EA Forum Bot Site
EA Forum

Comment Permalink
jackva3
0
0

If EA's can identify government actions with potentially high payoff, it could be a very good way to be effective.

This seems incomplete.

The criterion for effective action here would rather be something like a  (1) correctly estimated (2) high expected value of (3) marginal effort, e.g. that additional donations or work can affect the probability of important policy changes.

It  could be true that policies follow a power-law without this implying many effective actions (e.g.  this could be true in policy spaces that are crowded and where additional effort for one "side" leads to counteracting by another).

Most EAs working on issues outside global development seem to believe that funding marginal policy change in fairly technical issue areas (such as, before 2020, bio-risk, and AI policy, and also the top recs in climate) is very high EV,  with top recommended funding opportunities usually ones that influence policy in some way (in  a wide understanding of "policy", where this includes field building / coalition building). Matt's piece linked below gives good evidence for why that seems a reasonable assumption.

Reply
See in context
by felix.h
2 min read 6

12

Frontpage

This was a post I made in the EA subreddit (link here), but it might also receive some interesting input here, hence this crosspost. Would love to hear feedback on it!

 

One of the key insights of EA for me is that altruism is dependent on power laws. Certain interventions are much more effective than others, even though these effective options are sometimes ignored by the larger public.

For a while I have had the intuition this might also be the case for politics. I have been exhausted by the constant media cycle around it, with every week bringing a new controversy to be outraged about. But what if politics is dependent on power laws? What if, for example, 80% of the policy interventions have only 20% of the total impact in terms of lives saved, living quality improved, qalys extended...etc, but 20% of political actions produce 80% of the results?

Some cases I noticed seem to comply with this thought:

- George W. Bush, for all the bad things that he did, funded a little known anti-AIDS programme that probably saved more lives than were lost in the Iraq war.

- land reform in Asia: because of a quite specific alignment of people and forces the US supported a certain type of land reform in Taiwan, Japan and South-Korea after World War 2, which helped kickstart these countries' industrialisation, and economic trajectory out of poverty. (Sources: https://www.goodreads.com/book/show/16144575-how-asia-works & https://twitter.com/noahpinion/status/1052002737863385088)

- biotech R&D: South-Korea's industrial policy on biotech, active since the 1990ies, combined with the traumatic experience of MERS in 2015 built the basis for a quick setup of test-and-trace capacities when other countries fumbled with ramping up testing in the early stages of the pandemic. (source: https://www.korea-chair.eu/wp-content/uploads/2020/04/KFVUB_Policy-Brief-2020-04.pdf)

Of course most of this is anecdotal, and there's disagreement about how effective these policy actions were, but it would be logical that principles derived from development interventions carry over into politics. If EA's can identify government actions with potentially high payoff, it could be a very good way to be effective. It would also add some direction to what people such as Rob Wiblin already said about the social impact of voting, and about how EA's should engage with politics.

But again, this is mostly a gut feeling, so I would love input from EA's, and see if others have (not) been thinking in the same way.

12

0
0

Reactions

0
0
Comments6


Sorted by
New & upvoted
Click to highlight new comments since:
Matt_Lerner6
0
0

One thing to note here is that lots of commonly-used power law distributions have positive support. Political choices can and sometimes do have dramatically negative effects, and many of the catastrophes that EAs are concerned with are plausibly the result of those choices (like nuclear catastrophe, for instance). 

So a distribution that describes the outcomes of political choices should probably have support on the whole real line, and you wouldn't want to model choices with most simple power-law distributions.  But you might be on to something-- you might think of a hierarchical model in which there's some probability that decisions are either good or bad, and that the degree to which they are good or bad is governed by a power law distribution. That's the model I've been working with, but it seems incomplete to me.

Reply
Daniel_Eth4
0
0

Worth noting that if some political choices have very large negative outcomes, then choosing political paths that avoid those outcomes would have very positive counterfactual impact, even if no one sees it.

Reply
felix.h1
0
0

Good point! I can't say I have an immediate response, but I'm gonna think a bit more about this.

Reply
jackva3
0
0

If EA's can identify government actions with potentially high payoff, it could be a very good way to be effective.

This seems incomplete.

The criterion for effective action here would rather be something like a  (1) correctly estimated (2) high expected value of (3) marginal effort, e.g. that additional donations or work can affect the probability of important policy changes.

It  could be true that policies follow a power-law without this implying many effective actions (e.g.  this could be true in policy spaces that are crowded and where additional effort for one "side" leads to counteracting by another).

Most EAs working on issues outside global development seem to believe that funding marginal policy change in fairly technical issue areas (such as, before 2020, bio-risk, and AI policy, and also the top recs in climate) is very high EV,  with top recommended funding opportunities usually ones that influence policy in some way (in  a wide understanding of "policy", where this includes field building / coalition building). Matt's piece linked below gives good evidence for why that seems a reasonable assumption.

Reply
jackva2
0
0

There's some related discussion and empirical evidence here.

Reply
felix.h1
0
0

Thank you! Great post, with a lot of what I was looking for in there.

Reply
Curated and popular this week
 ·  · 16m read
 · 
This is a crosspost for The Case for Insect Consciousness by Bob Fischer, which was originally published on Asterisk in January 2025. [Subtitle.] The evidence that insects feel pain is mounting, however we approach the issue. For years, I was on the fence about the possibility of insects feeling pain — sometimes, I defended the hypothesis;[1] more often, I argued against it.[2] Then, in 2021, I started working on the puzzle of how to compare pain intensity across species. If a human and a pig are suffering as much as each one can, are they suffering the same amount? Or is the human’s pain worse? When my colleagues and I looked at several species, investigating both the probability of pain and its relative intensity,[3] we found something unexpected: on both scores, insects aren’t that different from many other animals.  Around the same time, I started working with an entomologist with a background in neuroscience. She helped me appreciate the weaknesses of the arguments against insect pain. (For instance, people make a big deal of stories about praying mantises mating while being eaten; they ignore how often male mantises fight fiercely to avoid being devoured.) The more I studied the science of sentience, the less confident I became about any theory that would let us rule insect sentience out.  I’m a philosopher, and philosophers pride themselves on following arguments wherever they lead. But we all have our limits, and I worry, quite sincerely, that I’ve been too willing to give insects the benefit of the doubt. I’ve been troubled by what we do to farmed animals for my entire adult life, whereas it’s hard to feel much for flies. Still, I find the argument for insect pain persuasive enough to devote a lot of my time to insect welfare research. In brief, the apparent evidence for the capacity of insects to feel pain is uncomfortably strong.[4] We could dismiss it if we had a consensus-commanding theory of sentience that explained why the apparent evidence is ir
 ·  · 40m read
 · 
I am Jason Green-Lowe, the executive director of the Center for AI Policy (CAIP). Our mission is to directly convince Congress to pass strong AI safety legislation. As I explain in some detail in this post, I think our organization has been doing extremely important work, and that we’ve been doing well at it. Unfortunately, we have been unable to get funding from traditional donors to continue our operations. If we don’t get more funding in the next 30 days, we will have to shut down, which will damage our relationships with Congress and make it harder for future advocates to get traction on AI governance. In this post, I explain what we’ve been doing, why I think it’s valuable, and how your donations could help.  This is the first post in what I expect will be a 3-part series. The first post focuses on CAIP’s particular need for funding. The second post will lay out a more general case for why effective altruists and others who worry about AI safety should spend more money on advocacy and less money on research – even if you don’t think my organization in particular deserves any more funding, you might be convinced that it’s a priority to make sure other advocates get more funding. The third post will take a look at some institutional problems that might be part of why our movement has been systematically underfunding advocacy and offer suggestions about how to correct those problems. OUR MISSION AND STRATEGY The Center for AI Policy’s mission is to directly and openly urge the US Congress to pass strong AI safety legislation. By “strong AI safety legislation,” we mean laws that will significantly change AI developers’ incentives and make them less likely to develop or deploy extremely dangerous AI models. The particular dangers we are most worried about are (a) bioweapons, (b) intelligence explosions, and (c) gradual disempowerment. Most AI models do not significantly increase these risks, and so we advocate for narrowly-targeted laws that would focus their att
 ·  · 10m read
 · 
Citation: McKay, H. and Shah, S. (2025). Forecasting farmed animal numbers in 2033. Rethink Priorities. The report is also available on the Rethink Priorities website. Executive summary We produced rough-and-ready forecasts of the number of animals farmed in 2033 with the aim of helping advocates and funders with prioritization decisions. We focus on the most numerous groups of farmed animals: broiler chickens, finfishes, shrimps, and select insect species. Our forecasts suggest almost 6 trillion of these animals could be slaughtered in 2033 (Figure 1).   Figure 1: Invertebrates could account for 95% of farmed animals slaughtered in 2033 according to our midpoint estimates. Note that ‘Insects’ only includes black soldier fly larvae and mealworms. Our midpoint estimates point to a potential fourfold increase in the number of animals slaughtered from 2023 to 2033 and a doubling of the number of animals farmed at any time. Invertebrates drive the majority of this growth, and could account for 95% of farmed animals slaughtered in 2033 (see Figure 1) and three quarters of those alive at any time in our mid-point projections. We believe our forecasts point to an urgent need to address critical questions around the sentience and welfare of farmed invertebrates. Our estimates come with many caveats and warnings. In particular: * Species scope: For practicality, we produced numbers only for a few key animal groups: broiler chickens, finfishes, shrimp, and certain insects (black soldier flies and mealworms only). * Sensitivity to insect farming growth: Our forecasts are particularly sensitive to the growth in insect farming, which is highly sensitive to the success of insect farming business models and their ability to attract future investment. The recent and forecasted estimates, with 90% subjective credible intervals, can be viewed below in Table 1.  Table 1: Estimates of recent and forecasted numbers of broiler chickens, finfishes, shrimps, and insects slau
Relevant opportunities
126
Please Donate to CAIP (Post 1 of 3 on AI Governance)
Jason Green-Lowe
· · 40m read
16
16
3
[Part-time AI Safety Research Program] MARS 3.0 Applications Open for Participants & Recruiting Mentors
Cambridge AI Safety Hub
· · 2m read
0
0
71
Can TikToks communicate AI policy and risk?
Caitlin Borke
· · 1m read
1
1
View more