I’m Ben! I’m a junior at Harvard studying math and computer science. When I’m not doing that, I enjoy making things, coding, reading, playing piano, singing, contra dancing, hiking, effective altruism, and other strange activities.


To stressed-out altruists


in saving the world, ethics, effective altruism

Lately several people in the effective altruism community have mentioned that thinking about EA all the time can be pretty stressful. If effective altruism is making you anxious, I think you’re probably being far too hard on yourself, and you should try to compartmentalize more. It may not seem optimal to deliberately go easy on yourself, and it might not be if you’re a perfect utilitarian robot, but you’re a human mind running on an imperfect brain and it’s more important not to break anything. If you’re stressed, here are some considerations for holding yourself to a somewhat lower standard.

First of all, “effective” is not binary-valued. It seems like some people are berating themselves for not meeting what I like to call the “perfect utilitarian robot agent” standard, where you try and behave like a completely rational actor whose sole concern in life is maximizing aggregate utility. Typically the imaginary person of comparison doesn’t have to worry about pesky mental things like social contact or emotions, either (to name a few). Nobody actually meets this standard, for hopefully obvious reasons. So in that sense, nobody is a “perfect” effective altruist, and it doesn’t make sense to fault oneself.

Suppose, for instance, you spend some time hanging out with a romantic interest instead of taking those hours and squeezing out the last drop of donated income. Sure, maybe it’s satisficing and not optimizing, and it you read a lot of Less Wrong (or a lot of certain EA stuff) you might get the impression that that’s some kind of cardinal sin. But not all instances of satisficing are created equal. The amount by which you are copping out by giving up those few hours is epsilonically small compared to—well, come to think of it, compared to almost everything ever. Like all the people who donate to Make-a-Wish instead of AMF, or “hand-illustrated personalized stories for orphans” (no joke, this is a Harvard student group) instead of CEA.

So the amount you berate yourself should be properly calibrated to the size of the difference, i.e., it should also be epsilonically small. If you fault yourself too much for the little things, you’ll run out of room on the degrees-of-fault scale to distinguish between small stuff, like wasting time on the Internet, and terrible ideas, like giving up on effective altruism entirely. And since giving up entirely would let you avoid a whole bunch of difficult things, you might burn out and quit, which would suck. So if you’re going to fault yourself for something, make sure you do it in proportion to the disutility it caused (and make sure you have an accurate conception of what that is!).

But I don’t think it’s even a good idea to fault yourself for small things at all. If what you’re trying to do stresses you out, I don’t think it’s even “optimal” to begin with. Or maybe it’s optimal for a perfect utilitarian robot agent, but it’s certainly not optimal for a human. Your brain has reactions to things that you just can’t help, like getting stressed out when you try to do too much, or depressed when you have no social contact. Trying to fight against these reactions, or suppress them like a perfect utilitarian robot agent would, is going to be about as effective as trying to get past a mountain headbutting a tunnel through it: if you were a perfect mountain-headbutting robot it would be the straightest route, but you (and the rest of us humans) just aren’t built like that.

The right answer in that case is to walk around the mountain instead. Similarly, trying to accomplish things under stress isn’t sustainable, and imposes both a drain on day-to-day productivity and a risk of burning out. Many of the most productive and effective people I know are barely stressed at all.

I think the culprit of stress for many anxious EAs is lack of compartmentalization. Now, to really understand the ideas of effective altruism, you need not to compartmentalize too much—for example, I have a roommate who buys the idea of altruism in the abstract, but doesn’t do anything about it because he separates his brain’s “abstract morality” module and its “decide what to do” module. Because of things like this, compartmentalization has an often-deserved poor reputation for letting people evade cognitive dissonance without really coming to terms with their conflicting ideas. But compartmentalization isn’t always maladaptive: if you do it too little, then whatever you care about completely consumes you to the point of non-functional misery.

Effective altruism requires less compartmentalization than the average person has, so standard effective altruism discourse, which is calibrated against the average person, tries to break down compartmentalization. But you probably aren’t the average person. If you’re stressed out about effective altruism, ignore the standard EA discourse and compartmentalize more! For instance, Jeff and Julia (and many other earners-to-give) made the deliberate decision to compartmentalize their spending on themselves and on charity, so that they weren’t agonizing over every decision.1

Of course, this kind of thinking can be used to rationalize being too self-interested and not working hard enough. I don’t know of a good rule for when it’s OK for us to lower our standards for ourselves and when we should suck it up and maximize aggregate utility. But if you feel like you’re at war with yourself, if trying to be an effective altruist is psychologically difficult, you’re almost certainly erring in the opposite direction. If that’s you, don’t worry all the time that you’re satisficing instead of optimizing: sometimes, satisficing is optimizing.


  1. See also Julia Wise’s essay Cheerfully

2 commentscomment


Effective altruism requires less compartmentalization than the average person has, so standard effective altruism discourse, which is calibrated against the average person, tries to break down compartmentalization. But you probably aren’t the average person. If you’re stressed out about effective altruism, ignore the standard EA discourse and compartmentalize more!

The general problem of meme targeting is described some in Scott Alexander’s post, all debates are bravery debates.


Thanks for this post! Until I read it, I thought work-life balance had to be meticulous, felt guilty about possibly taking “too much” leisure time, and dreamed of having a computer optimize and micromanage my work-life balance 24/7 the way some people were doing (and I was in no hurry to do) with nutrition and fitness.

I think this post is particularly important for those of us in graduate school. Unlike undergrads, we don’t have regular deadlines to benchmark ourselves against; and compared to people working full-time, I think I speak for most of us when I say we’re less concerned about the standards we’re setting for our peers (or else student unions would focus on negotiating breaks as much as labour unions do).

comment