Reading LessWrong but not winning

I've associated myself with LessWrong, and the rationalist community, for several years. It was one of the two communities I felt like I vaguely belonged to.

I saw a tweet recently:

I think LW overly weights their epistemology in their successes & under-accounts for their single-minded relentlessness.

The epistemology is above average, but not as good as they think it is. Relentless meta-game is excellent, tho, even if might not be good enough to eat itself

— @meditationstuff, 2020-12-26

Here is how I feel about it:

LW, as a community and a body of work, is popular—and this is a success. LW is good at entertainingly explaining philosophy and psychology in a way that works for computer nerds, who in their turn currently happen to have lots of money and cultural influence.

LW, as Eliezer's project to get people to think and care about AI risk, is also a success.

However, in terms of bringing success to its members, LW is probably a big failure. Reading LW but not winning should be approximately as unsurprising as reading Hacker News but not running a startup.

The biggest reason for that is that in the outer LW community, knowing that you should shut up and multiply is valued as highly as being able to actually shut up and multiply. And LW-the-community provides very few incentives to train that (genuinely useful for winning) skill.

So, instead people ask each other

— kill one or five?
— kill one
— ONE OF US ONE OF US

and that's about it.

In a situation where beliefs don't have to lead to actions, anybody can believe anything at all, and biting bullets becomes very easy. Kill one person to save five, give the vaccine according to utilitarian principles, explode the sun.

However, in the practice of shutting up and multiplying, the hard part isn't multiplying but shutting up. LW offers no guidance on how to shut up, beyond trying to philosophically convince you that you should shut up.

From More Dakka by Zvi:

Image

"Do things that work" is a great skill that can change one's life and set them on a path to greatness. LW is bad at training that skill. Instead LW instills the belief that "everyone should do things that work", which might be good as a first step but that's it.

And, well, this might be as good as it gets—you can't effectively train people by putting words in front of them.

Story time: I organized a LW meetup in Minsk, in the COVID times. I bought extra masks, I told everyone to wear masks, I made sure everyone did wear masks. Me and another guy were going to give talks. The guy said "I have a quiet voice so I'll take the mask off". Everyone but me was like "yeah sure fine". (In the end I just told him that he won't give his talk.)

Being shit at shutting up and multiplying, or at shutting up in general, isn't treated as a failure in the community. There's no social pressure to actually win, as long as you agree that "yes, winning is good".

And there's a good reason for this. Everyone is compassionate and gets that changing yourself is hard, and so when somebody makes a post like "I can't force myself do the most basic things, like getting up early", people just go "here are some tips that helped me". Which is the correct thing to say, if the only thing you can say are words.

And rationalists think that words/beliefs are the only things that exist, with some communities believing false things and some communities believing true things.

Which is nonsense! Proper communities have people who exemplify community values, and it works because people learn by imitation. Who is the shiny beacon of light in the LW community, when it comes to winning and not having the right beliefs? I dunno.

Oh. It's Harry Potter.

Oh boy, this does not bode well.

Addendum: the Dragon Army case study

Dragon Army was a rationalist sharehouse with supposedly draconian rules. See its charter:

What is Dragon Army [Barracks]? It’s a high­-commitment, high-­standards, high­investment group house model with centralized leadership and an up­-or-­out participation norm, designed to a) improve its members and b) actually accomplish medium-­to-­large scale tasks requiring long­term coordination. Tongue-­in­-cheek referred to as the “fascist/authoritarian take on rationalist housing,” which has no doubt contributed to my being vulnerable to strawmanning but was nevertheless the correct joke to be making, lest people underestimate rather than overestimate the constraints they were signing up for.

Everybody was like "oh no it might be a cult". Except that if you think about it, this is exactly what a workplace is. High-commitment? Check. High-standards? Check. Improve its members? Check. Longterm coordination? Check. Giving people orders? Check.

This is not a cult. This is commonplace. Most adult people are engaged in this kind of thing, because they have jobs.

Correspondingly, when I evaluate the retrospective from the point of view of a manager, I see someone who isn't a great manager:

While setting personal boundaries is a skill I could improve in its own right, having a “nobody in Dragon Army is responsible for anything Dragon Army” day each week would’ve been net positive for everyone.

[...] As a simple but representative example, we didn’t get formal weekly check-ins off the ground until several months in.

[...] I didn’t give enough orders. I didn’t provide as much structure as half the house showed up hoping for, in part to avoid transgressing with the other half of the house.

[...] Actually check in on the ~20h/wk requirement. One Dragon, late in the game, asked outright “So, do we actually have standards, or not?” and the answer was “No, and it’s my fault.”

This sounds fancy, so I will translate:

  • Didn't set up weekends (!)
  • Didn't do 1:1s
  • Didn't give orders because was afraid to give orders
  • Didn't even ensure people worked for 20h/week

I'd say that this is "below average" for a workplace, except that a lot of workplaces are shit, so it's above average but still not great. I am a self-taught manager with very little experience. I expect I will be at Duncan's level in half a year. But Duncan was the Curriculum Director at CFAR at the time!

"Knows the importance of standards and is willing to bite significant bullets about them, but can't set a standard and hold people accountable to that standard" is a version of "knows to shut up and multiply but can't shut up".

So I'd say it's a drastic mismatch between "has an understanding of the principles behind things" (high) and "has the ability to follow those principles" (not very high). Which is my whole point!