Has EA started its decline?
Just as hurricanes need a steady supply of warm water, groups need a continuous flow of time invested in them to remain strong. People’s time is a limited resource, especially since we still lack a cure for aging. If someone is rich, or famous, or especially skilled, then their time is even more scarce. A strong group, of course, has to convince lots of people to spend time there, rather than on all of the competing alternatives. For a for-profit business, this is simple, at least in theory: you pay people to invest time, which you then turn into money, which you then use to pay for more time, and so on. For a non-profit, or a loosely organized informal group like EA, there’s a lot more variance.
I think that, since 2013 or so, most of the new time invested in EA has come from one particular feedback cycle. There was a single core meme, which was roughly a mix of earlier ideas by Peter Singer, Holden Karnofsky, and Toby Ord, and it went something like: * Some conventional-looking nonprofits, especially in poor countries, can do a tremendous amount of good. Orders of magnitude more than other nonprofits, like preventing death or permanent disability at trivial cost. We can figure out which nonprofits these are by applying some relatively simple, well-known analysis tools, like cost-benefit spreadsheets and controlled scientific experiments. Therefore, average people in developed countries can do huge amounts of good, the type of good that would canonically make you a famous hero, by permanently donating a significant fraction of their salary to these charities. Hence, earning more can make you more moral if you donate a constant fraction.
I’m sure someone can nitpick the above, but I think it’s essentially correct. The content of * was memetically fit, for several reasons: it was on the border between being obvious, and being either heretical or too complex to understand easily; the idea of doing lots of good appeals to people, and strongly to certain subgroups, especially idealistic students; and a few people really disliked it, so the heat of debate added more energy. Hence, * spread exponentially as people heard about it, they were then inspired to write about it, causing more people to hear about it, and so on.
The banner of EA is bigger than *, of course. AI risk, for example, isn’t really part of *. But I think the “other parts” of EA are mostly effects, not causes. That is, the exponential spiral around * grew so quickly that AI risk, veganism, and several other things were partly sucked into it. One could, afterwards, redefine what EA meant — several have tried to — but that doesn’t change the historical fact that * was the group’s main energy source.
I claim that the flow of energy from * is now dying off, for many reasons. Lots of people have concluded that * isn’t fully accurate; conventional nonprofits, for example, can’t really do the amount of good that many expected . Several ideas adjacent to * are thought to be damaging; eg., that one should work at a morally repugnant job to get more donation money. Or that, if one also thinks saving a life and killing are morally equivalent, one would effectively be a murderer for taking an expensive vacation. The object-level successes of * have also been mostly ignored, dampening interest. I’ve never heard anyone else note, for example, that AMF now has the budget and organization to reach a significant fraction of all reachable African families. And any meme tends to burn itself out over time. Eventually, most people in the audience will have heard about it, and then it’s no longer worth writing about.
Since the time going into a group is hard to measure, this is difficult to definitively prove. But there are several lines of evidence I’m using: There don’t seem to be regular stories about EA in mainstream media outlets anymore. I couldn’t find any from the last month; the closest was on the Psychology Today website, but it was in an ancillary blog rather than the main magazine. Google suggested one, a piece in Vice, but it turned out that Google had the date wrong (it was from 2014). EA has been covered a lot before, including by the NYT, WaPo, NPR, ABC, The Atlantic, BBC, WSJ, etc., so I think this represents a decline.
I unfortunately don’t have exact numbers, but I believe that EA Global 2017 wasn’t significantly larger than 2016 or 2015, and many people I know went to 2015 or 16 but not 17. In terms of speaker lineup, EAG has never really topped having Elon Musk in 2015. This year had Joseph Gordon-Levitt, but he’s obviously being chosen purely for fame rather than subject matter experience. As above, press coverage of the conference appeared to peak in 2015, with pieces by HuffPo, Inc and Vox, and little since.
The EA Facebook group, which in five years went from nothing to 13,500 members, is now growing fairly slowly at ~20%/year annualized. In my experience, people rarely leave groups, so this may represent a steady state or even a decline in terms of real readership. There are other EA discussions, of course, most obviously the EA Forum, but I don’t think any of them have close to the Facebook group’s scale.
To my knowledge, local EA groups remain relatively small and disorganized, with the main exceptions being student groups. Where I live in the Bay Area is arguably the global center of EA, and the events here tend to be small and irregular, with the partial exception of student events at Berkeley and Stanford.
There was an EA Forum post a few weeks ago about various growth metrics , with many of the metrics now increasing slowly, at a steady state, or on a slow decline. The main exceptions are EA and 80K newsletter signups, which I believe have been driven by paid promotion, and OpenPhil donations, which of course come from a single “pot” of money rather than a large community.
I’ll make a few guesses as to what happens next. OpenPhil donations should continue to ramp up quickly, as they remain well below target level (5% / year * ~$12B = ~$600M/year). I expect that most people who signed the GWWC pledge will stick to it, at least for the next few years, and so donations from that should also continue increasing as pledgers enter the labor force (as of last year, ~50% were students at time of signup). I think enough energy is still coming from places at least partly in EA, like AI risk, veganism, and the political ideas OpenPhil is exploring, so as to keep conferences and so on in roughly a steady state. Interest in *, already at a fairly low level in “core” EA, will continue declining as the echoes of past writing slowly die down. Since EA skews so young, we may see less strong enthusiasm but also fewer talent shortages as people mature.
Beyond that, as I mentioned in the last thread, I think some other movement will eventually come along and scoop up much of the energy and talent now in EA, as has happened several times before. Such transitions are rarely “complete” — eg., the World Transhumanist Association still nominally exists as Humanity+, despite having few active members and doing very little. I therefore expect EA organizations to keep existing, at least nominally, for quite a long time, possibly decades. The GWWC pledge is supposed to be a lifetime commitment, and enough people take that seriously that I’d expect that to keep going for a while, or else be merged into some larger organization.