I do independent research on EA topics. I write about whatever seems important, tractable, and interesting (to me). Lately, I mainly write about EA investing strategy, but my attention span is too short to pick just one topic.
I have a website: https://mdickens.me/ Most of the content on my website gets cross-posted to the EA Forum.
My favorite things that I’ve written: https://mdickens.me/favorite-posts/
I used to work as a software developer at Affirm.
This principle has seemingly strange implications:
If and nothing has been done yet, then the first thing you do produces infinite utility (assuming you start by doing the best thing possible and then move to progressively worse things).
If , then a randomly-chosen opportunity has infinite expected utility.
I feel this way—I recently watched some footage of a PauseAI protest and it made me cringe, and I would hate participating in one. But also I think there are good rational arguments for doing protests, and I think AI pause protests are among the highest-EV interventions right now.
What do you mean by “if this is true”? What is “this”?
Jaan Tallinn, who funds SFF, has invested in DeepMind and Anthropic. I don’t know if this is relevant because AFAIK Tallinn does not make funding decisions for SFF (although presumably he has veto power).
Is PauseAI US a 501(c)(3)?
I’m confident in PauseAI US’s ability to run protests and I think the case for doing protests is pretty strong. You’re also doing lobbying, headed by Felix De Simone. I’m less confident about that so I have some questions.
There are no other organized groups (AFAIK) doing AI pause protests in the US of the sort you’re doing. But there are other groups talking to policy-makers, including Center for AI Policy, Center for AI Safety, and Palisade (plus some others outside the US, and some others that focus on AI risk but that I think are less value-aligned). What is the value-add of PauseAI US’s direct lobbying efforts compared these other groups? And are you coordinating with them at all?
What is Felix’s background / experience in this area? Basically, why should I expect him to be good at lobbying?
Given that a protein needs a very exact number of each amino acid to be synthesized, for essential amino acids like Methionine, I would expect their consumptions to be a bottleneck for muscle building which needs protein. Even if all the other proteins are in good amounts and thus the PDCAAS score is decent, you can drown in a river that’s on average 20cm deep.
I could be misreading but it sounds like you’re misunderstanding how PDCAAS works. If a food contains lots of every other essential amino acid but only (say) 30% of the required amount of methionine, then its PDCAAS will be 30. If the PDCAAS is (say) 90%, that means it contains at least 90% of the requirement for every essential amino acid, not just 90% on average.
You can buy methionine pills or powder: https://www.amazon.com/s?k=methionine&ref=nb_sb_noss
I believe they make methionine via chemical synthesis so it’s vegan.
For me personally, I get most of my daily protein from soy so I don’t worry about methionine, but I do have some methionine pills for when I need them. I also have lysine pills which pair well with wheat protein.
Recently I was annoyed at how difficult it was to figure out the protein quality of my diet so I built a calculator: https://mdickens.me/2024/08/29/DIAAS_calculator/
For me this seems more useful as a modus tollens: economists generally treat utility functions as isoelastic[1], which implies that opportunities are Pareto-distributed.
But it’s also useful as a sanity check: it’s intuitive to me that utility is isoelastic, and also that opportunities are Pareto-distributed, so it’s nice that these two intuitions are consistent with each other.
[1] Although this might be more out of convenience than anything else, since isoelastic utility functions have some nice mathematical properties.
Upvoted and disagree-voted.
Do you have any particular reason to believe that EAs overweight large problems?
Take the animal welfare vs global health debate: are we sometimes biased by the sheer magnitude of animal suffering numbers, rather than other relevant factors?
EAs donate much more to global health than to animal welfare. Do you think the ratio should be even higher still?
Can you explain how to use the discount rate parameter? From context, it looks like it’s meant as the “social discount rate”, not the “pure time preference” (which I would think should be 0). But for EAs, I think of the social discount rate as being positive mostly due to x-risk, and your model already separately accounts for AI x-risk, so should the social discount rate exclude discounting due to AI x-risk? In which case I would think it should be pretty low.
Edit: Never mind, I missed that you explained this in the post. Seems like you agree with my interpretation that the discount rate is based on non-AI x-risk.
(In which case 2% seems way too high to me, but that’s just details.)
I’m not sure if this is the same phenomenon, or a different phenomenon that uses the same word, but I see calls for “open-mindedness” in the woo community. When expressing my disbelief in ghosts/astrology/ESP/etc., I’m told I “need to be more open-minded”.
I interpreted disagreement to mean “EA is not reinventing the wheel with this concept”.
I get the sense that GiveWell would not recommend any animal welfare intervention (nor would they recommend any x-risk or policy intervention). But I don’t think that’s because they think any intervention that doesn’t meet their standards isn’t worth funding—they fund a lot of more speculative interventions thru Open Philanthropy. I think GiveWell wants to be viewed as a reliable source for high-quality charities, so they don’t want to recommend more speculative charities even if the all-things-considered EV is good.
(I’m just speculating here.)
As with ~all criticisms of EA, this open letter doesn’t have any concrete description of what would be better than EA. Like just once, I would like to see a criticism say, “You shouldn’t donate to GiveWell top charities, instead you should donate to X, and here is my cost-effectiveness analysis.”
The only proposal I saw was (paraphrased) “EA should be about getting teenagers excited to be effectively altruistic.” Ok, the movement-building arm of EA already does that. What is your proposal for what those teenagers should then actually do?
I only read the title and bolded summary of this post but I upvoted it because
The title and summary were sufficient for me to understand the argument being made
It was readily apparent to me that the argument makes sense
I had never considered the argument before
Juncture 1 – Animal Friendly Researchers
There is some reason to believe that virtually everyone is too animal-unfriendly, including animal welfare advocates:
Everyone is a human. Humans will naturally be biased toward humans over other species.
An uninformed prior says all individuals have equal moral weight. Virtually all people—including animal advocates—give humans higher weight than any other species, which is definitely a bias in a technical sense, if not in the sense people usually use the word.
Animal welfare has much higher EV even under conservative assumptions. IMO only plausible argument against is that the evidence base for animal welfare interventions is much worse, so if you are very skeptical of unproven interventions, you might vote the other way. But you’d have to be very skeptical.
“Slow-rolling mistakes” are usually much more important to identify than “point-in-time blunders”
After reading your post, I wasn’t sure you were right about this. But after thinking about it for a few minutes, I can’t come up with any serious mistakes I’ve made that were “point-in-time blunders”.
The closest thing I can think of is when I accidentally donated $20,000 to the GiveWell Community Foundation instead of The Clear Fund (aka GiveWell), but fortunately they returned the money so it all worked out.
This post has been sitting in my open tabs for 4 months and I am finally getting to it today.
I’m not entirely sure, but by my reading, the article is using this as an argument against η > 1. But I don’t think it’s really an argument. If η > 1 then indeed you should take into account absolute income levels, and I do in fact think you should do that. And yes that would change prioritization, and that’s a good thing because the current prioritization is probably wrong—it doesn’t assign enough weight to people with lower incomes.
You never actually say you’re arguing against η=1.87, but the title implies it.
FWIW I am not convinced by the evidence on what value of η to use, different lines of evidence point in different directions. I lean toward η > 1, so I think it’s reasonable to use η > 1 everywhere.
I think most people intuitively feel that income doublings matter more for poorer people, which requires η > 1.
η = 1 implies you can get arbitrarily high utility if you’re sufficiently rich, which seems wrong. It seems more likely that income doublings provide diminishing returns—once you reach very high income levels (like $100M+), income doublings hardly matter at all. This makes sense when you look at the space of consumable goods: doubling your income increases the size of the set of things you can buy, but each doubling increases the set size by less than the previous doubling.[1]
The best literature review I’ve seen comes from Gordon Irlam. The only real conclusion is that different methods see huge variance in estimates of η.
One paper that I particularly like[2], A New Method of Estimating Risk Aversion, estimates η using labor elasticity and finds η = 0.71 (see Table 1). But if you look at the various data sources it uses, the estimates of η vary greatly depending on source so the η = 0.71 average hides a lot of underlying uncertainty.
I often see people cite life satisfaction data showing η = 1. The commonly-cited paper, Stevenson & Wolfers (2013), didn’t perform any statistical tests for non-linearity on log-income vs. happiness. In Table 1, the paper did binary comparisons of the slope of log-income vs. happiness for rich vs. poor people and did not find clear differences, but it did find that the slope was generally steeper for rich people[3], which suggests η < 1 (I’m pretty sure on priors that η >= 1 so I don’t know what’s up with that result).
I briefly looked for more recent papers that test for non-linearity of log-income vs. happiness. I didn’t find exactly that, but I did find Happiness, income satiation and turning points around the world which finds that life satisfaction levels off at a certain income level. I didn’t read carefully but it looks like this paper used a sketchy spline curve-fitting method that I don’t trust (the fitted curves show that higher income decreases happiness above a certain point, which suggests that they’re using the wrong kind of curve; see Fig 1 and Fig 2[4]). But the fact that their spline curves level off suggests that happiness increases sub-linearly with income.
I feel like there’s room for a solid meta-analysis on income and life satisfaction, and I’m not satisfied with any of the existing literature.
In summary, the existing evidence is so high-variance that none of it meaningfully updates me away from my intuition that η must be greater than 1.
[1] This brings to mind a method for estimating η that I’ve never seen: Assume the prices of goods are Pareto-distributed and estimate the alpha parameter of the underlying Pareto distribution. Use that to estimate η (using this method).
[2] Even though I haven’t actually read most of it lol. I just like the concept. Maybe it contains a bunch of math errors, I don’t know.
[3] The paper did this comparison across a bunch of surveys. You’d need to do some kind of sophisticated non-standard significance test to determine if the overall difference is statistically significant, and the paper did not do that. (I think what you’d want to do is create a combined likelihood function that includes every survey and then get the p-value from the likelihood function. Or just skip the p-value and report the shape of the likelihood function because that’s more informative anyway.)
[4] Perhaps this is a property of the data, not the curve-fitting method. The paper says “in [some comparisons], the SWB [subjective well-being] level at satiation was greater due to turning-point effects (Bayes factor < 1⁄3).” They say they present this data in the supplementary appendix, which isn’t publicly available and isn’t on Sci-Hub (AFAICT), so it seems I can’t check.