller–buyer pricing example. Shows a sequential recommendation construction (flip coin who moves first) that implements full cooperation in Prisoner’s Dilemma under a simple payoff inequality, illustrating the extra power of sequential contact.

AI-written summary · about

" chapters (15)

1. 0:00Introduction to primitives
2. 2:50Mechanism design vs. info
3. 5:27Information design approach
4. 7:43Robustness and equilibrium
5. 10:14Formalizing information
6. 13:32Seller and buyer examples
7. 17:47Single agent motivation
8. 20:42Characterizing distribution
9. 26:05Obedience and representation
10. 31:21Linear program applications
11. 34:40Bayesian persuasion details
12. 37:48Beliefs as signals
13. 43:41Correlated equilibrium
14. 50:30Robustness in auctions
15. 56:48Sequential interactions