Finding your passion and altruism

I.

80,000 Hours recently revised their career guide and published it as a book, also confusingly called 80,000 Hours. 80,000 Hours¹ is now significantly more AI-forward, urging readers towards careers that make sense if we expect transformative AI to be developed in the next decade.²

Most of this review is, in classic style, an essay loosely inspired by the book in question, so I thought I’d say right up front who should read 80,000 Hours (book):

• If you believe that transformative AI will likely happen before 2035, and you feel scared or overwhelmed or unsure what to do, you should definitely read 80,000 Hours.

  • Pay particular attention to chapter 8, which explains how to prepare for the AI future and is useful even for readers with no altruistic inclinations.

• If you don’t believe transformative AI is likely to happen before 2035, want to devote your career to ambitiously trying to do good, and aren’t sure how, 80,000 Hours is a useful book to read, but you’ll need to think carefully about how to apply what they say to your own situation. Don’t take the advice literally; instead, use the broader principles and adjust the concrete advice to your own beliefs.

• If you want to do some good but don’t want to reorient your entire life around doing good, Giving What We Can is more your speed.

• If you’re already heavily involved in the effective altruism or AI safety communities, 80,000 Hours contains very little information which will be surprising to you.

• If you’re a mid-career or late-career professional and want to transition into doing good more effectively and/or prepare for the AI transition, 80,000 Hours can still help you. Although historically the career guide was aimed at early-career professionals, the new revision is aimed equally at people throughout their career.

(conflict of interest note: I was paid to provide feedback on drafts of this version of the career guide. I can assure you that everything you like about the book was because of me, and everything you don’t like was put in over my vociferous objections)

II.

Chapter 1 of 80,000 Hours opens with the advice not to follow your passions. Your interests change over time. You would probably be unhappy doing a job you loved if you got paid very little, worked extremely long hours, or feared being fired. What matters for career happiness isn’t your passions, but engaging work that helps others that you’re good at, and that you do alongside colleagues you really like. In short, you should try jobs you don’t feel passionate about right now, rather than dismissing them out of hand.

I agree with this advice, but I also feel uncomfortable whenever I read it.

When I was nine years old, I decided that I would have to choose between being an actor and a writer, and I picked writer. Ever since I was about twelve, I’ve written on about 80% of days (higher if you exclude bad depressive episodes where I also wasn’t leaving the bed or brushing my teeth). I have made occasional gestures towards some other career path—I did a coding bootcamp, I briefly worked on wild-animal welfare lit reviews—but doing anything other than writing feels torturous.

I absolutely did not consider the most pressing world problems and my comparative advantage in solving them given my abilities, personality, and the distribution of people already working on them. I wrote BE A WRITER on the bottom of the piece of paper and then backfilled some argument that this was the utility-maximizing thing to do.

One of my closest friends is an AI safety researcher. He liked robotics when he was a kid, and then realized that AIs were the coolest part of robotics, and never switched back. He works constantly.³ I have heard him say “I’m glad my company has given all the employees today off, so I can get some AI safety research done without having to be interrupted with meetings.” I have multiple other friends who have literally never heard him talk about something other than AI. I manage to keep our AI conversations to about a quarter of our interaction time, through techniques up to and including announcing “I’m bored of this, talk about something else.” At one point he said to me, “Max Harms’s Red Heart is really good! You should read it. It’s SO accurate about AI.”

(I then clarified that ‘accuracy about AI’ is not, in fact, my primary criterion for book quality)

The thing is. The thing is.

I can’t really assess the quality of AI safety research, but my friend certainly has an impressive resume. And I know other people who do AI safety research and bounce around from independent grant to independent grant, from gig position to gig position, never quite getting discouraging enough signals to quit, never getting their names on a paper I read about in Transformer. And all of those people are well-rounded people with diverse interests who can talk about as many as several topics other than AI.

And when I look at other people I know well who are very high performers in their fields—

I see programmers who complain “I’m working so much at my startup so I don’t have any time for hacking on open-source anymore.” I see writers who have fun by researching their books and writing fanfic. I see my friend the elementary school teacher who, as a form of recreation, writes fantasy lesson plans for high school. I don’t see a lot of people who know how to shut up about their jobs.

I see people who are weird and obsessive and, well, passionate.

In a piece on an entirely different subject, Ruxandra Teslo contrasts normal jobs—“fields with constant returns to hours – where each hour worked is about as productive on average as the last hour worked, like fast food workers or pharmacists”—with “greedy careers”:

[c]areers that are greedy for the employee’s time and pay more per hour to the people who work the most. Consider a corporate lawyer working on a deal. The initial hours are spent getting familiar with the material and the people involved. Later hours – once the lawyer understands the case – are much more productive than those at the start. A person working a 40-hour week in this scenario does more than twice as much work as a person working a 20-hour week. Lots of careers are like this.

Some people apparently manage to have greedy careers without any particular passion for their careers (they’re BigLaw lawyers or consultants). I haven’t met any of those people, maybe because they’re too busy working. Everyone I know who is wildly successful at a greedy career is successful because they want to work sixty hours a week and feel annoyed when circumstances arrange themselves such that this isn’t possible. They think about their work at parties and on dates and in the shower and when they’re falling asleep at night and first thing when they wake up in the morning.

We keep asking wildly successful people in greedy careers how to become a success, and so they keep telling people to “follow your passion.” It worked for them! But this is terrible advice. People with passions don’t need it: even if you wanted to, you wouldn’t be able to convince me not to write, or my friends not to do AI safety research or elementary-school teaching.

And most people don’t have a passion. If you ask them what their passion is, they’ll produce something that is more accurately termed “a pleasant daydream about a nice-sounding future with all the downsides left out.” Millions of Canadian teenagers say their passion is sports, and maybe two dozen put in thirty hours a week of training. Most people who say their passion is writing fiction aren’t passionate about sitting down at a keyboard to make up stories about imaginary people; they’re passionate about looking at a shelf full of books with their name on the cover.

Teenagers are particularly unlikely to have found their legitimate passion. Teenagers have been exposed to a tiny fraction of possible experiences. Most teenagers with the potential to be passionate about teaching or social work or development economics or AI safety research would have had no opportunity to discover this fact.

So I think it makes sense to advise people who don’t have passions. You can still do good work if you don’t have an insatiable itch in your soul. If only people with a passion for AI could do something about AI risk, humanity would be doomed and we might as well wrap up this whole effective altruism business and spend our last decade partying.

But... I do think you will perform at a higher level if you find work you can’t and don’t want to stop thinking about. And it seems like there ought to be some advice that helps people find this? 80,000 Hours sort of gives relevant advice. They recommend ‘cheap tests’ for personal fit; one cheap test is doing a small project that gives you a sense of what it’s like to actually do the work. If you test a dozen likely passions, you’re more likely to find yours. At least it works better than daydreaming about the [Your Name Here] shelf at the local library.

But 80,000 Hours doesn’t say “if you do a small project and you’re obsessed with it and you can’t stop thinking about it and you’d rather do it than anything else in the world, definitely pick that one.” Indeed, they’re pretty vague about what it means to have a high level of personal fit for a job.

I don’t think that we should assume people already know what their passions are, or that you can figure out your passion without trying things. But I do think people will do higher-impact work if they’re passionate about their jobs, and I worry that dismissing “find your passion” out of hand makes people less likely to find jobs at which they can genuinely excel.

III.

Who is the target audience of 80,000 Hours’s career guide?

Imagine ranking everyone in terms of how much they improve the world You’ll find that the 1% of people who Did The Most contribute more than the other 99%; the 0.1% of people who Did Even More Most contribute more than the other 99.9%; the 0.01% of people who Did Even Even More Most contribute more than the other 99.99%; and so on and so forth. A friend of mine donates 10% from her minimum-wage job and saves a bit more than one life per year. I move tens of thousands of dollars through my donation blog posts. Another friend of mine once donated almost twice the U.S. median household income in a single week.

In terms of moral praiseworthiness, I and my two friends are in a pretty similar place. We’re all making significant personal sacrifices to live out our values. But in terms of positive effect on the world, Rich Friend is doing forty or fifty times as much good as Minimum-Wage Friend.

So, like, if Rich Friend donates an additional percentage point of their income, they donate many times more money. If you convince Rich Friend to switch their donations to a more effective cause area, it affects much more money. It is far more important to convince Rich Friend to be effective and altruistic than it is to convince Minimum-Wage Friend to be effective and altruistic.

I want to emphasize that this is not a criticism of Minimum-Wage Friend. Many of the differences between the Rich Friends and Minimum-Wage Friends of the world aren’t because the Minimum-Wage Friends are lazier or made worse decisions. Some people grow up in loving upper-class households, while other grow up with abusive parents in poverty. Some people are naturally gifted at machine learning or quantitative trading; other people are naturally gifted at early childhood education or animal care; still other people have no particular gifts at all. Some people have disabilities that make it impossible for them to go to college or to work the most remunerative professions. It is tremendously, profoundly unfair that you can do far more good for the world if you’re a neurotypical, able-bodied person who grew up in a developed country with a loving, wealthy family and who has a natural flair for biosecurity policy.

From a moral perspective, what matters is whether you made the best use possible of your talents. From a prioritization perspective, we want to give advice to people who can have the highest potential to make the world better.

There’s also just more to say to the highest-achieving do-gooders. For 99% of people, the only chapter of 80,000 Hours they need to read is chapter 3, which covers how to do good in any job. You can donate money, advocate for important causes through social media and contacts with friends, vote, and make it easier for other people to do good. Chapter 3 leaves out some important details—for example, the importance of salary negotiation, which can easily leave many people with ten thousand more dollars to donate. But I think as a first approximation chapter 3 about covers it.

High-achieving people need the rest of the book. They need to choose between a priority path they have a low level of personal fit for, and a less important path that they’re better-suited for. They need to trade off between developing career capital and having a direct impact today. They need to decide whether they want to earn to give, do research, do policymaking, or advocate for important ideas. They need to not burn out!

So 80,000 Hours’s advice is going to disproportionately go to 1% of its potential readership—the potential readers whose decisions are more important, and who need more detailed advice about how to make them.

When I make this argument, people often say to me “well, why do they advertise the book to everyone else then? Can’t they just put in the preface ‘this book is for people who attended Ivy League schools and Oxbridge. If you’re a normal person, read chapter 3 and then go on with your life’?”

The problem is that there is no one single metric for how much potential someone has to make the world better. Many people who attended Harvard are complete incompetents who can best advance the goals of high-impact organizations by staying as far away from them as possible. Many unambiguously high-impact people in effective altruism or AI safety have resumes that look unimpressive or just weird. Eliezer Yudkowsky never attended high school!

Historically, 80,000 Hours had a bad habit of assuming all its readers graduated from an elite college. The recent revision does a good job of welcoming readers from a wider variety of backgrounds (although it still occasionally assumes that the reader is actively exploring the option of becoming a management consultant). I think having a vaguer target audience is important in order to capture the many people with high potential who have a resume full of holes.

In my experience, the second most important factor in whether you can have a high impact (after intelligence) is something that you might call “grit” or “ambition” or “determination” or “agency” or “self-confidence” or “hard work” or “stubbornness.” People who change the world set demanding goals for themselves, work long hours, try a hundredth solution to the problem if the first 99 don’t work, and keep bashing their head against the wall long after any reasonable person would have retired to start an emu farm.

(I think this is what a “passion” is: the activity where you have access to secret reserves of hard work and perseverance and agency that you can’t apply to anything else.)

It is often the case that you can transfer to a particular career path if you’re the sort of person who will put in twelve hours a day studying machine learning papers on your own, and not if you’d put in twenty minutes of effort and then spend the rest of the day binging Netflix.

But, like, how do you write a self-help book that says “this will work if you have hard work and perseverance and agency, and not otherwise”? Everyone’s going to read that as “you should be able to do this, if you were good enough, and also the fate of humanity hangs in the balance (no pressure).”

I don’t necessarily think you should be able to do this. We moralize a lot about hard work and perseverance and agency, but I have yet to see much evidence that you can make yourself have more of these qualities. When I have asked hard-working and persevering people of my acquaintance how they got that way, “I don’t know, I’ve always been like that” is way more common than “well, I used to be lazy and easily stymied, but then I discovered this One Weird Trick, Invented By A Mom—” I have hard work and perseverance and agency for writing and only writing. No matter how frustrated I am by the fact, I have not managed to evince any hard work or perseverance or agency for cause prioritization research, political campaigning, fundraising, not being suicidal, or answering emails in a remotely timely manner.

On the other hand—your talents, your disabilities, your family background, and your class are neither your fault nor morally wrong. But, speaking as a lazy person, being less hardworking and determined than you could be... is... kind of... your fault? Not in a sense where hating yourself about being lazy helps (hating yourself about your personality flaws never helps and usually makes the situation worse). But if you’re like “perseverance and hard work are innate, it doesn’t make you a worse person to give up easily and hate doing work,” I don’t know if anything makes you a worse person. Kindness and courage and humility and self-control and loyalty and not being racist are also all famously hard to get more of by wanting to.

But the 80,000 Hours career guide is always going to be alienating. As far as I know, 80,000 Hours tries to target readers at a certain level of functionality, competence or potential for competence, and legible success, so that more than 1% of its readers are in the 1% of people with the highest potential to improve the world. But ultimately, no matter how well-targeted the marketing is, most readers aren’t going to be able to take most of the advice—often not for any reason you can put in a resume, but because they lack the personality traits that make someone achieve extraordinary things, in a way that is at least in some sense their fault.

IV.

I am weirded out by the extent to which 80,000 Hours is a selfish book.

To be fair, the selfish component fades out over the course of the first section, and is basically absent in the second through fourth sections. But the first chapter is entirely devoted to how you can find a truly satisfying job, and concludes the answer is to find work that matters. The fourth chapter emphasizes that you can do a lot of good in the world at very little personal sacrifice, and doing good will probably leave you better off because of the happiness and sense of meaning it gives you.

Now, it is both true and important that you can do a lot of good at very little personal sacrifice. If you can spare $4,000 a year—well within the capacities of an average American household with a bit of budgeting and prioritization—you can save a life once a year, every year. Being an effective altruist means that I do interesting work, surrounded by kind and fascinating people, while continuing to be in the top 1% globally for household income but by a somewhat less embarrassingly large fraction. I have long said that being a virtuous person is good for you.

But... I’m not an effective altruist to meet nice people, have a more engaging job, or have a sense of meaning. I’m an effective altruist because I think it’s bad when parents bury their children because they couldn’t afford a five-dollar medication, or when people’s roofs collapse on them several times a year, or when chickens die of gangrenous dermatitis because they spend their lives standing in their own feces.

If all I wanted was a sense of meaning, why would I bother with this effectiveness stuff? My sense of meaning isn’t scope-sensitive. I could become a sex therapist and get paid $250 an hour to help the upper-middle-class have orgasms. It would be way less stressful and probably more meaningful, because I’d get to see the people I helped.

I feel stupid saying it, but I’m an effective altruist because I care about other people and animals, not as a galaxybrained self-help strategy.

I’ve said before that I think the right amount of resources to put into altruism is an amount that makes you wince a little bit. Not a lot! I’m not suggesting that you donate everything you own to GiveDirectly and live under a bridge, or work a job you despise, or live in a way that fills you with ennui and despair. But if you never write a check and go “ouch, I wish I could have had that vacation,” what are we even doing here?

I understand why 80,000 Hours doesn’t want to go into moral philosophy, a discipline mostly notable for how utterly it has failed to convince anyone of anything. But... shouldn’t we be trying to convince people to do good in the world with reference to how awful the world presently is?

I know some people who are passionate, not about AI safety or writing or teaching, but about doing good. They are some of the highest-impact people I’ve ever met, because instead of optimizing around their coincidental interests and enjoyments, they can just do the next good thing.

I don’t know how to become one of them myself; I don’t know how to make more of them. But if you’re writing a book aimed at trying to get people to do good with your careers, shouldn’t you be trying to appeal to those people?

80,000 Hours presents effective altruism as the sensible, prudent thing to do. But I don’t think anyone has ever done great good things because it seemed sensible and prudent. Because it seemed noble, or heroic, or glorious, yes. Out of righteous anger, or grief about suffering, or pride in their work, or love of God, or a desire to atone, yes. Because it might impress pretty people, or because everyone else was doing it, or because it seemed like a good idea at the time and they got a bit carried away, a shocking amount of the time. But rarely because doing great good things seemed sensible and prudent, because it’s not.

This is Eunice.

She was the first person on GiveDirectly Live, the day I wrote this, who had a picture.

She lives in Kenya. She is pregnant and has to walk three kilometers to the nearest hospital. Her spouse makes $5 a day as a day laborer; she can’t work anymore because she has stomach pains from her pregnancy. She loves her husband, and is excited to welcome her child into the world. She is grateful that, in spite of her poverty, she still has her health. She has currently received $56 of the $860 she will receive through the Kenyan Mothers and Babies program. She wants to use the money to buy a cow, open a small store, and pay for the hospital delivery of her baby.

From a selfish perspective, being an effective altruist is a stupid thing to do. But I’m not doing it for me. I’m doing it for Eunice.

1. Book, but like also organization.

2. Which isn’t the same thing as careers in AI safety research—for example, transformative AI increases the chance of a catastrophic bioweapons attack, so 80,000 Hours also recommends working in biosecurity.

3. When I fact-checked this section with him, he said that he worked 55 hours a week which is, quote, “really not crazy.”