worldoptimization:

some thoughts on leadership:

I used to think something like “it’s bad to be overconfident, and people who say things about how leaders should be confident are just promoting bad epistemic norms.”

now I think something more like–there are lots of situations where no one really knows what to do, and you need to make a decision. lots of people’s instinct in those situations is to ask the person in charge; but it turns out, the person in charge often doesn’t really know what to do either!

and leadership, in these cases, means not panicking or doing nothing or trying to find a person even more in charge, but instead coming up with your best guess–even if you’re super uncertain–and stating confidently and decisively “let’s do X.”

and if someone says “no actually, wouldn’t it be better to do Y instead because of Z” then great! none of this means you shouldn’t update on evidence or listen to other people’s opinions. just that there are a lot of situations that come down to “no one really knows, and someone needs to make a call and accept the responsibility if it turns out to be totally wrong” and being willing to do that is really valuable.

In coming to understand military leadership I realized “everyone pulling together on the same halfassed plan now” often yields better results than “argue about it for a while, then everyone pursues several better ones”