Effective altruism is not effective

Effective Altruism is a particularly base version of utilitarianism. It cheerfully accepts situations where you would be required to eat your grandmother or allow your child to starve in order to feed a village of people you've never met in another country. Peter Singer argues that it's okay to be a concentration camp guard because if you refuse, someone else will be found to take your position, and they might be more brutal than you. Do you believe the earth is overpopulated? If so, and if you cannot live in such a way as to offset the damage caused by your existence, effective altruism calls for suicide.

Of course, we can just assume everything goes well, we agitate for political change, we get a change for the better, everything is great.

But the risk of a different outcome never seems to be taken into consideration. For instance:

We believe the conditions in our country are unfair. We start agitating and take down vital institutions in the name of change. Our movement gets hijacked by a military strongman who happens to be charismatic and liked by the people, and now we're under a brutal military dictatorship.

Oops.

History shows the likelihood of the second kind of outcome is, at the very least, non-ignorable.

I would like to see more analysis on what causes bad outcomes of this kind, and to see them taken seriously, instead of this handwaving where poltical action always leads to good outcomes.

What you need to do that is politics.

Except if you look at actual EA organizations they are not “reasoning at the margin” when it comes to global poverty. GiveWell is targeting global health, which can absolutely be solved with donations (vaccines, for instance, don’t require changes to societal structure to eradicate diseases).

Open Philanthropy tackles more structural issues such as criminal justice reform - but they aren’t doing something “marginal” like donating to bail funds. Instead, they’re donating to organizations which build grassroots movements among crime victims and formerly incarcerated people and funding pilot programs for alternatives to prisons.

Optimizing the effectiveness of one's charitable donations is a reasonable goal. Doing basic research to choose efficient and effective charities makes sense. However, there are large parts of the EA community that have gone off the deep end of debating topics like AI Risk. This has given rise to weird situations where EA communities have wasted countless hours debating whether it's better to donate mosquito nets to combat malaria, or to give money to research institutions that are studying the risks of general AI. 2 of the top 3 biggest donations from Open Philanthropy are to study AI risk. The 80,000 hours community has spent years creating podcasts and websites, but it's still difficult to extract actionable information from it. It's all becoming increasingly nebulous, and frustratingly so.