Third Wave Effective Altruism

I think this is a useful exercise for a few reasons.

1. It's helpful for outsiders and new people to reconcile different material that they see from different points in time.
2. It's helpful for people to clarify if they hold a bundle of positions that's associated with a particular wave. The old waves don't necessarily go away, they just cease to dominate. It's quite handy to be able to say "To understand my positions it may help to know that I'm a first-wave EA" or whatever. 

I would be tempted to divide the second wave in two. I think there was a distinct period where career choice and talent constraints became a dominant theme, but before longtermism took off. So I'd say:

1. As you have it
2. Talent constraints become prominent, 80k becomes prominent, money becomes looser as OPP becomes a major player etc.
3. Longtermism becomes prominent, money becomes silly as FTX enters the scene, WWOTF

Not sure on dates. This gives us a very short third wave, but I think that matches reality: the FTX crisis and AI panic killed/substantially changed it.

Also, I think an interesting row for your table would be "Prominent thinkers". It's interesting that in other movements the waves are typically spearheaded by new people and thinkers, whereas in our case it's often the same people again. Some extremely incomplete lists:

1. Singer, Ord, Karnofsky ...
2. Todd, MacAskill ... 
3. Ord, Cotton-Barratt, MacAskill, Bostrom ...

I also think this can help us with the question of "where is the next wave going?". We can instead ask "who are the thinkers who are gaining prominence?". It seems to me that there's a bit of a void, except in the AIS space, so maybe that will dominate by default.

I agree that we're now in a third wave, but I think this post is missing an essential aspect of the new wave, which is that EA's reputation has taken a massive hit. EA doesn't just have less money because of SBF; it has less trust and prestige, less optimism about becoming a mass movement (or even a mass-elite movement), and fewer potential allies because of SBF, Bostrom's email/apology, and the Time article.

For that reason, I'd put the date of the third wave around the 10th of November 2022, when it became clear that FTX was not only experiencing a "liquidity crisis" but had misled customers, investors, and the EA community and likely committed massive fraud, and when the Future Fund team resigned. The other features of the Third Wave (the additional scandals and the rise in public interest in AI safety due to ChatGPT, GPT-4, the FLI letter, the CAIS statement, and so on) took a few months to emerge, but that week seems like the turning point.

I'm glad Ben shared this post! 

I'm not sure how much I agree with the framework — or at least the idea that we're entering a third wave, but this seems like a useful tool/exercise.

Here's one consideration that comes to mind as I think about whether we're entering a third wave (written quickly, sorry in advance!). 

I've got competing intuitions: 

1. We tend to over-react to (or over-update on) changes that seem really huge but end up not affecting priorities/ the status quo that much. 
   1. E.g. I think some events feel like they'll have a big effect, but they're actually just big in the news or on Twitter for a few weeks, and then everyone goes back to something pretty normal. Or relatedly, when something really bad happens and is covered by the news (e.g. an earthquake, or some form of violence): we might feel pressure to donate to a relevant charity, make a public statement, etc., when actually we should keep working on our mostly unrelated projects.
2. At the same time, I think we tend to under-react and are too slow to make important changes based on things happening in the world. It's too easy to believe that everything is normal (while in reality, futures are wild). We're probably attached to projects (don't want to stare into the abyss) and probably dismiss some ideas/predictions as too weird without giving them enough consideration. 
   1. COVID is probably an important example here (people weren't updating fast enough), and I can think of some other examples from my personal life. 

My best guess (not resilient and pretty vague) is that we're generally too slow to update on in-the-world changes (that aren't about other people's views or the like), and too quick to update on ~memes in our immediate surroundings or our information sources/networks. I tentatively think that (public) opinion does in fact change a lot, but those changes are generally slower, and that we should be cautious about thinking that opinion-like changes are big, since small/local changes can feel huge/permanent/global. 

So: to the extent that the idea that we're entering a third wave is based on the sense that AI safety concerns are going mainstream, I feel very unsure that we're interpreting things correctly. We have decent (and not vibes-based) signals that AI safety is in fact going mainstream, but I'm still pretty unsure if things will go back to ~normal. Of course, other things have also changed; specific influential people seem to have gotten worried, it seems like governments are taking AI (existential) risk seriously, etc. — these seem less(?) likely to revert to normal (although I'm just guessing, again). I imagine that we can look at past case studies of this and get very rough ~base rates, potentially — I'd be very interested. 

(I have some other concerns about using/believing this model, but just wanted to outline one for now.)

I'll also share some notes/comments I added on a slightly earlier draft. I haven't read the comments carefully, so at least some of this is probably redundant.

Some other possible "third waves" (very quick brainstorm)

1. Attempting to stay relevant: AI safety blows up, EA still has a lot of people who have been thinking about AI safety for a long time and feel like they should be contributing, but they don’t catch on to the fact that they’re now 100x smaller fraction of the field, and not the biggest players anymore. (Also seems possible that they're the experts and suddenly have lots of work, but it doesn't seem like a certain thing.)
2. EA grows: AI attention brings a lot of attention to EA somehow, and EA grows a bunch through unusual pathways (unusual for us), everything else is similar (maybe this is the 4th wave somehow — hinges on something that hasn’t happened). The main updates are about the size of the movement/network (what would EA look like if it had 20x more people?), and its composition (later-career folks, etc.)
3. "Effective AIS": Little changes from now from EA’s POV except that AI safety is big outside of EA, but most of that is ~ineffective for one reason or another. At the same time, there’s a fair amount of funding for “effective AI safety” work (possibly something similar to what happens with effective climate work)
   1. I.e. a lot of stuff gets "AI safety" that's not really AI safety (or is just not great). But big donors are interested in AI (existential) safety and there are people in ~EA-adjacent spaces who are attracted to EAxAIS because of competence and reasonableness of arguments; donors are excited about funding this kind of thing. We need to work on making work like this legible. We need a version of Longview/FP but for AI safety. 
   2. Alternatively: AI safety becomes super politicized and people don’t want to work with AI companies, so EAs are the only ones doing that. 
   3. Alternatively: AI safety (in the popular understanding) becomes very strongly about something like copyright issues/bias/unemployment ("we shouldn't be distracted from the real problems today")
   4. Etc.
4. "Back to normal+puddles": Attention on AI safety passes. Things are very similar except there’s a ~quiet and occasionally noisy archipelago of AI-safety-oriented communities/projects (think puddles after a storm). 
   1. Some people think that EA is “the AI safety thing” and confuse EA with that (like they still do with earning to give sometimes).
5. The "third wave" might be prompted by something that isn't AI-related. Some possible scenarios:
   1. Something potentially FTX-related leads to the EA brand becoming toxic. 
   2. The EA network sees a schism along something like GHD vs. non-GHD, “longtermism” vs not, "weird" vs not, etc. 
      1. Alternatively, there's an overall fracturing into loosely-grouped and loosely-networked focus areas, like effective GHD, effective FAW, WAW, maybe pandemic preparedness, AI safety, AI governance, ~cause prioritization research, etc. Some organizations and groups focus on letting donors evaluate projects across a wide space given their priorities and philosophies (or giving career advice).
   3. EA has just grown too big to be useful to coordinate around, and we’re seeing what looks like the beginnings of a ~healthy fracturing (which in reality might be past the point of no return); we’re shifting to a model where there are cause-specific communities that are friendly to each other, and some orgs work across them and keep an eye on them, etc. 
6. Stuff that might happen that could change things fast
   1. Big war
   2. Big politicization moment of AI, or AI safety becomes very strongly about something like copyright issues/bias/unemployment
   3. Really scary AI thing that makes people really freaked out
   4. Something weird happens with labs (E.g. government does something strange) and they become super uncooperative?
   5. New pandemic
   6. Significantly more bad press about EA
   7. Huge endorsement of EA somehow / viral moment
   8. ~research becomes automated
   9. Etc.

I see a world that still desperately needs Wave 1, and I see a lot of work still to be done in that area. 

I look at the effectivealtriusm.org homepage, and a lot of what is mentioned there is still what you're referring to as wave 1. 

I would even venture that in most of the world (perhaps outside the hubs), people are drawn to EA first by Wave 1 concepts. We get frustrated at the poverty and disease and war and poor governance and refugee crises and famines and ... and we wonder why can't we do more to fix these with the significant resources we do devote to them. We see a group like EA looking at how to use limited resources to help people in the most effective way possible and it seems like a critical answer to a long-neglected question. 

Is it possible that what you're describing here is the cutting-edge aspects of EA - the areas where EA is breaking new ground philosophically and analytically, the areas which create lively, passionate debates on this forum, for example? And so, naturally, the ideas for the future come from areas like AI and longtermism. But a lot of vital EA work doesn't have to be cutting-edge research. 

But IMHO there is still a massive opportunity to help most of the world's population in very concrete, tangible ways, and effective altruists can make vital contributions. You write the the goal of wave 1 was "donations to effective charities" - but this is a quite limited reading of what EA can do. How about influencing how governments spend their aid budgets, which is often very differently from how they would be most effective? There are a few groups doing this kind of work (e.g. Gates Foundation), but there is still so much aid and donations being inefficiently spent. Ideas as simple as how to convince governments to just give people in developing countries cash rather than spending 10X that much trying to solve their problems for them. 

I know that part of the vision is to focus on areas which are neglected, but I see a big difference between working in an area that is neglected (which is not true of global poverty and disease, for example) and working in a way that is neglected (quantitative, analytical, data-driven) even in areas which receive a lot of attention and even (badly-spent) money. 

Apologies if this feels ill-informed. I'm writing as someone who isn't in any of the hubs and so just seeing EA from the "outside." 

It seems that one assumption underlying this frame is that EA will largely continue as a unified enterprise, perhaps with some cause areas "spinning off" to mainstream. 

I think the possibility of breakup is real enough to include as a possibility in one's model, at least as a footnote. That could be a collaborative divorce, or a messy one:

• An example of a collaborative divorce would be a mix between the second and third possible future scenarios. "Forefront of weirdness" and "return to first wave" have some real tensions, and it's plausible that a significantly greater degree of separation would help each group better achieve its own goals.
• A messy divorce could be caused by (e.g.) a further loss of confidence in centralized institutions due to FTX fallout, or worsening reports of sexual assault in some subcommunities, leading a significant fraction of people to break off for various reasons.

Substantially less money, through a combination of Meta stock falling, FTX collapsing, and general market/crypto downturns^[3]...[3] Although Meta stock is back up since I first wrote this; I would be appreciative if someone could do an update on EA funding

Looking at this table, I expect the non-FTX total is about the same^[1]—I'd wager that there is more funding commited now than during the first ~70% of the second wave period.^[2]

I think most people have yet to grasp the extent to which markets have bounced back:

• The S&P 500 Total Return Index is within 6% of its all-time high; only ever spent ~4 months above today's value
• META is -26% from ATH, but now highest since Jan '22 and ~10 months ever above this price

1. ^{^}

   Dustin's net worth looks to be about about -$7Bn from the peak (per his preferred source). Meanwhile, GiveWell (2-3x), Founders Pledge (3x), and GWWC (1.5x) numbers all seem to be higher

2. ^{^}

   I still think that the waves framing is useful and captures the prevailing narrative tbc

Fwiw, I think this kind of thinking is good but shows a deep need for synthesis. There is so much to read on the forum and we are poor at succinctly nailing down exactly what we thinking on areas where we basically agree or there is little disagreement.

As an example, most people do not have detailed maps of the sea but on our maps we agree where it is. This is because the exact location of the sea is not important or widely agreed. If every person making a map of their town needed to draw all of the sea, that would be a big waste of time.

Similarly, due to the main form of community knowledge transfer being blogs, there is little "things we all agree on", so many blogs end up covering ground covered in other blogs, because the authors or readers haven't read them (because there is too much to read).

The solution might seem to be more overviews from elites, but while I think elites make better choices on average I think that knowledge synthesis requires back and forth. I am much more likely to engage with what "we as a community think" if I feel like if a large chunk of the community disagreed then it would change. 

Here, for instance, I think the push for longtermism was shorter than being the defining cause of this period.

To put it more concretely I think what the community should be is a job for elites, but what it is and currently believes is a job for some other process. I think we do not have that process and so everyone has to read way too much and write blogs that cover old ground. We have discussions many times without making ground because we can't focus on actual areas of disagreement.

This post is a partial solution - giving clear models of who we were and are - humans need narratives to move forward. But this process seems underrated - if all you have is the agreement of what the second wave was, then there is a lot of work needed to be done by current EAs to understand what typical EAs like themselves think and do. As humans that's a pretty good sign of what is safe to do. If we had a way to agree this information quicker, I would have more brain space for work, actual substantive disagreements etc and more trust that the community could come to concensus here.

Seems like underrated work, thanks for doing it.

One observation is that the first transition from 1st to 2nd wave was deliberate in that it was after a strategic review conducted by CEA, whilst the second transition was imposed by events. Perhaps the consequence of the first transition also has a influence (not sure how strong) on the trajectory of the second transition which is still unfolding.

Thanks for writing this up. I've often thought about EA in terms of waves (borrowing the idea from feminist theory) but never put fingers to keyboard. It'shard to do, because there is so much vagueness and so many currents and undercurrents happening. Some bits that seem missing:

You can identify waves within causes areas as well as between cause areas. Within 'future people', it seemed to go from X-risks to 'broad longtermism' (and I guess it's now going back to a focus on AI). Within animals, it started with factory-farmed land animals, and now seems to include invertebrates and wild-animals. Within 'present people', it was objective wellbeing - poverty and physical health - and now is (I think and hope) shifting to subjective wellbeing. (I certainly see HLI's work as being part of 2nd or 3rd wave EA).

Another trend is that EA initially seemed to be more pluralistic about what the top cause was ("EA as a question"), and then became more monistic with a push towards longtermism ("EA as an answer"). I'm not what that next stage is.

I like this framing, and here's some thoughts on movements from a movement veteran. First, it's obvious EA moved from raising money for effective charities focus to longtermism/X-risk and it interesting to see all the cultural flows of that in EA. And then it seems fairly obvious that the series of scandals from FTX to sexual harassment has had a reverberating shock wave affect in EA and that to me is the clearest sign EA is primed for a new wave, the third. I would date it not only by AI but also by some date averaging of these big difficult stories from Dec. 2022 to early 2023 where FTX was still everywhere while ChatGPT debuted globally and the sex stuff. The fact that EA is years ahead of any other organized movement in organizing on AI Safety means it can be a hub and that has value moving forward. 

I think the really big question is this - what will the effect of all the trauma and embarrassment and people reassessing themselves and EA as a whole end up producing on the steering rudder of EA...where will it turn, what ways will it change? Comments on that would be very interesting. 

Here's my comment, which if you were to read all my posts and comments you could see the trend: I don't know at all what direction the rudder will steer us toward, but I hope that it includes a huge cultural reformation surrounding Utilitarianism. One of the iconic quotes that would give me this idea is of an interview with SBF where he says he's a Benthamite Utilitarian very soon before he's revealed to be a historically awful fraudster who was spawned and enabled by a bunch of Benthamite Utilitarians calling themselves Effective Altruists.  

Now I know some leading lights have spoken out recently that, "Aw we haven't really been hard core Benthamites...we've always been more balanced". I would say that's classic blind spot talk because EA you have no idea how strongly Utilitarian you come across to anyone from the outside...you are not at all balanced, you are hard core Utilitarian...if you think you're balanced you're just too inside to know how things look from the outside. I think what's happening mentally is you are smart enough to imagine the freakish' crazy side of extreme Utilitarianism and you know you aren't that...but that's because nobody is that freakish' except literally psychopath outliers who don't count, but instead you are still firmly situated in a kind of Utilitarianism which though balanced with some common sense which is both socially and literally unavoidable, is still very far over from most of your peers in non EA culture worlds. 

I can imagine all the defensive comments saying it's not that bad, we're more balanced, but as I said above that's just being too inside to see from the outside - if there was any one major cultural thing that typified EA and EA people it would be utilitarianism...eating Huel alone at your desk so you can grind on to be more effective is to me the iconic image of that. 

I know this is tough love, but I do dearly love EA...and I just want all to be happy and stop eating Huel alone at your desk and discover the joy of being with others and having an ice cream cone now and then. You'll be far more optimized by that to do your good work. Utilitarianism is optimizing for robots not for humans. Effective Altruists are humans helping humanity...optimize for humanness.

If EA decided to pursue the politics and civil society route, I would suggest that it would likely make sense to follow a strategy similar to what the Good Ancestors Project has been following in Australia. This project has done a combination of a) outreach to policy-makers b) co-ordinating an open letter to the government c) making a formal submission to a government inquiry d) walking EA's through the process of making their own submissions (you'd have to check with Greg to see if he still thinks all of these activities are worthwhile).

Even though AI Policy seems like the highest priority at the moment, there are benefits of working on multiple cause areas since a) you can only submit to an inquiry when one is happening, so more cause areas increases the chance that there is something relevant b) there's a nice synergy that comes from getting EA's who have different cause areas as their main focus to submit to the inquiries for other areas.

Greg has a great explanation where he talks about EA having spent a lot of effort figuring out how to leverage our financial capital and our career capital to make the world better, but that we've been neglecting our political capital. Obviously there's the question of whether we have good ways to deploy that capital, but I suspect that this answer is that we do.

I'm not claiming that this is necessarily the route forward, but it is likely worth exploring in countries with well-developed EA communities.

Thanks for the model - I think it's useful. 
I think it'd probably be more appropriate to say that wave 2 was x-risk (and not broad longtermism) and/or that longtermism became x-risk. Before reading your thoughts on the possibilities for the third wave, I spent a few seconds developing my thoughts. The thoughts were:
1. Target audience: More focus on Global South/LMIC.
2. Culture: Diversification and more ways of living (e.g., the proportion of Huel drinkers go down).
3. Call-to-action: A higher level community/set of ideas (e.g., distilling and formalizing the method of EA and/or longtermism) and cause-specific community (AI safety, etc.).
4. Other: EA as a label gets reduced (e.g., if CEA changes its name).

Substantially less money, through a combination of Meta stock falling [...] 

I have also being talking about META stock falling. But when I looked it up recently, I noticed META is close to an all time high. 

Thank you for writing this!
I just wanted to flag that this format could fixate us on the structure of our past strategy.

For instance (and this is just one example out of many), I believe that the past strategy of the movement was inherently incoherent; 
From the way you describe the second wave, it's clear that the movement was focused on "career changes" of "talent" into "longtermism". However, the movement didn't describe itself as a longtermism hiring agency, it described itself as a community of people who seek the most impactful courses of action. This post and comment describe this criticism in length. 

If this criticism makes sense, then we might consider a third wave that looks something like:

• We've split the brand of EA from a few cause-specific brands that focus on the most effective interventions in their area (e.g., effective AI safety, effective climate change, and so on).
• The new EA brand focuses on the education of EA tools and principles (and maybe also refers people to the resources of the other cause-specific community).
• Each other cause-specific brand has its own call-to-actions, focus audiences, and so on. 
  For instance, AI safety's call-to-action could be research, animal advocacy's call-to-action could be donations, and so on. 

I think that this format of waves is great for brainstorming, and I'm very happy that auch brainstorming is happening. However, this is just one example of why our former strategy might have been suboptimal, and how this format could fixate us on similar directions. 

Great post Ben, and I think the idea of 'EA waves' is a useful framing even if not ~100% historically accurate

Object-level answer: I have a lot of sympathy with Zoe Cremer's idea that the next phase of EA should be to embrace an 'institutional turn', both in our own institutions but also regarding how we approach being effective at our other cause areas. However, as IIDM is probably the area I'm most interested in, take this with a large degree of bias and discount accordingly!! I would still suggest Forum readers check out sources that Cremer highlights as promising - e.g. the work of Audrey Tang in Taiwan, or Helene Landemore's research agenda at Yale.

Meta-level question: I think it's interesting to me that this frame is easily accepted in terms of the first 'bednet' wave being replaced by the second 'AI/longtermism' wave. I agree that there has been a change here, but I think this change may have been reified somewhat. To what extent was early EA eden before the fall compared to EA now? Surely some of that change is honestly people changing their minds? Furthermore, a lot (most?) EA funding still goes to GH&D, we're still all about bednets! (I know you talk about 'flagship' cause areas in this post, but I often see people push this point to its biggest extreme in discussion, but maybe I'm overreacting here)

Conditioned on this frame being roughly correct, it raises the question of whether to expect a fourth wave at some point -- and what the strategic implications of predicting a wave roughly every seven years or so (albeit with low confidence in the frequency estimate) might be.

To address that, I propose a (literal) toy model of EA as a board game. I have Settlers of Catan in mind, but other games may work. The game has the following characteristics:

• Resource allocation is an important part of the game. Your resources are wood, brick, sheep, wheat, and ore (money, talent, new ideas, public support, organizational capacity?). You can invest in facilities that will produce a certain type of resource, but the lead time before those facilities start producing can be considerable. There are complex rules by which you can convert existing resources into others, but it's often not very efficient and usually takes a a turn (year) or two.
• About every seven turns, the game enters a new phase in which the objectives and scoring rules of the game, as well as some game mechanics, significantly change. Some resources may become easier or harder to acquire, and some may become much more or less important to scoring well.
• The person running the game will tell you when a new phase has started . . . but may not tell you for a few turns more what the new objectives/scoring rules are. You can make educated guesses about what new objectives and scoring rules might now be in play, but cannot have high confidence in your guesses.
• The consequences of being short on a specific resource will depend on the specific rule changes. Resources may be complements (e.g., you need grain and ore to build a city), co-factors (e.g., you need a little sheep to build a port), or have other interactions. 
  • But in many cases the resource:impact relationship will not be linear. For instance, it's really painful to not have any wood or brick production in the standard rules of Catan; you get a lot of utility out of the first marginal units of production.

In this model, a rational player will significantly consider the universe of plausible new phases when making decisions about current resource stocks and production. In particular, the player will likely devote some attention to procuring a minimum level of supply for each resource that may be critical under a plausible rule change. 

Coming back to the real world, the toy model suggests that once the contours of the third wave are better defined, it might be better not to focus EA "resource production" as heavily on its particular needs as was perhaps done at certain points in the second wave. Rather, ensuring a basic supply of potentially critical resources for the fourth wave would also be an important goal.