Why I now identify as an Effective Altruist
For a long time, I’ve considered myself an existentialist. But existentialism isn’t a complete philosophy. In its emphasis on freedom and authenticity, it tells me that any path I choose consistent with my identity is sufficient. What is my identity? Of the available paths, are some better than others? What does better mean in a world without intrinsic value?
Two weeks ago, I stumbled on mindingourway.com and found a perspective I’d never seen before. The author, Nate Soares, approaches the question of how to live from evolutionary, scientific, cognitive psychology, and emotional perspectives. His writing is clear and immediate, full of concrete instantiations of general concepts. He doesn’t prescribe. He delineates who his writing might help, and who not. He’s explicit about his levels of uncertainty. He doesn’t depend on any philosophical or ethical frameworks.
The Listless Guilt
You don’t have to care about anything. If you don’t care, you may feel the Listless Guilt. The Listless Guilt is a nonspecific guilt: “shouldn’t there be more to life than grinding my day job, binging Netflix, and gastronomically touring restaurants?” It is distinct from caring about something and not working on it hard enough, which is a specific guilt. If you feel the Listless Guilt and don’t like it, the antidote is to care about something. You can assuage the specific guilt by working on the problem.
Working on various startup ideas, I often asked myself if I was missing something. Surely there’s more to life than making a 0.0001% improvement in economic efficiency. Perhaps if I took more risk, I’d feel something. Perhaps I was feeling the Listless Guilt.
Seeing the dark world vs tolerifying
There are all sorts of awful things in the world. In San Francisco, I see it every day. Homeless people sitting on the sidewalk of major thoroughfares shoot up with meth and heroin as techies like me walk by and try not to look. They walk around like zombies in various states of disrepair. We’ve all heard the statistics of how many people live in poverty: 1.4B living on less than $1.25/day. In order to get on with our lives, we weave narratives that tell us it’s okay for us not to help.
- “It’s not my problem.”
- “If I helped it might create bad incentives. It might tell people they don’t have to help themselves.”
- “The problem is too big. If I helped, I’d only be able to affect a negligible amount of people.”
- “Most volunteer organizations are about making volunteers feel good, not about creating lasting change.”
- “My donations would probably be spent on advertising instead of actually helping.”
These statements are all “tolerifying,” finding ways to tolerate badness and thus suppressing the emotional response.
I thought I didn’t feel anything for people suffering halfway around the world. When I thought about it and made even the briefest attempt to imagine what they must go through every day, “the Grinch’s small heart grew three sizes” (Dr Seuss). I do care, and I bet you do too. Try it.
When I meditation retreated, I learned that both joy and suffering come from attachment. To me, these emotions are part of the human experience, not something to be avoided, as long as I’m actively accepting them. Choosing to suffer generates meaning and emotion. Now when I suffer, I ask myself, “What attachment is this suffering coming from?” In seeing the dark world, I see a path to something better, but we’re not there yet. The emotional acknowledgment that things are bad underpins my new motivational system. This is a suffering worth keeping.
Tribalism vs aesthetics
Our emotional reactions come pre-installed by a combination of biology and experience.
The biology piece made a lot of sense in the environment of evolutionary adaptedness (EEA). Humans lived in small tribes that competed against one another for scarce, limited resources. They vied for social and political status and made decisions that pitted in-groups against out-groups . There’s actually not much evidence to support evolutionary psychology claims like this, but the point still stands that people exhibit maladaptive behavior in modern society.
In the West, we’ve somehow defaulted to a culture of maximizing personal wealth. Instead, imagine the culture was about making the lives of all humans as good as possible. Imagine a world where everyone wants to help everyone. Someone standing next to me would have no particular advantage over someone hundreds of miles away, except insofar as I can understand their problems. Even if coordination is hard, doesn’t that seem strictly better? As we lift entire regions out of poverty, imagine the resulting amplification. Those who would have spent their lives incapacitated by malaria instead enter the world stage, making critical breakthroughs in science and technology.
When I try to imagine a world better than what we have, I don’t have to try hard. In Enlightenment Now, Stephen Pinker’s extensive data shows we’re moving solidly towards abundance. It’s not evenly distributed, and it won’t be for a long time. Giving abundance a boost seems like something worth fighting for.
When I introspect and compare my desire to help my friends and family against my desire to help distant strangers I’ll never meet, the former win, hands down. The comparison I’m actually doing is “tribal loyalty + saliency bias” against “aesthetic dreams for a better world.” In an environment optimizing reproductive fitness, it seems obvious which of those should win: tribal loyalty. We no longer have to optimize for reproductive fitness! We can choose to act in accordance with our human values instead. It’s okay that our biology and our values are sometimes in conflict with one another. In fact, we can change our minds any time. Either way, it’s important to acknowledge that these forces are often conflicting. I’m choosing to favor my aesthetic values.
NB: I’m using tribalism and aesthetics differently here than the usual definitions. By tribalism I mean “bias towards friends and family” and by aesthetics, “my preferences for what the world should look like.”
Self care and great expectations
Since realizing I care, it’s easy to feel crushed by the enormity of the problem. Every minute someone is dying. Every extra dollar I spend on myself could have bought someone else an entire day of livelihood. If I burn out, though, my impact will decrease until I recover. So I need to find a way to maximize my long term impact.
I’m just one person. My mind can only change so fast. I’ll make mistakes, and sometimes hurt when I’m trying to help. I can only ask myself to do my best. It will have to be enough. My goal is to maximize expectation (probability times outcome) over time, not at any particular moment. It’s a marathon, not a sprint.
How I’ll change the future
Nate ends with a plea for us to create the future filled with goodness, however it is we define that. My current goal is to reduce human and maybe animal suffering. I think in terms of leverage. If I give a dollar today, I can have 1x impact. If my investments beat the market, I might have 1.1x impact. If I eventually start some kind of grant program that changes others’ life paths towards making the world better, perhaps I’ll have 10x impact.
Now consider that creating powerful AI might be the last invention humanity ever needs to make. If it goes poorly, it could mean a bad outcome for humanity forever. If it goes well, the sooner it goes well, the better. Even if there’s a tiny chance it will fail, if I can shift that by even a tiny amount, my positive contribution could be huge.
Two years ago, I was working at OpenAI for a similar reason. After trying my hand at OpenAI and MIRI, I became disillusioned about my ability to have impact. Part of the problem was the Sour Grapes Fallacy, and the other part was my longish timeline, which has since shortened because of GPT-2.
I think we’re still relatively far from human level AI, dangerous or not, but now is the time for me to get in front of the problem. If I wait until it’s obvious how to get there, it may be too late for me to ramp up. For this reason, I’m planning to shift my focus back to AI safety. I’ll write another post explaining why I’ve chosen this path instead of the many others available to me.
Meanwhile, I’m considering the implications of being an Effective Altruist. I’ll start with small changes that I mostly won’t notice, like being flexitarian: choosing vegetarian options on menus and making vegetarian choices when grocery shopping. Once I have an income (soon!), I’ll donate 5% to a GiveWell charity or a grant program. Easing into these changes reduces the chance I’ll want to leave because I’m feeling burdened. As I grow more comfortable, I’ll do more.
I could have distanced myself from the official movement, which conflates its mission with the sociopolitical makeup and idioms of its members. I believe that’d do a disservice to the movement. By taking up the values along with the label, I hope to show that it’s possible to be part of the movement in my own way. Perhaps I’ll inspire others to do the same.
This post is a rough approximation of how I changed my perspective on what I’m doing on this planet, inspired by mindingourway.com. The ideas are still fresh and raw in my mind and may have some bumps or inconsistencies. I’ve gotten excited about philosophies before only to abandon them after a year or two. I hope this time I approach these ideas with more circumspection. If you’re intrigued by, disagree with, or feel anything about what I’ve presented, let’s chat! To a better world!