This was a talk at Effective Altruism Global (London 2019) by Mahendra Prasad. It's a non-technical introduction to the ideas in his working paper I previously posted. It also covers some additional ground. For instance, from Republican polling data, we can see the difference voting methods can make with Trump in first place in plurality voting vs last place with approval voting. Towards the end, there's also a sketch of a research program applying algorithmic game theory to better understand and improve the background conditions of multiagent systems like deliberative democracy.