Note: This post was written hastily off the dome with barely any editing. I would have just posted it as a quick take except it got too long. Please don’t nitpick specific wording and instead try to respond to the core ideas.
When we talk about EA, we’re often talking about two different things that we should distinguish. On one hand, there’s the pre-2013, global poverty-oriented version of EA. We might call that EA 1.0. On the other hand, there’s the post-2017 version of EA that’s much more oriented around AI, longtermism, and existential risk. We could call that EA 2.0.
An important part of the story is that a cult, Leverage Research, organized the first EA Summit in 2013 and eventually gained full control of the Centre of Effective Altruism in 2018 when one of its members became the CEO. Nobody ever talks about this, but a cult infiltrated and took over EA. That’s a major part of EA’s history and development. It might help explain a lot of what’s wrong with EA today.
So, we could also call the pre-2013 era the pre-cult era of EA and the post-2012 era the cult era of EA. The first EA conferences were organized by a cult. Major EA programs like the Pareto Fellowship — which I applied to! — were run by a cult and in a cult-like fashion. This is complete insanity and I almost never see anybody talk about this. Although Leverage Research has mostly — but not entirely! — been ejected from EA now, other cults and extremist groups are still in EA’s orbit. Including dangerous, violent ones that have killed people and tried to kill people.
Anyway, EA 1.0 still exists within EA 2.0. In absolute size, EA 1.0 might have even grown, although in relative terms it’s gone from 100% of EA to maybe 50% or less (depending how you measure).
Sometimes people want to criticize EA 2.0 specifically, without criticizing EA 1.0. Since we don’t have clear terminology (such as “EA 1.0” vs. “EA 2.0”) to distinguish these two parts of effective altruism, the critics often just say they’re criticizing effective altruism. People who want to defend the distinctive characteristics of EA 2.0 sometimes dodge getting to the crux of the disagreement by invoking the nice qualities of EA 1.0 that the critics often actually like.
For instance, a critic might say EA hasn’t put forward enough high-quality evidence or arguments for the supposedly imminent advent of superintelligence. A defender of EA 2.0 might then find a way to invoke GiveWell in order to defend EA’s intellectual rigour and empiricism.
The critic might say the EA community doesn’t focus enough on input from experts, peer-reviewed research, scientific evidence, etc. The criticism is specifically aimed at EA 2.0. But the defender will then use GiveWell as an example of an organization that gets lots of input from experts, reads a lot of peer-reviewed research in order to make decisions, and prizes randomized controlled trials and data from the field.
In this case, the critic is criticizing EA 2.0 (not EA 1.0) and the defender is defending EA 1.0 (not EA 2.0) and they’re just talking past each other. This is why terminology is helpful. This is why distinctions are helpful.
When you break it down, most critics of EA have a lot to criticize about EA 2.0 but very little, if anything, to criticize about EA 1.0. EA 1.0 had (and has) its critics as well — some make great points about, e.g., the importance of institutions in economic development, others make more dubious points related to Marxism or nationalism or communitarianism — but they are much fewer in number than the critics of EA 2.0.
Also, critics of EA 1.0 more often make criticisms that are fundamentally constructive, e.g., look into funding pro-democracy and anti-corruption interventions in developing countries in addition to global health. By contrast, critics of EA 2.0 more often find very little that is salvageable from EA thinking on AI, longtermism, and existential risk. The advice is commonly: go back to the drawing board, start over from scratch.
I don’t want to speak for anyone else, but my impression is that, for example, frequent EA Forum poster titotal and the philosopher David Thorstad have relatively little (if any) criticism for EA 1.0 and a whole heap of criticism for EA 2.0. Speaking for myself, I’m pro-EA 1.0 and anti-EA 2.0. I’m a fan of EA’s pre-cult era and then things seem to have gotten worse after the cult infiltration began. Critics and defenders of EA 2.0 can stop talking past each other so much if we adopt the EA 1.0 vs. EA 2.0 distinction.
This is particularly important given that EA is not an all or nothing, black or white, binary proposition. People sympathetic to EA can make fine-grained choices about what they support and invest themselves in — and what they don’t. If we’re thinking on the margin, it doesn’t make sense to try to weigh up all the good and bad of EA and see if it’s a net positive. That’s a pointless exercise. We’re not faced with a binary choice. The thing is to think about the next right step. Where should our next incremental unit of resources go? What communities and organizations should we focus our time and energy on? Among all possible actions we could take, which are the best ones? Or at least, good enough?[1]
The point of this post is just to a) make the distinction between EA 1.0 and EA 2.0 and b) encourage people to start using the “EA 1.0” vs. “EA 2.0” terminology. If you want to get into the reasons I’m against EA 2.0, I’ve written extensively about that elsewhere. Even better, you could read the philosopher David Thorstad’s years of thoughtful criticism on EA 2.0. The point here is just to make the distinction, and encourage the terminology.
Also, since the EA Forum is overwhelmingly EA 2.0 (to the point of suppressing anti-EA 2.0 sentiment, including through moderation), I want to encourage anyone with the time and energy to work on alternative platforms for discussion to do so. There should be an EA 1.0 space for discussion. (If you email me or message me on Substack, I might be able to help or connect you with people who can.)
- ^
Incidentally, black and white thinking like this is a frequent trap in political discussions. Is capitalism or mixed market economies good or bad? Is nationalism good or bad? Are regulations good or bad? Is the government good or bad? Is liberal democracy good or bad? Is social justice good or bad? If you get stuck in the binary role of a critic, you’ll often end up advocating worse alternatives to what you’re criticizing, such as Marxism-Leninism, libertarianism, authoritarianism, illiberal populism, etc. If you get stuck in the binary role of a defender, you wind up becoming an apologist for problems and failures that people are rightly angry about. Even if you have good intentions, you don’t come across as a credible reformer who is going to fix things and make them right.
