Remember to make inferential distance work for you: if you try and cover more than about 5 blogpostsworth of material at once, it'll end poorly because it'll be too much for your audience to take in - that's that bad side of inferential distance. But there's a good side too: if you just do a few things with very high quality, even if it doesn't seem like that much to you, if can be really impressive to your audience because most of it will be new to them.
Also, if you ever do a talk on human biases (as opposed to philosophy / AI design), I recommend that you make sure that your talk teaches the audience the preliminaries of talking about biases, while still seeming to just be about biases. Otherwise...
Just recently at my high school, a group of classmates and I started a science club. A major component of this is listening and giving peer lectures on topics of physics, math, computer science, etc. I picked a topic a bit off to the side: philosophy and decision making. Naturally, this includes rationality. My plan is to start with something based off the sequences, specifically "How to Actually Change Your Mind" and "A Human's Guide to Words".
I was hoping the Less Wrong community could give me some suggestions, tips, or even alternative ways to approach this. There is no end goal, we just want to learn more and think better. All our members are among the top 5% academically of their own grade. Most of us are seniors and have finished high school math, taking AP Calculus this year. We have covered basic statistics and Bayes' Theorem, but only applied it to the Disease Problem.
Any help or ideas are appreciated.
Update: Thank you for all these suggestions! They are incredibly helpful for me. I will attempt to make a recording of the lecture period if possible. I will make another discussion post sometime next weekend (the first lecture is next Friday) to report how it went.
Update 2: Report here.