Hrm. I find myself wanting to disagree with this comment while agreeing with your original post. I think there's three distinct levels worth thinking about.

There's worthiness as in "self-worth" or "worth helping." The world is probably better if the bulk majority of people have this, and I have heard people express the idea that every human is worthwhile in this sense. That's not to say me or you can't prioritize who we care about, but "such and such people aren't worth the air they breath" is a dangerous line of reasoning. Complements aren't particularly useful here, as "You can use language, therefore your existence is positive" can frankly come off as a backhanded insult of sorts, since that's a really low bar.

There are correct steps in the right directions well*, including most personal growth and including hill climbing towards better states. This is the place I think complements are best deployed; an adult human taking a ten minute walk outside is better than that human sitting on the couch watching reruns. (I recognize I'm making a value claim there that may not be globally correct.) Guessing "red" all the time in the probability question above is better than guessing "red" 60% of the time and guessing "blue" 40% of the time. Progress is worth appreciating, both on the personal level ("hey, congrats on beating your mile-run time!") and on the group level (I watched a time-lapse map of malaria cases in a room full of EAs recently, and I am kinda disappointed that nobody cheered.) It doesn't even have to be a new achievement! In martial arts, I eventually reached the point where every other session the instructor would nod and say "good stance" before moving on. This never stopped feeling good to hear, and it kept the basics in my mind even as I moved on to more advanced steps.

Then there's being correct on an absolute scale. The kind of rightness that involves local validity and correct premises, the kind that gets measured against the real world and succeeds. A successful rocket launch, a healthy patient after surgery, an AI that does what we meant and not what we said. The universe does not grade on a curve and gives no awards for effort. I think if we as aspiring rationalists lose sight of this, then we will eventually go astray no matter how good we are at the first two*. Complements here are rare, but powerful.

My suggested heuristic for the community would be to complement someone when you know them and see them advance along the path, or when they do something which helps you advance*. I also offer complements when someone does something I want them and/or others to do more of even if it is not novel, and I suspect that this kind of complement is what you are seeking to encourage; if so, then we are in agreement. "Good stance" is important to hear, as is "good job updating" and even "hey, good job organizing the meet up yesterday! I think you pretty good moderating, you jumped in at the right moment when me and Bob were getting derailed." Praise for getting things right, with the promise of more encouragement as we climb higher.

*To be clear, I don't think there's a single linear ladder we climb straight up from ignorance to superrationality. There are probably multiple paths to the summit, and there may well be more than one peak. That's a different topic however.

Give praise

by toonalfrink 1 min read29th Apr 201887 comments


The dominant model about status in LW seems to be one of relative influence. Necessarily, it's zero-sum. So we throw up our hands and accept that half the community is just going to run a deficit.

Here's a different take: status in the sense of worth. Here's a set of things we like, or here's a set of problems for you to solve, and if you do, you will pass the bar and we will grant you personhood and take you seriously and allow you onto the ark when the world comes crumbling. Worth is positive-sum.

I think both models are useful, but only one of these models underlies the emotional need we call status. I think it's the latter.

Another assumption: humans are satisficers. Those that claim to the contrary have never been satisfied. An unsatisfied satisficer acts like a maximizer. I think that Maslov got pretty close with his hierarchy of needs. Not the specific set of needs, not necessarily their order, but the idea of humans working on one need at the time, until satisfaction, so that the next need comes up.

It seems to me that many of us are stuck at the status level, and I think getting past it makes us surely happier and plausibly more effective.

How is worth generated? Quite simply, by giving praise. You either find your behavior exceeding a standard that the community agreed on, or someone actually tells you you're doing well. The latter seems more powerful.

I've asked around, and it even seems to be considered "normal" in most (non-systematizing) communities to habitually give praise. It's even apparently something people regard as necessary for proper psychological health. But honestly, apart from volunteering at CFAR, I can't recall getting much praise for anything I've done for the community. As a result I never quite feel like I'm doing enough, edging on burnout a few times. Reminds me of pica. Does working on AI Safety ever get me the sense of worth I'm looking for, or should I give up?

So I'd like to suggest we try for Giving Praise as a rationalist virtue. It might just be a staple of group rationality.