Lincoln's blog

[ Home | RSS 2.0 | ATOM 1.0 ]

Sun, 24 Aug 2014

Mon, 17 Jun 2013

Effective Altruism

Over the last several years, a number of organizations have popped up: GiveWell, 80,000 Hours, The Life You Can Save, Giving What We Can, Centre for Effective Altruism, Future of Humanity Institute, Center for Applied Rationality, Leverage Research, and probably a few I've missed.

The purpose of these organizations is to strategically make the world a better place -- figure out the best interventions, or the best actions to take, in order to reduce the most suffering and allow humanity to reach its greatest potential.

Perhaps surprisingly, not everyone agrees on the best way to achieve this. A quick look through the above websites: donate to proven cost-effective interventions, convince other people to donate a big fraction of their income, worry about existential risks, teach people to think more carefully, develop psychological insights, create friendly artificial intelligence, and so on.

I found this lack of coordination surprising because when people have a common goal, they benefit from working together, and so it seems they would do better if they were able to reach agreement on the best path forward. On the other hand, it makes sense that they have different paths: humans often exhibit flaws in rationality which would prevent such cooperation -- reasons like tribalism, status quo bias, and simple inability to work with people. Also, different people have different skills and comparative advantages, so it might make sense to work on a lot of different things.

Anyway, a lot of the folks in these organizations are convinced that improving their own rationality could have high payoffs in their ability to achieve their goals. Tribalism can also be addressed by getting a lot of people from different organizations in the same room and hoping they make friends with each other. So Leverage Research and CFAR got together and created the Effective Altruism Summit. 50 people -- many from the above organizations, but also a bunch of unaffiliated community members (including me) are going to stay in a big house in California for a week and try to produce a bit more rationality, goal-alignment and friendship amongst the effective altruism community.

I guess we'll see how it goes!

Currently it seems to me that the best way forward is for humanity (or a small group of humans) to become smart and rational enough to develop safe, powerful artificial intelligence. We can't go too slow on this: bad things are happening every day (people are dying, new advances are being made on dangerous technology). But if we go quick and screw it up, we'd be making things worse. I'm writing this down in order to see if my beliefs change over the course of the Summit.

posted at: 23:34 | path: /ea | permanent link to this entry

Tue, 29 Nov 2011

Givewell's 2011 recommendations

Since Rationality Camp this past summer, I have been interested in "optimal philanthropy": figuring out the best ways to do good in the world, which generally means saving lives.

Lots of people want to "do good things", but they don't think carefully or strategically about the best ways to achieve this. They fire and forget: they pick a charity which sounds like it is trying to achieve something positive in the world -- perhaps they put money into the bellringer's bucket at the grocery store, or they buy livestock gifts in response to a brochure. It's absolutely awesome that these people decided to help out. But they could easily increase the impact of their do-gooding by a factor of at least ten, maybe a hundred, by thinking about it a little bit.

Hold on a second, you might be saying. A factor of a hundred? Where's that coming from?


  1. Charities succeed and grow if they can convince people to give them money. They die if they can't. In contrast, there's no external pressure which would cause charities to optimize for effectiveness.
  2. Most charities have not chosen a high-impact cause. Providing toys to hospitalized children is good, but preventing the diseases which got them sick in the first place is much better.
  3. Even if a charity has chosen a high-impact cause, its directorship may not be acting strategically to maximize effectiveness. Most people are not strategic by default. For example, the directors might not realize when they needed to do a study on their interventions to determine effectiveness.
  4. Even if a charity in a high-impact cause has good leadership, public perception or competition could restrict the actions the charity can take. For example, donors often choose charities based on how much money goes directly to program expenses. But often an organization's marginal dollar is better spent on training, evaluation, or fundraising.

So yeah, it seems like most charities are probably going to do a really bad job at actually doing good! But all is not lost. There are a small number of excellent charities which have managed to navigate the above minefield and present provably high-impact interventions. GiveWell exists to perform in-depth reviews on charities, uncover these gems, and make recommendations. Just this week, GiveWell announced a new set of charity recommendations.

So if you're thinking about doing charitable giving this year, consider following GiveWell's recommendations. GiveWell also provides guidelines on evaluating charities yourself, if you prefer to act locally.

If you want further reading, check out:

posted at: 23:23 | path: /ea | permanent link to this entry