At a conference this past weekend in DC, a friend invited me to share some thoughts about how to avoid “burning out” while working to mitigate AI’s dangers. Below, moderately edited, are some remarks I prepared about how I keep my own fire alive. Thanks to my colleague Peter for some key insights.
“Burnout” is an evocative term for a loss of passion in one’s work. We often have a fire for the work we do, but without the right fuel, that fire can flicker and die. In a short post on Morale, J Bostock suggests that the fuel we need is a clear signal that good things come from effort; that our work drives real progress towards the things we value.
In the four years since I decided to quit my job and throw myself into the problem of AI, I think that I’ve managed to personally reduce the chance we all die by a few fractions of a percent. That may not sound like much, but in a world of billions of people, it’s enormous. I’m deeply proud of it, and I hope to do more. But I didn’t always feel like I’d be able to make a difference.
For the better part of two years, I flailed around without much direction. It sucked. I had precious little purchase and no real mentorship save the written words of people far away. It felt like climbing a slippery wheel in the dark.
Overcoming this took a mixture of luck, persistence, financial runway, the enduring patience and support of my wife, and being willing to locate and perform me-shaped gruntwork.
I did a spot of data analysis; I volunteered to teach courses in AI safety; I got a part-time job with the folks who wrote the course material I’d been using; I applied to 50-ish positions and got rejected from 47 of them. I said yes to a few too many things.
Eventually, I found my feet. Everyone is different, and your path will not look like mine, but several factors have helped me stay focused and whole.
1. My center is secure.
My wife loves and supports me. I have a life outside my work and it’s amazing. I visit my friends, I play with their kids, I delight in the joy in their eyes.
It is technically possible to do good work while being a miserable wreck. But if you don’t feel physically safe, emotionally whole, and financially secure, it’s okay to prioritize getting your life in order.
2. I play to my strengths.
I have skills and fascinations, the things I can do and the things I live and breathe. I can speak well because I’ve practiced speaking. I can write well because I’ve practiced writing. I can run a group meeting because I’ve done it for years as a facilitator; and also because running tabletop roleplaying games for rules-lawyering nerds transfers surprisingly well.
I thrive on bringing people and ideas together in combinations that bear fruit. I have a deep-seated obsession with saying complicated things in few and simple words.
These are my superpowers, my tools of the trade. I have found a niche that lets me use and temper them.
Ask yourself: What do I know, and what do I breathe?
3. I take breaks and mean them.
Two weeks ago, I was hiking in the Catskill mountains with one of my best friends in the world. I stared at the stars, wore the soles off my boots, and turned a muddy puddle into fresh drinkable water with the twin powers of friendship and science. I talked about Star Wars metaphysics and listened to the epic tale of a ten-day Avatar LARP. I barely thought about the future at all.
Do what you can when you can do it, and spend the rest of your hours living.
4. I’m following my heart.
I read Atlas Shrugged when I was a teenager. Ayn Rand’s philosophy is a mixed bag, but one thing I think it got right is the attitude towards moral duty. From Rand, I learned something that really ought to be common sense: What I consider “the right thing to do” isn’t imposed on me by the universe, it’s derived from my own values. I want to better the world and save its people, not because I must, but because I care.
There is an astounding difference between work you see as a moral duty and work that leaves an imprint of your soul’s deepest yearnings on the world, even when it’s the same act of labor either way.
When I ask my soul what it wants to do about the threat of extinction from AI, the answer is clear: Face the abyss with laughing eyes and a flickering candle clutched in my fist.
We may all die, but you can choose to die fighting on the side of life.
5. It’s not about my feelings.
The feelings and attitude with which we approach our work can matter enormously.
But it’s also important to remember why we are here. We’re here because thousands of people and who knows how many machines are working to build superhuman, alien minds that may one day kill every man, woman, and child on Earth. The mission is to chart a better course. It’s not about your feelings. It’s about saving the world.
Your final utmost measure is the shape you hammer into the future by your own hand. At the end of the day, you should find the most important step that you, personally, could possibly be taking, and do it. Then do the next step, and the next, at whatever pace you can lastingly sustain.
If you have the slightest political leverage, use it to push for an international treaty that halts the race to build superintelligence. If you are technical, figure out how to verify and enforce a treaty. If you have an audience, warn them what’s coming and how they can help. If you have better ideas, do those instead.
If none of the above, start reaching out to every human being you know. Try to have a frank and honest conversation with them, and invite them to speak their own concerns to power. Keep it up until you run out of humans or have a better idea.
On this, morale and effectiveness agree: pursue clear, legible progress on an urgent problem. Years ago, it was perhaps discouraging to work on the extinction threat, because opportunities were scarce and progress abstract. For good or ill, that’s less true today. We have more urgent calls than we know what to do with. In a burning world, it’s easy to find your fire.