I have signed up to give a LW mini-talk. I could in theory talk about various topics in Bayesian statistics and MCMC, but what I am actually interested in discussing right now are the ideas in this post: https://www.lesswrong.com/posts/Rh8v4ba9w5KTb6onD/ongoing-thoughts-on-proportional-voting-methods
If you have read that post, what topic there do you think would be best? It should be something that’s not common knowledge or “obvious” to the audience, but also doesn’t require too much setup. Assume that I am able to do a decent job of structuring my talk — that is, a substantially better job of setting up and exploring a single idea in a 5-minute chunk than I’ve done in that essay which is not intended to be a 5-minute chunk.
Feel free to post “answers” here which are general advice and not specific topic suggestions.
As a very quick answer, I’d basically be interested to hear about any part of it that you’re thinking a lot about, and for you to explain at least two different perspectives on the question.
It also helps if we can get the basic picture in 5 mins so that in the Q&A and convo after we can provide thoughts that would be actually useful to you, where even if we’re wrong you can explain why to us. Basically this idea is for you to optimise for you getting to explain things to us in a way that you find most useful.
I thik I’m gonna do something about why high-dimensional is hard. I’ll mention the voting context, but mostly discuss the problem in abstract.