Dan_Keys

Karma: 568

Dan_Keys Mar 12, 2025, 9:30 PM
2 points
2 ∶ 0
in reply to: Toby Tremlett🔹’s comment on: Announcing: Existential Choices Debate Week (March 17-24)
How about ‘On the margin, work on reducing the chance of our extinction is the work that most increases the value of the future’?
As I see it, the main issue with the framing in this post is that the work to reduce the chances of extinction might be the exact same work as the work to increase EV conditional on survival. In particular, preventing AI takeover might be the most valuable work for both. In which case the question would be asking to compare the overall marginal value of those takeover-prevention actions with the overall marginal value of those same actions.
(At first glance it’s an interesting coincidence for the same actions to help the most with both, but on reflection it’s not that unusual for these to align. Being in a serious car crash is really bad, both because you might die and because it could make your life much worse if you survive. Similarly with serious illness. Or, for nations/cities/tribes throughout history, losing a war where you’re conquered could lead to the conquerors killing you or doing other bad things to you. Avoiding something bad that might be fatal can be very valuable both for avoiding death and for the value conditional on survival.)

Dan_Keys Feb 13, 2025, 2:48 AM
4 points
1 ∶ 0
on: The standard case for delaying AI appears to rest on non-utilitarian assumptions
Nate Soares’ take here was that an AI takeover would most likely lead to an “unconscious meh” scenario, where “The outcome is worse than the “Pretty Good” scenario, but isn’t worse than an empty universe-shard” and “there’s little or no conscious experience in our universe-shard’s future. E.g., our universe-shard is tiled with tiny molecular squiggles (a.k.a. “molecular paperclips”).” Whereas humanity boosted by ASI would probably lead to a better outcome.
That was also the most common view in the polls in the comments there.

Dan_Keys Nov 13, 2024, 5:35 AM
4 points
2 ∶ 0
on: “Direct cash looks 3-4x more cost-effective in a new GiveWell assessment”
Wouldn’t the economic spillover effects depend on macroeconomic conditions? Government stimulus is more useful when there is more slack in the economy and more inflationary when there’s a tight labor market. I’d expect cash transfers to be similar.
I don’t know the conditions in the specific places studied, but in a lot of places there was significant slack in the economy from the Great Recession until Covid, and the labor markets are now tighter. So studies conducted in the 2010s might overestimate the present-day net benefits of economic spillovers.

Dan_Keys Aug 19, 2024, 8:19 PM
1 point
0 ∶ 0
on: Does “Ultimate Neartermism” via Eternal Inflation dominate Longtermism in expectation?
This sounds like one of those puzzles of infinities. If you take the limits in one way then it seems like one infinity is bigger than another, but if you take the limits a different way then the other infinity seems bigger.
A toy version: say that things begin with 1 bubble universe at time 0 and proceed in time steps, and at time step k, 10^k new bubble universes begin. Each bubble universe lasts for 2 time steps and then disappears. This continues indefinitely.
Option A: each bubble universe has a value of 1 in the first time step of its existence and a value of 5 in its second time step. (Then it disappears, or forever after has value 0.)
Option B: each bubble universe has a value of 3 in the first time step of its existence and a value of 1 in its second time step. (Then it disappears, or forever after has value 0.)
This has the same basic structure as the setup in the post, though with much smaller numbers.
We could try summing across all bubble universes at each time step, and then taking the limit as the total number of time steps increases without bound. Option B is 3x as good in the zeroth time step (value of 3 vs. 1), 2.125x as good through the next time step (value of 34 vs. 16), about 2.072x as good through the next time step (value of 344 vs. 166), and in the limit as the number of time steps increases without bound it is 2.0666… times as good (31/15). That is how this post sets up its comparison of infinities (with larger numbers so the ratio would be much more lopsided).
Instead, we could try summing within each bubble universe across all of its time steps, and then sum across all complete bubble universes. Each bubble universe has a total value of 6 in Option A vs. 4 in Option B, so Option A is 1.5x as good for each of them. Option A is 1.5x as good for the first bubble universe that appears (6 vs. 4), and for the first 11 bubble universes it is 1.5x as good (66 vs. 44), and for the first 111 bubble universes it is 1.5x as good (666 vs. 444), and if you take the limit as the number of bubble universes increases without bound it is 1.5x as good. This matches the standard longtermist argument (which has larger numbers so the ratio would be more lopsided).

Dan_Keys Jun 10, 2024, 8:13 PM
9 points
2 ∶ 1
on: Reducing the nearterm risk of human extinction is not astronomically cost-effective?
I disagree. One way of looking at it:
Imagine many, many civilizations that are roughly as technologically advanced as present-day human civilization.
Claim 1: Some of them will wind up having astronomical value (at least according to their own values)
Claim 2: Of those civilizations that do wind up having astronomical value, some will have gone through near misses, or high-risk periods, when they could have gone extinct if things had worked out slightly differently
Claim 3: Of those civilizations that do go extinct, some would have had wound up having astronomical value if they had survived that one extinction event. These are civilizations much like the ones in claim 2, but who got hit instead of getting a near miss.
Claim 4: Given claims 1-3, and that the “some” civilizations described in claims 1-3 are not vanishingly rare (enough to balance out the very high value), the expected value of averting a random extinction event for a technologically advanced civilization is astronomically high.
Then to apply this to humanity, we need something like:
Claim 5: We don’t have sufficient information to exclude present-day humanity from being one of the civilizations from claim 1 which winds up having astronomical value (or at least humanity conditional on successfully navigating the transition to superintelligent AI and surviving the next century in control of its own destiny)

Dan_Keys Apr 23, 2024, 12:50 AM
2 points
0 ∶ 0
on: Extinction risk and longtermism: a broader critique of Thorstad
I’ve raised related points here, and also here with followup, about how exponential decay with a fixed decay rate is not a good model to use for estimating long-term survival probability.

Dan_Keys Feb 26, 2024, 5:33 AM
30 points
5 ∶ 1
on: UPDATE: Critical Failures in the World Happiness Report’s Model of National Satisfaction
Does your model without log(GNI per capita) basically just include a proxy for log(GNI per capita), by including other predictor variables that, in combination, are highly predictive of log(GNI per capita)?
With a pool of 1058 potential predictor variables, many of which have some relationship to economic development or material standards of living, it wouldn’t be surprising if you could build a model to predict log(GNI per capita) with a very good fit. If that is possible with this pool of variables, and if log(GNI per capita) is linearly predictive of life satisfaction, then if you build a model predicting life satisfaction which can’t include log(GNI per capita), it can instead account for that variance by including the variables that predict log(GNI per capita).
And if you transform log(GNI per capita) into a form whose relationship with life satisfaction is sufficiently non-linear, and build a model which can only account for the linear portion of the relationship between that transformed variable and life satisfaction, then within that linear model those proxy variables might do a much better job than transformed log(GNI per capita) of accounting for the variance in life satisfaction.

Dan_Keys Jan 24, 2024, 9:17 PM
4 points
1 ∶ 0
in reply to: Otto’s comment on: Is fear productive when communicating AI x-risk? [Study results]
It looks like the 3 articles are in the appendix of the dissertation, on pages 65 (fear, Study A), 72 (hope, Study B), and 73 (mixed, Study C).

Dan_Keys Nov 17, 2023, 7:11 PM
5 points
0 ∶ 0
on: What social science research do you want to see reanalyzed?
The effect of health insurance on health, such as the old RAND study, the Oregon Medicaid expansion, the India study from a couple years ago, or whatever else is out there.
Robin Hanson likes to cite these studies as showing that more medicine doesn’t improve health, but I’m skeptical of the inference from ‘not statistically significant’ to ‘no effect’ (I’m in the comments there as “Unnamed”). I would like to see them re-analyzed based on effect size (e.g. a probability distribution or confidence interval for DALY per $).

Dan_Keys Nov 10, 2023, 10:48 PM
4 points
1 ∶ 0
in reply to: Vasco Grilo🔸’s comment on: Rethink Priorities’ Cross-Cause Cost-Effectiveness Model: Introduction and Overview
I’d guess that this is because an x-risk intervention might have on the order of a ¹⁄_100,000 chance of averting extinction. So if you run 150k simulations, you might get 0 or 1 or 2 or 3 simulations in which the intervention does anything. Then there’s another part of the model for estimating the value of averting extinction, but you’re only taking 0 or 1 or 2 or 3 draws that matter from that part of the model because in the vast majority of the 150k simulations that part of the model is just multiplied by zero.
And if the intervention sometimes increases extinction risk instead of reducing it, then the few draws where the intervention matters will include some where its effect is very negative rather than very positive.
One way around this is to factor the model, and do 150k Monte Carlo simulations for the ‘value of avoiding extinction’ part of the model only. The part of the model that deals with how the intervention affects the probability of extinction could be solved analytically, or solved with a separate set of simulations, and then combined analytically with the simulated distribution of value of avoiding extinction. Or perhaps there’s some other way of factoring the model, e.g. factoring out the cases where the intervention has no effect and then running simulations on the effect of the intervention conditional on it having an effect.

Dan_Keys Nov 4, 2023, 5:46 PM
57 points
3 ∶ 0
on: Clean Water—the incredible 30% mortality reducer we can’t explain
I believe the paper you’re referring to is “Water Treatment And Child Mortality: A Meta-Analysis And Cost-effectiveness Analysis” by Kremer, Luby, Maertens, Tan, & Więcek (2023).
The abstract of this version of the paper (which I found online) says:
We estimated a mean cross-study reduction in the odds of all-cause under-5 mortality of about 30% (Peto odds ratio, OR, 0.72; 95% CI 0.55 to 0.92; Bayes OR 0.70; 95% CrI 0.49 to 0.93). The results were qualitatively similar under alternative modeling and data inclusion choices. Taking into account heterogeneity across studies, the expected reduction in a new implementation is 25%.
That’s a point estimate of a 25-30% reduction in mortality (across 3 methods of estimating that number), with a confidence/credible interval that has a lower bound of a 7-8% reduction in mortality. So, it’s a fairly noisy estimate, due to some combination of the noisiness of individual studies and the heterogeneity across different studies.
That interval for the reduction in mortality just barely overlaps with your number that “Sub-saharan Africa diarrhoea causes 5-10% of child mortality.” (The overlap might be larger if that rate was higher than 5-10% in the years & locations where the studies were conducted.)
So it could be that the clean water interventions prevent most children’s deaths from diarrhoea and few other deaths, if the mortality reduction is near the bottom of the range that Kremer & colleagues estimate. Or they might prevent a decent chunk of other deaths, but not nearly as many as your Part 4 chart & list suggest, if the true mortality reduction is something like 15%.
There is generally also a possibility of a meta-analysis giving inflated results, due to factors like publication bias affecting which studies they include or other methodological issues in the original studies, which could mean that the true effect is smaller than the lower bound of their interval. I don’t know how likely that is in this case.
Here’s a more detailed look at their meta-analysis results:

Dan_Keys Oct 26, 2023, 11:53 PM
37 points
2 ∶ 0
on: UK Prime Minister Rishi Sunak’s Speech on AI
More from Existential Risk Observatory (@XRobservatory) on Twitter:
It was a landmark speech by @RishiSunak: the first real recognition of existential risk by a world leader. But even better are the press questions at the end:
@itvnews: “If the risks are as big as you say, shouldn’t we at the very least slow down AI development, at least long enough to understand and control the risks.”
@SkyNews: “Is it fair to say we know enough already to call for a moratorium on artificial general intelligence? Would you back a moratorium on AGI?”
Sky again: “Given the harms and the risk you pointed out in this report, and some of those are profound, surely there must be some red lines we can draw at this point. Which ones are yours?”
@TheSun: “You say we shouldn’t be losing sleep over this stuff. If not, why not?”
@theipaper: “You haven’t really talked about whether your government is actually going to regulate. Will there be an AI Bill or similar on the The King’s Speech?”
iNews again: “On the details of that regulation: does the government remain commited to this idea of responsible scaling, whereby you sort of test models after they’re being developed, or is it time to start thinking about how you intervene to stop the most dangerous models being developed at all?”
Who would have thought one year ago? The public debate about AI xrisk so far outdoes everyone’s expectations. Next step: convincing answers.
https://www.youtube.com/watch?v=hSup6mgNhzQ

Dan_Keys Oct 23, 2023, 9:11 PM
14 points
3 ∶ 0
in reply to: Dan_Keys’s comment on: How bad would human extinction be?
One way to build risk decay into a model is to assume that the risk is unknown within some range, and to update on survival.
A very simple version of this is to assume an unknown constant per-century extinction risk, and to start with a uniform distribution on the size of that risk. Then the probability of going extinct in the first century is ¹⁄₂ (by symmetry), and the probability of going extinct in the second century conditional on surviving the first is smaller than that (since the higher-risk worlds have disproportionately already gone extinct) - with these assumptions it is exactly ¹⁄₃. In fact these very simple assumptions match Laplace’s law of succession, and so the probability of going extinct in the nth century conditional on surviving the first n-1 is 1/(n+1), and the unconditional probability of surviving at least n centuries is also 1/(n+1).
More realistic versions could put more thought into the prior, instead of just picking something that’s mathematically convenient.
What links here?
- Dan_Keys's comment on Extinction risk and longtermism: a broader critique of Thorstad by Matthew Rendall (Apr 23, 2024, 12:50 AM; 2 points)

Dan_Keys Oct 23, 2023, 9:11 PM
49 points
8 ∶ 5
on: How bad would human extinction be?
Why are these expected values finite even in the limit?
It looks like this model is assuming that there is some floor risk level that the risk never drops below, which creates an upper bound for survival probability through n time periods based on exponential decay at that floor risk level. With the time of perils model, there is a large jolt of extinction risk during the time of perils, and then exponential decay of survival probability from there at the rate given by this risk floor.
The Jupyter notebook has this value as r_low=0.0001 per time period. If a time period is a year, that means a ¹⁄_10,000 chance of extinction each year after the time of perils is over. This implies a 10^-43 chance of surviving an additional million years after the time of perils is over (and a 10^-434 chance of surviving 10 million years, and a 10^-4343 chance of surviving 100 million years, …). This basically amounts to assuming that long-lived technologically advanced civilization is impossible. It’s why you didn’t have to run this model past the 140,000 year mark.
This constant r_low also gives implausible conditional probabilities. e.g. Intuitively, one might think that a technologically advanced civilization that has survived for 2 million years after making it through its time of perils has a pretty decent chance of making it to the 3 million year mark. But this model assumes that it still has a ¹⁄_10,000 chance of going extinct next year, and a 10^-43 chance of making it through another million years to the 3 million year mark.
This seems like a problem for any model which doesn’t involve decaying risk. If per-time-period risk is 1/n, then the model becomes wildly implausible if you extend it too far beyond n time periods, and it may have subtler problems before that. Perhaps you could (e.g.) build a time of perils model on top of a decaying r_low.
What links here?
- Dan_Keys's comment on Extinction risk and longtermism: a broader critique of Thorstad by Matthew Rendall (Apr 23, 2024, 12:50 AM; 2 points)

Dan_Keys Oct 19, 2023, 11:01 PM
6 points
1 ∶ 0
on: Comparing Two Forecasters in an Ideal World
You can get a sense for these sorts of numbers just by looking at a binomial distribution.
e.g., Suppose that there are n events which each independently have a 45% chance of happening, and a noisy/biased/inaccurate forecaster assigns 55% to each of them.
Then the noisy forecaster will look more accurate than an accurate forecaster (who always says 45%) if >50% of the events happen, and you can use the binomial distribution to see how likely that is to happen for different values of n. For example, according to this binomial calculator, with n=51 there is a 24% chance that at least ²⁶⁄₅₁ of the p=.45 events will resolve as True, and with n=201 there is a 8% chance (I’m picking odd numbers for n so that there aren’t ties).
With slightly more complicated math you can look at statistical significance, and you can repeat for values other than trueprob=45% & forecast=55%.

Dan_Keys Sep 8, 2023, 1:39 AM
4 points
3 ∶ 0
on: There is Little Evidence on Question Decomposition
Seems like a question where the answer has to be “it depends”.
There are some questions which have a decomposition that helps with estimating them (e.g. Fermi questions like estimating the mass of the Earth), and there are some decompositions that don’t help (for one thing, decompositions always stop somewhere, with components that aren’t further decomposed).
Research could help add texture to “it depends”, sketching out some generalizations about which sorts of decompositions are helpful, but it wouldn’t show that decomposition is just generally good or just generally bad or useless.

Dan_Keys Jul 4, 2023, 2:12 AM
16 points
2 ∶ 0
on: Mistakes in the moral mathematics of existential risk (Part 1: Introduction and cumulative risk) - Reflective altruism
However, an absolute reduction of cumulative risk by 10^-8 requires (by definition) driving cumulative risk at least below 1-10^-8. Again, you say, that must be easy. Not so. Driving cumulative risk this low requires driving per-century risk to about 1.6*10^-6, barely one in a million.
I’m unclear on what this means. I currently think that humanity has better than a 10^-8 chance of surviving the next billion years, so can I just say that “driving cumulative risk at least below 1-10^-8” is already done? Is the 1.6*10^-6per-century risk some sort of average of 10 million different per-century numbers (such that my views on the cumulative risk imply that this risk is similarly already below that number), or is this trying to force our thinking into an implausible-to-me model where the per-century risk is the same in every century, or is this talking about the first future century in which risk drops below that level?
On the whole, this decay-rate framing of the problem feels more confusing to me than something like a two-stage framing where there is some short-term risk of extinction (over the next 100 or 1000 years or similar) and then some probability of long-term survival conditional on surviving the first stage.
e.g., Suppose that someone thinks that humanity has a 10^-2 chance (1%) of surviving the next thousand years, and a 10^-4 chance (.01%) of surviving the next billion years conditional on surviving the next thousand years, and that our current actions can only affect the first of those two probabilities. Then increasing humanity’s chances of surviving a billion years by 10^-8 (in absolute terms) requires adding 10^-4 to our 10^-2 chance of surviving the next thousand years (an absolute .01% increase), or, equivalently, multiplying our chances of surviving the next thousand years by x1.01 (a 1% relative increase).

Dan_Keys Jul 3, 2023, 10:56 PM
11 points
3 ∶ 0
on: Existential risk pessimism and the time of perils
Constant per-century risk is implausible because these are conditional probabilities, conditional on surviving up to that century, which means that they’re non-independent.
For example, the probability of surviving the 80th century from now is conditioned on having survived the next 79 centuries. And the worlds where human civilization survives the next 79 centuries are mostly not worlds where we face a 10% chance of extinction risk each century and keep managing to stumble along. Rather, they’re worlds where the per-century probabilities of extinction over the next 79 centuries are generally lower than that, for whatever reason. And worlds where the next 79 per-century extinction probabilities are generally lower than 10% are mostly worlds where the 80th extinction probability is also lower than that. So, structurally we should expect extinction probabilities to go down over time, as cumulative extinction risk means filtering for less extinction-prone worlds.
Dice roll style models which assume independence can be helpful for sketching out some of the broad contours of the problem, but in many cases they don’t capture the structure of the situation well enough to use them for quantitative forecasts. This is an issue for modeling many domains, like for projecting how many games a sports team will win over the upcoming season. If your central estimate is that a team will win 30% of its games, worlds where the team winds up winning more than half of its games are mostly not worlds where the team kept getting lucky game after game. Rather, they’re worlds where the team was better than expected, so they generally had more than a 30% chance of winning each game.
A model that doesn’t account for this non-independence across games (like just using a binomial distribution based on your central estimate of what fraction of games a team will win) has built in the assumption that the only way to win more games is to have a bunch of 30% events go their way, and implicitly rules out the possibility of the team being better than your central estimate, so it will give inaccurate distributions. For example 3 of the 30 NBA teams had less than a 1/1000 chance of winning as many games as they did this year according to the distributions you’d get from a simple binomial model using the numbers here.
Similarly, Sam Wang’s 2016 election forecast gave Trump less than a 1% chance of winning the US presidency because it failed to correctly account for correlated uncertainty. By not sufficiently tracking non-independence, the model accidentally made extremely strong assumptions which led to a very overconfident prediction.
What links here?
- Dan_Keys's comment on Extinction risk and longtermism: a broader critique of Thorstad by Matthew Rendall (Apr 23, 2024, 12:50 AM; 2 points)

Dan_Keys May 7, 2023, 7:26 PM
22 points
4 ∶ 0
on: On Child Wasting, Mega-Charities, and Measurability Bias
GiveWell has a 2021 post Why malnutrition treatment is one of our top research priorities, which includes a rough estimate of “a cost of about $2,000 to $18,000 per death averted” through treating “otherwise untreated episodes of malnutrition in sub-Saharan Africa.” You can click through to the footnotes and the spreadsheets for more details on how they calculated that.

Dan_Keys Mar 25, 2023, 1:37 AM
13 points
4 ∶ 0
in reply to: titotal’s comment on: Exploring Metaculus’ community predictions
Is this just showing that the predictions were inaccurate before updating?
I think it’s saying that predictions over the lifetime of the market are less accurate for questions where early forecasters disagreed a lot with later forecasters, compared to questions where early forecasters mostly agreed with later forecasters. Which sounds unsurprising.