I appreciate that this comes from a good place, but it seems like a strawman argument and a bit misleading.
For one, EA hasn't been around for 35 years, making your title quite misleading, and you when you do acknowledge the work that EA organisations have done to accelerate vaccine development and approval its tucked in a small footnote at the end rather than in the body of the article.
This argument, that systems work is important, also isn't new to the community and has been engaged on extensively. But the arguments, both for and against, aren't included in this article whatsoever.
To be honest, this feels like you're both aiming for engagement and pushing abundance as a new cause area (all the while bundling vaccine approval advocacy with housing reform in the most wealthy country in the world), rather than providing good-faith feedback to the EA community.
I really appreciate this feedback, thanks Andy! I've now clarified the footnote - the reason it's a footnote is because the EA efforts concentrated on accelerating the *second* malaria vaccine (not the 35-year one, the first one) and post-*approval* deployment. Also, the efforts were still aiming to accelerate parts of the process, rather than changing the process itself. I've clarified the footnote now! :)
Definitely EA hasn't been around for 35 years. I wrote it that way because 1. Since the vaccine was considered safe since the 1980s, EA had a lot of time to flag it. I meant to highlight just how long the vaccine's been around, though maybe the bigger reason is: 2. I aim to follow Inkhaven/Scott Alexander's advice to not write for the harshest critics because then you'll end up caveating everything, and become less clear + harder to read. This is how I have and do say it in person, though I realize I'm often saying it to friends who know I adore EA.
Perhaps the statement comes off as misleading if I'm seeming anti-EA or if I'm talking to anyone but the EA community, though neither is true. I intentionally put EA in the title so that my readers who aren't following EA will skip it. I didn't explain why "neglected tractable and important" are criteria, or why AI risk is relevant. This piece would be confusing to folks not following EA, and I accepted that. I've been with EA for a while, I've started an EA city group. This is geared for people engaged enough to join a group like mine.
Also, I agree the systems change critique wasn't engaged on extensively - I didn't mean to introduce it as new! Becuase it's old, I didn't spend much time explaining that criticism - but I should have caveated that the criticism is well-known in EA. I figured everyone's read the EA forum posts, and knows that's every new EA or former EA's top 10 critique. Maybe I should have caveated that for people newer to EA? I really appreciate this feedback, to be clear! :)
Lastly, I'm not aiming the piece to be "feedback" to the EA community, as much as explaining why I chose to pivot to an area that I consider meets the EA bar but that others don't see as "really EA" - even though CG funds Abundance stuff. This post explains why Abundance connects to EA prioritization for me. Relatedly, I've heard some people regret that EA focused so little on politics because Trump's election derailed a lot of EA goals. I know abundance issues' sounds far-fetched together (this is probably what EA sounds like to people outside of EA, animals + AI?), but it connects because better housing policy probably would have meant Trump didn't get elected, meaning AI / democracy / great power competition efforts are better off.
Again, I really appreciate it Andy! This helps me a lot. :)
The article linked to support the claim there has "been a malaria vaccine since the 1980s" says that the protein used by the current malaria vaccine was identified in 1980 and had its gene sequenced in 1984. The trial showing the efficacy of an actual vaccine based on that protein (albeit only 34% efficacy, which is abysmal) took place in 1998 and was published in 2001. We accept the low efficacy of RTS,S now because it's all we've got after decades of trying and developing vaccines against parasites is hard, but I would image for a long time there was hope that a better vaccine could be developed. RTS,S also requires 4 does and needs a cold chain which kind of sucks if you want to roll it out across malaria-endemic regions. Bednets are a more straightforward solution to implement. Obviously, in an ideal world, both are provided to as many people as possible but everything has an opportunity cost.
Thank you Jacob! I really appreciate it. I was mostly going by GAVI the global vaccine alliance starting the clock from late 1980s to 2021, from when the US government requested a vaccine after the protein was identified in 1984 as you stated. Their headline is 35 years, though I can see how it would technically be 1998-2021 so 26 years. The 1984-1998 wait is still interesting because that's when it became clear there wasn't much urgency to get those trials. I really appreciate you pointing this out! https://www.gavi.org/vaccineswork/it-took-35-years-get-malaria-vaccine-why
Totally agree that bednets are straightforward - ideally, we'd have had both much quicker deployed in all malaria zones. :)
I massively agree that there should be more focus on growth and innovation (inside and outside of EA). But the post misses that CoGi (formerly OP) have been funding vaccine research, metascience, YIMBY and other related areas since 2014.
It's a shame that a lot of EA discussion and ideas were still focused on RCT backed health interventions with very little space for growth until more recently but I'm glad that the largest funder was taking it seriously early on.
Yes, absolutely!! CG has always been ahead of the curve. This post is more geared at EA circles that are still warming up to abundance. I’ve heard some people doubt whether abundance is really “EA” even though CG is funding it.
I think that question is a category error, and your post seems to implicitly argue that abundance is effective in EA lights? Because abundance will involve a lot of area- and policy-specific work and not all areas and policies have high altruistic effects! If you care about AI risk, don't work on US housing policy...
Thanks Siebe - I do aim to argue that if we care about AI risk, someone should work on US housing policy because it seems to influence public sentiment toward democracy/public trust a lot. If all efforts toward AI are focused on AI, those will be hampered by weaker/populist/unstable government leaders. When using a more systems-view, abundance helps other cause areas. I'm not sure if I'm understanding your question.
Yeah so I think this reasoning is mistaken, because the choice isn't over what everyone does, but what the few people that are on the fence do. And generally, if you believe that one cause is far more important than another, it's most effective to work directly on it unless something indirectly is VERY neglected! But housing abundance isn't that neglected at all! Sure, more people should work on it, but people worried about AI risk should probably work directly on AI risk.
Thanks Siebe. Though I'm not presenting housing policy as the thing people should focus on. I'm presenting "the abundance approach" which focuses on responsive / better governance e.g. in science, things usually ignored that matter for how people perceive demoracy. To choose only housing of my list feels like a cherry-pick; I spent 1000+ words listing that abundance is an approach, with housing used as one of many examples, because all these examples improve / maintain democracy. In so far as democracy, is important for AI risk and enables protection from gradual disempowerment / responsiveness to loss of control / etc, then it's going to be important that some people work on it even if it's not directly AI. It enables everything else.
What I'm arguing is that "responsive governance / maintenance of democracies " is neglected.
Thanks for engaging Abi :) I'm just having a hard time (maybe it's the brain fog) narrowing down which claims you're making. I care a lot about abundance and about EA, and I believe some abundance work is comparatively cost-effective compared to some EA work. But some of your examples feel underargued, like the statement about housing and AI risk we discussed. I'll try to clarify. Could you let me know which ones of these points you're making?
A. 'it could be very effective to do Abundance community building from EAs lights, look at all these effects!' - then I agree.
B. 'the abundance way of looking at things has some things to offer to EAs that they should adopt' - then I also agree
C. 'All these abundance initiatives are competitive with EA initiatives' - then I'm skeptical
D. "responsive governance / maintenance of democracies is neglected." - then I'm on the fence, because while I agree that this would be a multiplying factor for EA causes, it doesn't feel like a single lever to push on.
Any way this is becoming a long back and forth and you may have better things to do! Don't feel obliged to reply and good luck with the talk!
About a decade ago I needed to get a special permit to put a moving container in front of my house, so I went to the city department to do that. The guy ahead of me was in line to close a lane on one of the largest streets in the city so he could put up a crane to do repair work on a skyscraper. The person processing our paperwork used the same bureaucratic flow. At the time, you could only file online if you were a contractor applying for more then x number of permits a year.
I wanted to run for city council on the platform that we needed to reform bureaucratic processes to make sense, but that was clearly not a winning platform.
I will also say that over the last decade, this bureaucratic morass has been fixed, but the broader point, that basic technocratic competence isn’t how we choose elected officials.
I'll just note that GiveWell has written about their decision making around bednets and vaccine deployment, and it's not clear to me that's the best example of your point. But of course one can't argue with the potential of systemic change, as you call out.
Thanks Matt! I did check that out and was thinking of citing it. Their article focused on deployment after approval, not approval acceleration. Perhaps there’s another one you mean? :)
No I meant their post-2021 work. I guess I spot them their shorter term outlook as OP/CG was explicitly founded as the partner organization to look more speculatively and long-term.
Why clothing manufacturing has knock-on effects: factories need reliable energy to function. Accessesible energy for factories can lead to accessible energy for the community, and for entrepreneurs seeking to diversify away from the garment sector.
Great insights as always. I think there is also a (necessary) opportunity to build an optimistic media movement for progress & abundance. There are some encouraging developments but we need so much more!
I wholeheartedly agree on importance and neglectedness. Our biggest challenge will be tractability. Winning the needed political victories at scale will require pitching all this in ways that resonate with the public mood, and navigating or weakening the political power centers (ex: labor, big business, rent seekers) that have bent our government into such hideous shape to begin with.
Right now, I worry those requirements are in tension with the AI optimism that abundance thinkers are drawn to. It seems like a comparable "blind spot on systems change" as the one you describe. Identifying policies causing bad outcomes is only the first step to changing a broken system, and I think later steps will require rallying coalitions around a simpler, angrier story.
The presentation here could do with a bit of background. WHO was focused on bednets long before Effective Altruism was even heard of. The big dispute in the early 2000s was between advocates of bednets and proponents of DDT. Some DDT advocates were honest, but the most prominent were secretly financed by the tobacco industry, which wanted to discredit WHO campaigns against smoking.
Slow progress on vaccines is a largely separate issue, reflecting things like the lack of commercial interest in a once-off treatment that only benefits poor people. An abundance agenda means more public funding for essential research, not a hands-off approach by government.
Also relevantly for current debates about abundance it would be easy to reframe the npw-refuted arguments of DDT advocates in terms of an obvious solution, obstructed by environmentalist quibblee, bureaucratic inertia etc. In reality, the environmentalists were right all along, and the DDT advocates have moved on to other forms of science denialism
Here's a piece Tim Lambert and I wrote in 2008, by which time bednets had largely won the day.
This is super fascinating, John. Thanks for flagging! I only learned a couple of days ago that DDT was once used as the malaria effort. A really fascinating post! That's awful how those companies tried to discredit DDT harm evidence.
I'm curious whether you see something similar here with Golden Rice. Is that an example where environmental groups aren't following the data?
Also, yes, the Abundance Agenda includes public funding for central research, but the Abundance approach would mean that the government is noticing the trade-offs that it's forcing by its risk aversion. The answer might be that it should be risk averse, but at least to have calculated those lives lost. Similar to how the Abundance book and Institute for Progress are focusing on how federal grants in the U.S. are decided by consensus, which means that moonshots won't happen. There needs to be some way to decide grants, let's trials others because upsides with this one are less likely to happen.
I do think, by proposing things like an examination of the system at large and also things like advanced market commitments, are Abundance, though perhaps it's in the Ezra Klein-Derek Thompson style Abundance. I see Abundance more as improving governance in things with big impacts, rather than being handsoff necessarily. WHO could have allowed rollout with real-world evidence collection, for example.
I appreciate that this comes from a good place, but it seems like a strawman argument and a bit misleading.
For one, EA hasn't been around for 35 years, making your title quite misleading, and you when you do acknowledge the work that EA organisations have done to accelerate vaccine development and approval its tucked in a small footnote at the end rather than in the body of the article.
This argument, that systems work is important, also isn't new to the community and has been engaged on extensively. But the arguments, both for and against, aren't included in this article whatsoever.
To be honest, this feels like you're both aiming for engagement and pushing abundance as a new cause area (all the while bundling vaccine approval advocacy with housing reform in the most wealthy country in the world), rather than providing good-faith feedback to the EA community.
I really appreciate this feedback, thanks Andy! I've now clarified the footnote - the reason it's a footnote is because the EA efforts concentrated on accelerating the *second* malaria vaccine (not the 35-year one, the first one) and post-*approval* deployment. Also, the efforts were still aiming to accelerate parts of the process, rather than changing the process itself. I've clarified the footnote now! :)
Definitely EA hasn't been around for 35 years. I wrote it that way because 1. Since the vaccine was considered safe since the 1980s, EA had a lot of time to flag it. I meant to highlight just how long the vaccine's been around, though maybe the bigger reason is: 2. I aim to follow Inkhaven/Scott Alexander's advice to not write for the harshest critics because then you'll end up caveating everything, and become less clear + harder to read. This is how I have and do say it in person, though I realize I'm often saying it to friends who know I adore EA.
Perhaps the statement comes off as misleading if I'm seeming anti-EA or if I'm talking to anyone but the EA community, though neither is true. I intentionally put EA in the title so that my readers who aren't following EA will skip it. I didn't explain why "neglected tractable and important" are criteria, or why AI risk is relevant. This piece would be confusing to folks not following EA, and I accepted that. I've been with EA for a while, I've started an EA city group. This is geared for people engaged enough to join a group like mine.
Also, I agree the systems change critique wasn't engaged on extensively - I didn't mean to introduce it as new! Becuase it's old, I didn't spend much time explaining that criticism - but I should have caveated that the criticism is well-known in EA. I figured everyone's read the EA forum posts, and knows that's every new EA or former EA's top 10 critique. Maybe I should have caveated that for people newer to EA? I really appreciate this feedback, to be clear! :)
Lastly, I'm not aiming the piece to be "feedback" to the EA community, as much as explaining why I chose to pivot to an area that I consider meets the EA bar but that others don't see as "really EA" - even though CG funds Abundance stuff. This post explains why Abundance connects to EA prioritization for me. Relatedly, I've heard some people regret that EA focused so little on politics because Trump's election derailed a lot of EA goals. I know abundance issues' sounds far-fetched together (this is probably what EA sounds like to people outside of EA, animals + AI?), but it connects because better housing policy probably would have meant Trump didn't get elected, meaning AI / democracy / great power competition efforts are better off.
Again, I really appreciate it Andy! This helps me a lot. :)
The article linked to support the claim there has "been a malaria vaccine since the 1980s" says that the protein used by the current malaria vaccine was identified in 1980 and had its gene sequenced in 1984. The trial showing the efficacy of an actual vaccine based on that protein (albeit only 34% efficacy, which is abysmal) took place in 1998 and was published in 2001. We accept the low efficacy of RTS,S now because it's all we've got after decades of trying and developing vaccines against parasites is hard, but I would image for a long time there was hope that a better vaccine could be developed. RTS,S also requires 4 does and needs a cold chain which kind of sucks if you want to roll it out across malaria-endemic regions. Bednets are a more straightforward solution to implement. Obviously, in an ideal world, both are provided to as many people as possible but everything has an opportunity cost.
Thank you Jacob! I really appreciate it. I was mostly going by GAVI the global vaccine alliance starting the clock from late 1980s to 2021, from when the US government requested a vaccine after the protein was identified in 1984 as you stated. Their headline is 35 years, though I can see how it would technically be 1998-2021 so 26 years. The 1984-1998 wait is still interesting because that's when it became clear there wasn't much urgency to get those trials. I really appreciate you pointing this out! https://www.gavi.org/vaccineswork/it-took-35-years-get-malaria-vaccine-why
Totally agree that bednets are straightforward - ideally, we'd have had both much quicker deployed in all malaria zones. :)
I massively agree that there should be more focus on growth and innovation (inside and outside of EA). But the post misses that CoGi (formerly OP) have been funding vaccine research, metascience, YIMBY and other related areas since 2014.
It's a shame that a lot of EA discussion and ideas were still focused on RCT backed health interventions with very little space for growth until more recently but I'm glad that the largest funder was taking it seriously early on.
https://coefficientgiving.org/research/a-quantitative-approach-to-vaccine-funding/
https://coefficientgiving.org/funds/abundance-and-growth/?grant_year=2016#featured-grants
Yes, absolutely!! CG has always been ahead of the curve. This post is more geared at EA circles that are still warming up to abundance. I’ve heard some people doubt whether abundance is really “EA” even though CG is funding it.
I think that question is a category error, and your post seems to implicitly argue that abundance is effective in EA lights? Because abundance will involve a lot of area- and policy-specific work and not all areas and policies have high altruistic effects! If you care about AI risk, don't work on US housing policy...
Thanks Siebe - I do aim to argue that if we care about AI risk, someone should work on US housing policy because it seems to influence public sentiment toward democracy/public trust a lot. If all efforts toward AI are focused on AI, those will be hampered by weaker/populist/unstable government leaders. When using a more systems-view, abundance helps other cause areas. I'm not sure if I'm understanding your question.
Yeah so I think this reasoning is mistaken, because the choice isn't over what everyone does, but what the few people that are on the fence do. And generally, if you believe that one cause is far more important than another, it's most effective to work directly on it unless something indirectly is VERY neglected! But housing abundance isn't that neglected at all! Sure, more people should work on it, but people worried about AI risk should probably work directly on AI risk.
https://www.effectivealtruism.org/marginal-thinking
Thanks Siebe. Though I'm not presenting housing policy as the thing people should focus on. I'm presenting "the abundance approach" which focuses on responsive / better governance e.g. in science, things usually ignored that matter for how people perceive demoracy. To choose only housing of my list feels like a cherry-pick; I spent 1000+ words listing that abundance is an approach, with housing used as one of many examples, because all these examples improve / maintain democracy. In so far as democracy, is important for AI risk and enables protection from gradual disempowerment / responsiveness to loss of control / etc, then it's going to be important that some people work on it even if it's not directly AI. It enables everything else.
What I'm arguing is that "responsive governance / maintenance of democracies " is neglected.
Thanks for engaging Abi :) I'm just having a hard time (maybe it's the brain fog) narrowing down which claims you're making. I care a lot about abundance and about EA, and I believe some abundance work is comparatively cost-effective compared to some EA work. But some of your examples feel underargued, like the statement about housing and AI risk we discussed. I'll try to clarify. Could you let me know which ones of these points you're making?
A. 'it could be very effective to do Abundance community building from EAs lights, look at all these effects!' - then I agree.
B. 'the abundance way of looking at things has some things to offer to EAs that they should adopt' - then I also agree
C. 'All these abundance initiatives are competitive with EA initiatives' - then I'm skeptical
D. "responsive governance / maintenance of democracies is neglected." - then I'm on the fence, because while I agree that this would be a multiplying factor for EA causes, it doesn't feel like a single lever to push on.
Any way this is becoming a long back and forth and you may have better things to do! Don't feel obliged to reply and good luck with the talk!
About a decade ago I needed to get a special permit to put a moving container in front of my house, so I went to the city department to do that. The guy ahead of me was in line to close a lane on one of the largest streets in the city so he could put up a crane to do repair work on a skyscraper. The person processing our paperwork used the same bureaucratic flow. At the time, you could only file online if you were a contractor applying for more then x number of permits a year.
I wanted to run for city council on the platform that we needed to reform bureaucratic processes to make sense, but that was clearly not a winning platform.
I will also say that over the last decade, this bureaucratic morass has been fixed, but the broader point, that basic technocratic competence isn’t how we choose elected officials.
That's great news that the bureaucracy has been fixed. Can I ask where that was?
Totally agree that technical competence isn't what wins elections, but it helps with being an effective person in office. :)
This was in Seattle.
This is an absolutely fantastic piece. Thank you for putting it together!
This means a lot - thanks Matt! :)
I'll just note that GiveWell has written about their decision making around bednets and vaccine deployment, and it's not clear to me that's the best example of your point. But of course one can't argue with the potential of systemic change, as you call out.
Thanks Matt! I did check that out and was thinking of citing it. Their article focused on deployment after approval, not approval acceleration. Perhaps there’s another one you mean? :)
No I meant their post-2021 work. I guess I spot them their shorter term outlook as OP/CG was explicitly founded as the partner organization to look more speculatively and long-term.
Why clothing manufacturing has knock-on effects: factories need reliable energy to function. Accessesible energy for factories can lead to accessible energy for the community, and for entrepreneurs seeking to diversify away from the garment sector.
Thanks J! Great example of knock on effects.
Great insights as always. I think there is also a (necessary) opportunity to build an optimistic media movement for progress & abundance. There are some encouraging developments but we need so much more!
Positive Sum TO THE MOON
1000% Thanks so much Andy! :) I really appreciate it!
I wholeheartedly agree on importance and neglectedness. Our biggest challenge will be tractability. Winning the needed political victories at scale will require pitching all this in ways that resonate with the public mood, and navigating or weakening the political power centers (ex: labor, big business, rent seekers) that have bent our government into such hideous shape to begin with.
Right now, I worry those requirements are in tension with the AI optimism that abundance thinkers are drawn to. It seems like a comparable "blind spot on systems change" as the one you describe. Identifying policies causing bad outcomes is only the first step to changing a broken system, and I think later steps will require rallying coalitions around a simpler, angrier story.
Totally agree Andrew, thanks! :)
The presentation here could do with a bit of background. WHO was focused on bednets long before Effective Altruism was even heard of. The big dispute in the early 2000s was between advocates of bednets and proponents of DDT. Some DDT advocates were honest, but the most prominent were secretly financed by the tobacco industry, which wanted to discredit WHO campaigns against smoking.
Slow progress on vaccines is a largely separate issue, reflecting things like the lack of commercial interest in a once-off treatment that only benefits poor people. An abundance agenda means more public funding for essential research, not a hands-off approach by government.
Also relevantly for current debates about abundance it would be easy to reframe the npw-refuted arguments of DDT advocates in terms of an obvious solution, obstructed by environmentalist quibblee, bureaucratic inertia etc. In reality, the environmentalists were right all along, and the DDT advocates have moved on to other forms of science denialism
Here's a piece Tim Lambert and I wrote in 2008, by which time bednets had largely won the day.
https://www.prospectmagazine.co.uk/world/environment-news/52371/rehabilitating-carson
This is super fascinating, John. Thanks for flagging! I only learned a couple of days ago that DDT was once used as the malaria effort. A really fascinating post! That's awful how those companies tried to discredit DDT harm evidence.
I'm curious whether you see something similar here with Golden Rice. Is that an example where environmental groups aren't following the data?
Also, yes, the Abundance Agenda includes public funding for central research, but the Abundance approach would mean that the government is noticing the trade-offs that it's forcing by its risk aversion. The answer might be that it should be risk averse, but at least to have calculated those lives lost. Similar to how the Abundance book and Institute for Progress are focusing on how federal grants in the U.S. are decided by consensus, which means that moonshots won't happen. There needs to be some way to decide grants, let's trials others because upsides with this one are less likely to happen.
I do think, by proposing things like an examination of the system at large and also things like advanced market commitments, are Abundance, though perhaps it's in the Ezra Klein-Derek Thompson style Abundance. I see Abundance more as improving governance in things with big impacts, rather than being handsoff necessarily. WHO could have allowed rollout with real-world evidence collection, for example.