Opening a Soul-Saving Shop

Effective Altruism that shows people how we got here, not where we ended up

Last year, I lamented that the Soul of EA is In Trouble because EAs were downplaying their true motivations in favor of expedient ones. This hits a growing movement especially hard because, when people interface with it, they’ll think the movement is about whatever vaguely palatable conception of “AI safety” they were pitched, and people who think that will make up more and more of the movement over time.

I disagreed then and I disagree now that it is actually expedient to repress our true motivations. Those motivations are beautiful: care for literally everyone, willingness to sacrifice personal advantage, the desire to believe true things as best you can, no matter what you thought earlier. These things have produced all the best EAs in the past. They are the messages we should be sending to the EAs of the future. It’s not hard to convince people of this intellectually. It is hard to convince them of it socially. We live in a detached and cynical era. Overly earnest and wholesome people come off as at least naïve, if not stupid, if not weak, if not crazy.

As many EAs have experimented with, perhaps the move is to toughen up. Play the cool, technical expert. Talk a big game about your ties to decacorns. It’ll serve your softer, harder-to-articulate goals better. It will make sense to listeners who will want to be your allies.

The problem is that if you don’t talk about the vulnerable, aspirational, moral, spiritual(?) calling that inspired you to safeguard the future, you’ll never reproduce yourself. You might not even produce good allies. You’re only a handful of people. Once you retire, switch projects, burnout, give up, or exhaust your capital, the spirit of altruism that sent you down this road and guided your choices goes with you.

I’ve beaten the drum about how you should lean into your most genuine, aspirational beliefs for some time now. Maybe I’m wrong and it actually is expedient for *some* people to tamp down their souls. There is a need for a serious counter-balance to that though. So, in the greatest abundance of pragmatism I can muster, I’ve now founded The Network for EA Support & Training (Nest).

The hope is to invigorate the earnest, empathetic, truth-seeking core of Effective Altruism. To save the soul of a movement by igniting the souls of its members.

What is Nest?

Nest supports in-person EA communities, primarily by coaching EA group organizers. I founded it this year alongside Sarah Bluhm. In practice, we’re going to partner with 5-10 university groups and host three multi-day, in-person workshops for a broader range of community builders.

We’ll primarily help groups by: 1) being a repository of substantive EA knowledge, and 2) being a source of advice and coaching on how to communicate EA ideas.

On substantive knowledge, we’re quite interested in the abstract concepts that inform the ~EA worldview. See our working list of ~200 concepts from economics, philosophy, game theory, and statistics here, along with the beginnings of a long list of important facts and statistics. There have also been dozens of insightful, accessible write ups applying these concepts that we regard as seminal which we’ve begun to collect (and rate) here.

We’re more opinionated on the communications coaching side of things. We think organizers should aim primarily to pose important questions and fill in missing context as opposed to pitching specific causes, programs, or jobs. The world is very complicated and we’re still fundamentally bottlenecked on good ideas. Aiming to put asses in jobs does not relieve that bottleneck; helping smart people develop their own independent models might. The challenge for organizers is to balance epistemic humility with intensity about the questions that matter.

To serve these ends, we’ll also do much of the legwork of crafting plans and events aimed at finding the people at universities who are going to be interested in these questions. We have a regranting budget to help groups execute these and I am personally willing to show up and lead sessions on EA Philosophy, AI risk, AI policy, Individual career choice, and the value of community building work.

Why am I doing this?

Some kind of liberal, epistemically virtuous beneficence has to win, at least in a few critical areas. Ideas like these are downstream of the people putting them into practice, and people are downstream of the groups and communities they find themselves in. If the groups and communities aren’t great, the ideas wither and society drops increasingly important balls. I don’t take a view on whether this means catastrophe, hell, extinction, or something more moderate, but this is where the most potential value lies.

There are other ways to create a lot of value in the world, particularly by just solving big problems directly without being preoccupied by abstract ideas or community. These approaches complement mine. No one is going to be inspired by a community that doesn’t accomplish anything; no one is going to accomplish anything if no one tells them about it or (socially) rewards them for it.

I happen to think the explicit community side of things is more neglected at the moment, but I’m also biased by the fact that this is what excites me personally. I generally endorse people leaning into their personal biases and excitement to do things though, at least above some bar for how strong the impact case is.

I. Liberal, epistemically virtuous beneficence

There’s a fundamental tension between power and goodness in the world. Everyone wants goodness for themselves (and usually at least a bit for others), but you need power to get it. You need some amount of power just to survive or, put another way, to avoid being overpowered by your environment, which is constantly changing. If you’re nervous that the world is going to overpower you, you’re going to trade away lots of goodness for the power to avoid that.

A simple example is taking a higher-paying job that’s more stressful, time-consuming, and worse-for-the-world compared to alternatives. Despite all those downsides, money is power that can help you weather economic downturns or help you be more well-regarded by people who can do you favors later. The problem is that the world is big and the stakes can always be raised. There is always another, even higher-paying job just within reach that will get you more of these resources. One gets tempted to play the game to the very end, never cashing out their chips for more leisure or charity because they choose to invest them in accumulating more and more security.

Even a more beneficent version of yourself might run for school board to improve schools, but you realize you can do more as mayor, and then as a congressperson, and maybe later as president. You spend your whole career doling out favors to special interests who can help you win the next election. Eventually, you lose an election and lose the power to call in any of those favors to improve the schools. You probably lost to someone playing the same game even harder than you were.

These are extreme examples, but you can imagine the whole continuum. Someone else forswears power entirely, refusing to even invest in a college degree. They work at Starbucks and live off of reheated lentils giving every spare dollar to the local homeless. Everyone says they’re wonderful, but no one lends them a dime to help the homeless. This person would have been much better off pursuing some power of their own.

Effective Altruism is about getting this trade-off right. A flourishing world requires people who will competently pursue power, but only to the extent it serves goodness. They will cash out their power as strategically as possible to do the most good, according to some relatively robust definition of good, to which they can hold themselves and each other accountable.

II. People & Groups

The existing EA community offers some rewards for doing this well, though it has recently narrowed its scope, with the most energetic and enthusiastic rewarders of Big Good focusing ever-more exclusively on transformative AI. While it may be right to do this, it has left a gap for a broader, more vigilant EA movement to be ever on the lookout for people thinking about and executing on other global public goods and high-leverage interventions to improve lives.

Nest aims to build a more conservative version of EA that focuses on rewarding and supporting people who pursue projects inspired by high-level EA principles despite differences on narrower empirical beliefs about AI, or philosophical positions on future generations, or the value of various non-human animals, etc. Explicitly inviting more people to the table more credibly signals commitment to the broad norms from which so much hard-to-predict good in the future may come.

The character of groups that effectively send this signal will be reflected in their understanding of the perspectives of everyone who might contribute and in their ability to communicate the ideas productively to these audiences.

My theory is that the primary medium of communication will be one-on-one or small-group participatory conversations. Other talks, events, activities, and reading groups primarily serve to set up these conversations and give them content.

The conversations themselves should be anchored in perspective-sharing, where at least one party is principally interested in understanding what the other believes about which problems are important and which interventions might best address those problems. The goal is to mutually improve each other’s understanding. That will go better when people have:

1. More knowledge of relevant facts and concepts,

2. Good calibration and self-awareness about the limits of their knowledge and strength of their beliefs,

3. Genuine interest in the person in front of them,

4. Mutual willingness to change their minds,

5. The desire to reward thoughtfulness and sincerity rather than specific conclusions.

What Does This Look Like at a University?

Universities matter because they’re the place people come to answer questions of this shape at the time in their lives where they are most likely to act on the answers.

Of course, not everyone comes to university with these particular questions or even this shape of question in mind. In fact, relatively few people even have the disposition to engage seriously with these questions. There’s nothing affirmatively wrong with that, but if you imagined scenarios where *everyone,* or even more than 10% of people at your university were going to enthusiastically engage with this, you should adjust your ambitions downward. Getting 2% of your campus to meaningfully engage with your group should be considered *wild* success.

As a practical matter, the first thing to do is find people with the relevant disposition. This is disproportionately menial-labor-intensive, but there is both some skill involved and the rewards are great. Some people on your campus are already explicitly aware of and sympathetic to capital E, capital A Effective Altruism. There are at least 20 of them. There are 3-10 times as many people who think of themselves as “big questions” types, who will find the kinds of conversations you have in an EA context very stimulating.

You want both groups to know you exist and to know that your group is a good place to be. I include some basics in this footnote.¹

I. Great Conversations

The life of a group will be the discussions you have with people who enjoyed your hook events, liked the ideas, and wanted to follow up. The first of these conversations will often be a 1-1 with a board member, where you’re mostly exploring the interested-person’s ideas. Ideally, you’ll have 1-2 new ideas or readings to recommend that are both highly relevant to EA and specifically interesting to the person. Maybe it’s something that presents a head on challenge to what they’re most excited about, maybe it rearticulates a point they made from a new angle. The goal is to demonstrate that there are lots of interesting ideas here.

EA groups should be the place people go to discuss and debate how the world works, what’s going on, and what really matters. The vibe is one where mutual understanding, charity, and steelmanning are necessary stops on the way to truth. Whatever people think or want to discuss, these are places where they will feel heard and taken seriously. Part of taking ideas seriously is of course disagreeing with them, but one shouldn’t leap to disagreement before they’re confident they’ve understood the point at issue.

In practice, this will often look like short, member-led discussions, where someone writes something up or prepares some remarks on something important to the world. I’m demonstrating one model of that with this post and the in-person sessions I’ve led on it.

Group leaders should also design their own reading group curricula that are tailored to their interests and what they want to talk about. You can also organize retreats. The ultimate point of EA groups is to give people a social reason to follow high stakes events in the world, read inspiring or illuminating texts, and learn about the concrete projects making progress on things that matter.

If you create environments that do these things well, you don’t need to worry about holding people’s hands or pushing them into certain jobs. Jobs and programs may be your honest advice for some people and not others. That will come out naturally as you understand them and their values better. And understanding is the thing to aim for.

1

• Online Surface Area

  • Website w/ faces, emails, Calendlies, next event info

    • The first thing people find when they google “[university] effective altruism”

  • Linked on the EA Forum, verified active groups on Nest site

• Extremely approachable hook events.

  • Reverse donation tabling w/ fun gimmick

    • People do some action that determines how much *you* donate to the charity of *their* choice

    • Choices are spread out among EA cause areas

    • Ask people to explain their choice

    • Register their action by filling out a form where they opt in to a 1-1 and the mailing list

    • Actions include, e.g., $0.20 per push up, three dart throws with different values on the board, draw your charity, get someone else to recognize it

  • Popular Nerd-brand meetups

    • Run meetups for popular over-lapping brands like Kurzgesagt, Veritasium, Hank & John, VSauce, ACX

    • Let people know the EA club organized these and give a spiel about what you do

  • Provocative talks

    • “Will AI Kill Everyone?” “Should *You* Give Half Your Money Away?” “Should Vegans Eat Beef?”