Summary
- Get clarity on your options. There’s no such thing as a good or bad idea. It’s always good or bad compared to something else.
- Don’t just decide if this charity idea is “good” or “bad”. Compare it to all the options
- There’s an AI safety job bottleneck. If you can start an AI safety charity, it’s quite likely that you should because then you “unlock” a ton of marginal people’s talent who wouldn’t otherwise be able to get the very few AI safety jobs available.
- We need to try many things to avoid x-risks. Given that we haven’t found the answer yet and the astronomical stakes, we need to try a lot of things.
- Freedom. You get to work on what you think is highest impact.
- Do you need more experience? Probably not. There are more efficient ways of learning.
- Most jobs give you very narrow experience. Say you work at a think tank as a researcher. You may get tons of experience researching, but none in lobbying, management, fundraising, operations, etc.
- It’s a crapshoot whether your manager will be a good teacher. Your manager might teach you a lot, but they very often won’t.
- Get experience by doing the thing you want to learn. It’s way more relevant to what you’ll actually need.
- Mentors > experience. Get a bunch of mentors and coaches who can give you advice specific to your situation. Way more efficient
- Read and take classes. Set up a practice of constant and lifelong learning. There are plenty of classes and books out there to teach you all you need to know.
- Just hire for the skills you need. You don’t need to know how to do most of the skills. As a founder, you need to be able to hire and manage people who have those skills. Here’s a crash course on how to hire people.
- There are always going to be people telling you to get experience before starting something. Those same people will often be calling you an “expert in the field” after you’ve been running an org in the area for a year.
- Entrepreneurship is more in the realm of art than science. There is no single “best practice”. Nobody knows for sure what is best. There’s no one way fits all.
- Depends on your timelines. If you put non-negligible probability on shorter timelines, experience by doing + mentorship/coaching + reading/classes will make you learn faster.
- The experience of entrepreneurship varies wildly based on the org. Not all orgs are the same. Some you might love running, others might make you miserable.
- You don’t have to start a big org. You can stay small if you want.
- Larger organizations tend to have more impact but less freedom for the founders.
- You can start a “lifestyle business” but for charity, where it’s not massively ambitious, but does a lot of good while also maintaining a balanced lifestyle.
- Reasons to not start a charity
- If you’re generally afraid of things. You can still start something if you’re prone to anxiety, but assume your anxiety will go much higher.
- Mental health issues. You can start a charity and have mental health issues, but it is going to be a lot harder. Generally speaking, the more mentally robust you are, the more suited you are to charity entrepreneurship. It is hard and it will push you to your limits.
- If you interpret other people’s doubt as you being wrong/bad. There will always be doubters, especially in EA/AI safety. You need to be able to have the attitude of “I disagree and think you’re wrong and I’m going to do it anyway”.
- If you don’t really really want to. There will be endless challenges, and you need to have an unshakeable answer to the question “Why am I even doing this?” when those happen.
- The lack of security in entrepreneurship is overplayed.
- If you have a job, if just one person, your boss, doesn’t like what you’re doing, you’re out of luck. With an org, if one funder doesn’t like what you’re doing, you can get other funders and keep doing what you were doing.
- For most people, the worst that will happen is having to sleep on the couch of a friend or family.
- You can apply to Nonlinear’s career advice. I give free career advice sessions to people who are considering starting an org in AI safety. You can learn more and apply here.
- List of AI safety ideas. If you decide to start a charity, check out this prioritized list of ideas or add your ideas to the sub-reddit to get feedback.