What is an effective altruist

One of the things I did while working at 80,000 Hours in 2012 was to write what I believe to be the first statement of what it is to be an effective altruist.

An effective altruist is someone who makes helping others a significant part of their lives and tries to make their help as effective as possible.

Effective altruists believe that helping people is a good thing. They believe that it matters a lot to make people’s lives better. It matters even when those people are far away and are people you’ve never met.

That desire to make people’s lives better isn’t a political idea. We have members who are hard-core libertarians, far-left socialists and everything in between. Some of us are religious, agnostic, or atheist. In short, there are a lot of different reasons for wanting to do good.

The other thing that unites effective altruists is that they want to make as much of a difference as they can. They think it’s worth thinking your options through carefully, to make an estimate about which actions will make the world the best place. So when they choose between two jobs, they think about the difference each job would make compared with what would happen if they didn’t take it. When they give to a charity, they take estimates of how much good the charity will do with that money very seriously.

An effective altruist recognises that estimates only get you so far. But they are continually surprised by how far they go – our research shows that even on a rough estimate some careers do tens of thousands of times more good than others.

Although there are a few things that I would now emphasise differently, I’m pleased by the broad brush strokes. In particular, it mostly avoided maximalizing claims which I tend to think are often counterproductive (although it does talk about making ‘as much of a difference’ as one can), and it doesn’t over-emphasise quantitative comparison (although the number ‘tens of thousands’ was both unnecessary and probably incorrect). It also doesn’t commit to a specific value set, which I think would have been easy to do back then when the pool of people talking about this was very small.