I have encountered and lived in several communities that have the following two properties:

1. They produce a lot of weird, novel ideas.
2. A reasonable set of the people in them are extremely overconfident.

I think this is no coincidence, and I think having norms of “epistemic modesty” likely breeds out the production of novel ideas. Here’s why, basically:

1. Sharing new ideas is costly, because you might be wrong. Perhaps in some ideal world, people could segregate their ideas into “super new ideas that I’m really unsure about” and “really actually good ideas that you should judge me for”, but I think this doesn’t happen in practice and so people are correctly apprehensive about bringing new things up.
2. One way to combat this costliness is to have the irrational belief that you’re right, even if you’re not.
3. If you have strong norms towards epistemic modesty, people are more likely to think they’re wrong, and so they share less. Even worse, I think epistemic modesty norms probably push people to stop even generating new ideas, much less sharing them. See Socratic Grilling.

I legitimately think I encounter this effect on a personal level– when I’m feeling really good about myself, I feel like a fountain of original thought. When my self-esteem is lower, I have to talk myself into taking my own ideas seriously with high-level policy arguments about how it’s good to think new things even when they’re wrong.

Let’s be clear here: people who are overconfident are wrong, and they lose money in bets. It is bad to base your decisions on things that are wrong. But I think producing truly novel ideas is the first step to producing really good ideas, and there’s an incredibly important niche of communities and individuals that do this and are overconfident. I don’t think there’s an easy alternative where everyone generates the same number of new ideas and is correctly confident all the time. I have historically been very judgmental of overconfident communities and I now think this was very lame of me.