By now, there should be enough data to automate dang? On a more serious note: Is he really doing this all by himself? If so, I am sure he could train some junior people to off-load the work.
Proper moderation requires strength of character to moderate evenly people you do and do not agree with. Not something you can impart on others, even less so on juniors.
Being a waiter, working in a casino, being a cab driver, and running a coffee shop... you know, the secret to banning people is that you're always fair and more equitable than everyone else who's screaming, so when you do finally do it, everyone is either happy or scared they'll be next. People are surprisingly easy to manipulate this way. It would be wonderful if more sites activated that part of people's fear-lizard-brain more often; good manners are partly taught and partly enforced.
Yes you're right that it requires strength. And, at the same time, people have been training other people in every profession to do their jobs for ages. It's how we have survived and thrived as a species.
Ok. But where does one learn this strength, and subtlety?
This is an amazing idea.
For me it was as a waiter, and a taxi driver. Lots of people are assholes, but they're all people. You have to treat them with respect; but you have to be willing to ban them.
How do you teach such a thing?
Okay. Here's my great idea. Let's make a site where people can mod HN. Use an algorithm to coach them to be better moderators. Train them to learn patience.
How long do you think it would take to write that? How good do you think your trainees would be in the real world? Or would it just be better to have someone go live in the world for a long time?
If so... doesn't that mean it's a rarely acquired skill?
I was thinking that dang is going to be one of the most represented voices in GPT-4 and later. Hopefully you can retrieve him from the mix by nickname.