2020 AI Alignment Literature Review and Charity Comparison 2020-12-21T15:25:04.543Z
Avoiding Munich's Mistakes: Advice for CEA and Local Groups 2020-10-14T17:08:13.033Z
Will protests lead to thousands of coronavirus deaths? 2020-06-03T19:08:10.413Z
2019 AI Alignment Literature Review and Charity Comparison 2019-12-19T02:58:58.884Z
2018 AI Alignment Literature Review and Charity Comparison 2018-12-18T04:48:58.945Z
2017 AI Safety Literature Review and Charity Comparison 2017-12-20T21:54:07.419Z
2016 AI Risk Literature Review and Charity Comparison 2016-12-13T04:36:48.060Z
Being a tobacco CEO is not quite as bad as it might seem 2016-01-28T03:59:15.614Z
Permanent Societal Improvements 2015-09-06T01:30:01.596Z
EA Facebook New Member Report 2015-07-26T16:35:54.894Z


Comment by Larks on Concerns with ACE's Recent Behavior · 2021-04-21T19:33:00.396Z · EA · GW

Thanks very much for this comment, and for correcting my mistaken assumption.

Comment by Larks on EA Forum feature suggestion thread · 2021-04-21T19:28:04.662Z · EA · GW

An approach some forum use is the ratio of up and downvotes: -38+40 is not the same as +2 ! This allows you to have a smooth measure of the degree of controversy rather than a binary classification.

Comment by Larks on EA Forum feature suggestion thread · 2021-04-21T14:53:13.934Z · EA · GW

I would consider something to reduce the karma users can get from commenting on controversial posts. Right now it seems easy to get very high scores by making not really that great comments in such places.

As an example, I think this comment I made is decent. It makes a true and relevant point that no-one else had mentioned . But it's not great; the topic of that thread is not that important, and the all the comments in it, let alone mine alone, do not resolve the issue. Most importantly, that comment is definitely not over 50% as good as this article I wrote. I would say the article is at least a thousand times more important, and took at least a thousand times longer to write.

I'm not sure how exactly you would do this though, as all the most obvious methods have significant drawbacks.

Comment by Larks on If Bill Gates believes all lives are equal, why is he impeding vaccine distribution? · 2021-04-21T06:09:26.828Z · EA · GW

This struck me as a catastrophic move, turning a vaccine developed by a nonprofit institution into a way to make a company lots of money, with no clear upside. (I think if Oxford stuck to their original plan, they could control the IP and require that their manufacturing partner sell the vaccines at cost). 

Both AstraZeneca and J&J have agreed that they will sell the vaccines at cost during the pandemic; this is mentioned in one of the articles you linked.

Unfortunately, this agreement has had major downsides. By preventing AstraZeneca from making a profit, it has undermined their ability to invest in capacity, and is one of the causes of the numerous production setbacks we have seen. I think it would have been much better if AstraZeneca was allowed to charge higher prices to incentivise them to produce more! If there is a shortage, that suggests that prices are too low, not too high.

Companies that own rights to vaccines developed with enormous public investment and massive guaranteed government contracts are fighting any efforts to share the knowledge that could be used make more vaccines in more parts of the world. 

AstraZeneca signed a voluntary licensing agreement with various Indian manufacturers to allow them to produce its vaccine. This fact was pointed out to you when you wrote about this subject on facebook.

In fact, the novel technology at the heart of the Moderna vaccine, for example, was developed partly by the National Institutes of Health using U.S. federal funds. Moderna then received a total of some $2.5 billion in taxpayer money for research support and as preorders for vaccines; by the company’s own admission, the $1 billion contribution it received for research covered 100 percent of those costs.

Given you highlight Moderna, why did you not mention that they have in fact agreed not to enforce their vaccine IP rights during the pandemic?

As it happens, I doubt this will be that important a move. Manufacturing mRNA vaccines is a very advanced business that requires a lot of expertise; these are not small molecule drugs! Even with the IP waiver I suspect most others would struggle to do what Moderna and Pfizer have achieved. There just is not a huge amount of currently idle capacity in this supply chain. For more detail on the manufacturing details, I recommend reading this article.

In practice, pharma companies are far more focused on boosting their share price than on developing new drugs. For example, in 2018, "the 10 biggest pharmaceutical companies spent nearly 170 percent of their net income (tapping cash reserves and borrowing) on payments to shareholders and executives through stock buybacks and dividends".

Developing new drugs and 'boosting the share price' are not in opposition. If a company develops a promising new drug, the expectation of future profits will cause the share price to increase. This is why biotech companies can see their share prices go up even if they are not doing any cash return at all.

I also think you have misunderstood how accounting works. Net Income is after R&D expense. The more a company spends on R&D, the lower Net Income will be. The fact that Net Income is low compared to cash return doesn't mean that the company isn't doing R&D.

As an example, take a look at Pfizer. In 2020 they spent about $9.4bn on R&D vs Net Income of around $9.6bn. This exceeded their dividends and buybacks, which were around $8.4bn. And we should probably also included M&A, which is basically indirect R&D. This is lumpy but large - e.g. over $10bn for 2019 for example.

I actually think that it is a major problem how low the prices for the vaccines have been. These drug companies have provided us with a way out of the pandemic that many people thought was impossible - and very few expected as quickly as this. For this they should have been richly rewarded - we want to incentivise companies to work on the world's biggest problems for the future! Unfortunately, if you look at the share prices of Pfizer or AstraZeneca and compare them to the broader market, you can see this was not the case. Given the amount these companies have been attacked by politicians, I think it is plausible that AstraZeneca leadership might regard this entire endeavour as a mistake.

Comment by Larks on As an EA, Should I renounce my US citizenship? · 2021-04-18T22:04:12.299Z · EA · GW

Do you currently have significant assets? I am not an expert in this but historically the US has imposed exit taxes on people who renounce their citizenship, to try to recoup the taxes they think you 'should' have paid.

Comment by Larks on College and Earning to Give · 2021-04-18T02:20:25.426Z · EA · GW

While I previously suggested students get married so their parents' finances wouldn't be considered, schools that do this kind of price discrimination request the CSS Profile in addition to the FAFSA, which does include finances for parents of married students.

I've also heard that a strategy is to have your kids get adopted, but as I look into it I'm not sure why this would work, as it seems the CSS profile would then look at your child's adoptive parent's assets. I guess you could have your children be adopted by someone with low income/assets. (I'm also not sure if this strategy is legal)

Comment by Larks on Concerns with ACE's Recent Behavior · 2021-04-16T13:16:10.276Z · EA · GW

Having written a similar post in the past, it's worth keeping in mind the amount of time they take to write is huge. Hypatia seems to have done a very good job expressing the facts in a way which communicates why they are so concerning while avoiding hyperbole. While giving organisations a chance to read a draft can be a good practice to reduce the risk of basic factual mistakes (and one I try to follow generally), it's not obligatory. Note that we generally do not afford non-EA organisations this privilege, and indeed I would be surprised if ACE offered Connor the chance to review their public statement which pseudonymously condemned him. Doing so adds significantly to the time commitment and raises anonymity risks[1], especially if one is worried about retaliation from an organisation that has penalized people for political disagreements in the past.


[1] As an example, here is something I very nearly messed up and only thought of at the last minute: you need to make a fresh copy of the google doc to share without the comments, or you will reveal the identity of your anonymous reviewers, even if you are personally happy to be known. 

Comment by Larks on Concerns with ACE's Recent Behavior · 2021-04-16T12:42:56.741Z · EA · GW

Thank you for writing this extremely detailed and thoughtful post on a very concerning topic.

Comment by Larks on If I pay my taxes, why should I also give to charity? · 2021-04-15T17:34:04.402Z · EA · GW

I'm glad you wrote this article, as I think this is a common objection, especially among more conservative people, so it is good to have a stock response. However, I think it could be significantly improved by first building up a steel-man of the 'myth' and then precisely targetting the response.

Most people around the world pay 20–40% of their income in taxes to support government services.

This link goes to a table of top quoted marginal income tax rates. However, this is an over-estimate, because the average is below the marginal. It is also an under-estimate, because it does not include sales taxes, property taxes, VATs, inheritance taxes, corporation taxes, payroll taxes, petrol taxes, tobacco taxes, mandatory insurance purchases, licence fees, utility taxes or import tariffs. I would recommend looking at total government spending / GDP instead.

This obligatory and often large payment can feel like the end of our civic duties — the culmination of our "fair" contribution to society, nullifying the need for further charitable giving.

I think this is a bit of a straw-man; I would expect people making this objection to also regard voting, jury duty, educating their children, national service, not committing crimes etc. to also be part of their civic duties.

Also if your point is that charity is different from taxes I'm not sure it makes sense to talk about 'further' charitable giving.

While lower-income countries suffer from an overall lack of funding, wealthier countries are liable to misallocate the funding they have.

This seems like a strange argument. Yes, small countries would benefit from more funding, but so would rich countries. Similarly, while rich countries are liable to spend on wasteful things, so are poor countries. The correct argument is not 'need vs misallocation', it is about diminishing marginal utility of money.

Also, the citation you provided seems rather niche. There are many major examples of government waste in the first world - extremely inefficient procurement being a very clear and non-partisan example. In contrast, the fact that some authors argue that too much climate change research funding has gone to research climate change seems like a relatively small and not obviously compelling example.

Political mobilization can influence governments' budgetary spending and help correct such misallocations,

This does not seem obviously non-trivially the case to me. There are many examples of political campaigns pushing for increasing in spending on something, but I struggle to think of many examples of similar successful campaigns to cut spending. And on average I would generally expect populist campaigns to reduce the average efficiency of spending, even if they 'could' in theory improve it.

Rather than viewing taxes as charity, think of them as payment for services that benefit you and your fellow citizens. Infrastructure and social services are important!

I'm not sure why you are pushing this view, it seems very confused to me:

  1. If they are a 'payment for services that benefit you' it is a very strange sort of payment. The amount of taxes you pay is not very closely linked at all to the amount of benefit you get, and you have almost no ability to opt out. Indeed, many people will pay taxes that go to fund the government hurting them.
  2. If the taxes really are a payment for services to benefit your fellow citizens, this seems somewhat like charity. In this case, why shouldn't I treat taxes to fund a fellow citizen's education as being similar to donations to a domestic education charity?
  3. Only a relatively small fraction of taxes go towards infrastructure. Even Biden's recent 'Infrastructure Bill' is mostly non-infrastructure spending.
  4. Public serves are a larger component of spending, but are still often dominated by welfare spending of one kind or other.
  5. This section takes an unnecessarily political view that risks alienating some readers despite making no difference to the core argument. Even if taxes were 100% waste there would still be a strong argument to give to charity.

Paying your taxes is the right thing to do, but paying your taxes alone is not sufficient to significantly improve the lives of others.

Similarly, I'm not sure what argument this is trying to make. The idea that we have an obligation to pay taxes is not uncontroversial, but to the extent people believe it, it is largely because they think doing so makes people better off. A lot of people probably believe that if they were excused from paying taxes that would significantly improve their life; if paying taxes doesn't significantly help anyone else, but it does hurt them, why should paying be the right thing for them to do?

But again, this seems extraneous. Giving to charity is a good idea regardless of whether paying taxes significantly improves the lives of others, and regardless of whether it is the 'right thing to do'.

Comment by Larks on "Insider giving" - An unfortunate donation strategy used by corporate insiders to avoid losses · 2021-04-14T03:41:34.812Z · EA · GW

Interesting theory!

certainly inflates donation figures

Presumably this effect would be significantly reduced if charities sold the stock as soon as they received it, as they would also sell at the 'inflated' price?

Comment by Larks on Confusion about implications of "Neutrality against Creating Happy Lives" · 2021-04-11T19:01:26.810Z · EA · GW

I have never seen a survey on this, but I think most people here adopt a totalist view on which creating new happy people is good, because of e.g. the classic transitivity argument. So you were correct to be confused!

Comment by Larks on Status update: Getting money out of politics and into charity · 2021-04-08T02:36:10.742Z · EA · GW

Is there a better way to go about outreach to donors than what we’re doing (i.e. sending emails, trying to figure out what works, iterating)? I figure some of you might have experience with this.

I would have expected that ultimately building a website and trying to build media hype would be more effective. Of course, you will have to be careful not to accidentally build up lopsided hype - e.g., for ever article in the NYT you need one in the WSJ. But presumably this is for after you settle on a product.

Comment by Larks on HaukeHillebrandt's Shortform · 2021-04-06T16:22:19.631Z · EA · GW

Ahh, in that case I agree that HKers, or even better Uighurs, would be well placed. But my impression was that 80k etc.'s concerns about China mainly revolved around things like improving Western-Chinese coordination to reduce the risk of war, AI race or climate change, rather than human rights. I would think that putting pressure on them for human rights abuses would be likely to make this worse, as the CCP views such activism as an attack on their system. It is hard to cooperate with someone if they are denouncing you as evil and funding your dissidents. 

Comment by Larks on HaukeHillebrandt's Shortform · 2021-04-06T14:27:26.223Z · EA · GW

I would have thought they would be unusually badly placed, because the regime will view them as traitors, for the same reason I would not recommend using apostates for outreach to muslims. 

Comment by Larks on Jakob_J's Shortform · 2021-04-05T00:49:52.069Z · EA · GW

Childcare is a very big cost - if you think that you are trying to spend a portion of your income on childcare, but for the worker this is their entire income, and the number of children one person can look after is limited (both by practicalities and also regulation), you can see why it tends to be expensive.

I would however keep in mind that most people who work in the public or NGO sector do manage to raise families though!

Comment by Larks on Mundane trouble with EV / utility · 2021-04-04T18:45:15.761Z · EA · GW

I still disagree because I am assuming I only get one chance at doing the action and personally I don't value a 1 in a million chance of being saved higher than zero.

Would you be interested in selling me a lottery ticket? We can use an online random number generator. I will win with a one-in-a-million chance, in which case you will give me all your worldly possessions, including all your future income, and you swear to do my wishes in all things. I will pay you $0.01 for this lottery ticket. 

If you really believed that one-in-a-million was the same as zero, this should be an attractive deal for you. But my guess is that actually you would not want to take it!

Comment by Larks on Mundane trouble with EV / utility · 2021-04-04T04:10:45.071Z · EA · GW

Here is one answer to one part of your question.

Naively (?), I'd say that if you invest your resources in an action with a 1 in a million chance of saving a billion lives, and a 999,999 in a million chance of having no effect, then (the overwhelming majority of the time) you haven't done anything at all.

I think the idea you are getting at here is you'd rather have a 10% chance of saving 10 people than a 1/million chance of saving a billion lives. I can definitely see why you would feel this way (and there are prudential reasons for it - e.g. worrying about being tricked), but if you were sure this was actually the maths, I think you should definitely go for the 1/million shot of saving the billion.

To see why, imagine yourself in the shoes of one of the 1,000,000,010 people whose lives are at risk in this scenario. If you go for the 'safe' option, you have a 10%*1/1,000,000,010 ~= one in a billion chance of being saved. In contrast, if you go for the 'riskier' option, you actually have a one in a million chance - over a thousand times better!

Comment by Larks on Why start a family planning charity? (Founders needed) · 2021-04-02T16:22:39.967Z · EA · GW

I'm quite concerned about your cost-effectiveness analysis. It seems to have been done in a quite naive way that massively biases the conclusions. 

When we do cost-benefit analysis, we need to consider both the costs and the benefits. Yet while your analysis and spreadsheet describe at length the costs of new people (financial, environmental etc.), it does not seem to analyse the benefits at all.

This would not be a big deal if these benefits were small. But they are actually very large!

Firstly, there are a lot of benefits to existing people from larger population sizes:

  • Existing people get the benefit of building relationships with these new people. The experience of being a parent, or a grandparent, is one of the biggest sources of meaning in most people's lives, and this is true even for accidental pregnancies. And certainly when people grow old their grandchildren and great-grandchildren seem to provide a source of both joy and support long after they have ceased participating in much of society.
  • Many things have increasing returns to scale, and so are more efficient with larger populations - e.g. mass transit, factory size, power plant size.
  • Division of Labour - whereby people specialise in one specific area they become more efficient at it. The larger the population, the more specialisation it can support.
  • The new people can become inventors and scientists or artists. Because ideas can be copied with ~ zero cost, these people can provide a benefit to everyone, so the higher the population the better.

It is of course possible that these benefits might be outweighed by the costs outlined in the report. But we cannot simply assume  that this is the case. As far as I can see the spreadsheet does not contain any reference to these considerations, and the report contains only a single throw-away sentence under 'other effects we did not research'.

But even more importantly, there is a huge class of people whose interests are closely tied to future population growth: those future people! As life is good for most people, this is a major advantage. They get to experience the joys of playing and growing and love and all the other good things in the world. I think the vast majority of people in the world live good lives and do not regret being born, so this is a massive positive.

Now, some people adopt ethical views according to which future people do not count. I, along with many academic philosophers and other EAs, reject such views, but they definitely exist, and to the extent that you had credence in such views this would reduce the benefits of population increases.

However, this report does not seem to endorse such a view, because it looks at the animal welfare implications of incremental people's entire lives - even though most of these animals have not yet been born, and indeed might never exist at all. Similarly, it considers the health benefits to newborns who have not yet been conceived. And it talks about the impacts of climate change - impacts which largely fall on future people. Given these discussions of the costs to future people/animals, it seems hard to justify not even mentioning the benefits of existence for future people.

Indeed, such concerns were actually mentioned on the CE website in the past:

One of the main metrics we focus on for this area of research is the number of unintended births averted. The cost-effectiveness of this intervention - as well as the way it is estimated - depends on one’s ethical theory. How much should we care about a person's happiness and suffering (hedonic well-being)? Or should we ultimately value fulfilling what someone wants, whatever that may be (preferences)? If the latter, then how can we make a direct comparison between the preferences of the mother, and the preferences of a child whose birth was averted? Should the preferences of a being that will not come to existence be counted at all? 

Despite this, these considerations did not seem to make it into the report - even in the 'Limitations' or 'Other effects' sections.

(And of course, even if you did reject such totalist views, the instrumental benefits of larger populations for existing generations would remain.)

As such, I would strongly encourage you to re-visit the analysis and try to incorporate both the costs and the benefits of the policy. Given the excellent work CE has done on other issues I would not want to see such an omission risk potentially promoting a negative expected value program.

Comment by Larks on Proposed Longtermist Flag · 2021-03-27T17:49:27.245Z · EA · GW

This is a pretty stylish flag, and I liked the thought process behind it. I must admit that when I first saw it I thought it looked like the sort of flag the villains would have though! Would it be possible to check out how it might look if the central star was blue? I feel like the middle bit instinctually represents 'us', and we want to be the good guys.

Comment by Larks on Proposed Longtermist Flag · 2021-03-25T03:49:48.010Z · EA · GW

It also resembles the light cone which is nice. I would consider putting more sand in the top though, and less in the bottom. Hopefully we have more time left than that!

Comment by Larks on Proposed Longtermist Flag · 2021-03-25T03:47:32.474Z · EA · GW

Well we are working on making the dot bigger but that takes time; realistically we want to have a flag design before the generation ships reach their destinations. 

Comment by Larks on ElizabethE's Shortform · 2021-03-22T18:15:37.355Z · EA · GW

A classic post from Gwern on the subject:

The “expanding circle” historical thesis ignores all instances in which modern ethics narrowed the set of beings to be morally regarded, often backing its exclusion by asserting their non-existence, and thus assumes its conclusion: where the circle is expanded, it’s highlighted as moral ‘progress’, and where it is narrowed, what is outside is simply defined away. When one compares modern with ancient society, the religious differences are striking: almost every single supernatural entity (place, personage, or force) has been excluded from the circle of moral concern, where they used to be huge parts of the circle and one could almost say the entire circle. Further examples include estates, houses, fetuses, prisoners, and graves.

Comment by Larks on Should we consider the sleep loss epidemic an urgent global issue? · 2021-03-21T04:31:00.855Z · EA · GW

I have recently read Why We Sleep (a nice book review) by Matthew Walker PhD (AKA Sleep Diplomat).

The book explains about the benefits of sleeping enough and the negative consequences of not sleeping enough, based on scientific research.

Please keep in mind that this book is exceptionally poor scholarship and should not be relied upon as a scientific source according to Guzey's detailed refutation. Here is the summary, but you should read the whole article:

In the process of reading the book and encountering some extraordinary claims about sleep, I decided to compare the facts it presented with the scientific literature. I found that the book consistently overstates the problem of lack of sleep, sometimes egregiously so. It misrepresents basic sleep research and contradicts its own sources.

In one instance, Walker claims that sleeping less than six or seven hours a night doubles one’s risk of cancer – this is not supported by the scientific evidence (Section 1.1). In another instance, Walker seems to have invented a “fact” that the WHO has declared a sleep loss epidemic (Section 4). In yet another instance, he falsely claims that the National Sleep Foundation recommends 8 hours of sleep per night, and then uses this “fact” to falsely claim that two-thirds of people in developed nations sleep less than the “the recommended eight hours of nightly sleep” (Section 5).

Walker’s book has likely wasted thousands of hours of life and worsened the health of people who read it and took its recommendations at face value (Section 7).

The myths created by the book have spread in the popular culture and are being propagated by Walker and by other scientists in academic research. For example, in 2019, Walker published an academic paper that cited Why We Sleep 4 times just on its first page, meaning that he believes that the book abides by the academic, not the pop-science standards of accuracy (Section 14).

Any book of Why We Sleep’s length is bound to contain some factual errors. Therefore, to avoid potential concerns about cherry-picking the few inaccuracies scattered throughout, in this essay, I’m going to highlight the five most egregious scientific and factual errors Walker makes in Chapter 1 of the book. This chapter contains 10 pages and constitutes less than 4% of the book by the total word count.

Comment by Larks on Against neutrality about creating happy lives · 2021-03-15T18:52:42.788Z · EA · GW

It’s uncontroversial that creating a person with a bad life inflicts on them a serious moral wrong. 

This seems possibly true to me, but not obviously the case, and definitely not uncontroversial. I would guess many people who lived unfortunate lives would nonetheless disagree that their parents inflicted a moral wrong upon them by conceiving them. Similarly, I don't think I have ever heard anyone suggest that children who suffer at the hands of abusers or terrorists were first wronged, not by their tormentor, but by their parents. Even in bleak circumstances, so long as the parents didn't intend to make things bad for the children, I think most people would refrain from such a judgement. 

Comment by Larks on Is Democracy a Fad? · 2021-03-14T18:27:13.471Z · EA · GW

Nice article!

I thought this sub-footnote deserved a bit more prominence:

It’s also conceivable that there will be only be a short window of time between complete automation and the world becoming a pretty much unrecognizably post-human fever dream. In that case, my analysis of the impact of complete automation might only apply to a brief (but still potentially consequential) moment in history.

When we think about what jobs are automatable, it seems that many relatively unintellectual jobs (plumber, nurse, carehome worker) are surprisingly hard to automate, especially those requiring a social component. In contrast, 'AI developer' seems potentially easier to do - we already have things like evolutionary algorithms and ML which are, in some sense, algorithms designing algorithms, and AI developers are often bizarrely interested in accelerating their own obsolescence. So it's not obvious to me that there will be any positive length window of time between full automation and the end of human supremacy. 

Even if there was a short positive window, it's also possible that status quo bias might carry democracy over, as political convergence on locally optimal policy seems to be a slow process at best (e.g. the long coexistence of Parliamentary and Presidential systems, or of North and South Korea).

Comment by Larks on What Makes Outreach to Progressives Hard · 2021-03-14T17:25:46.499Z · EA · GW

EAs tend to reject person-affecting views of population ethics. This, however, has uncomfortable implications for some hot-button issues on the left, like reproductive rights and environmental ethics.


I can see why left wing views on abortion would biased people against totalist views, because they do not want to accept the implication that someone's desire to abort their child could be 'outweighed' by the interests of a possible-person. And I guess totalism would also imply we should have more children, in contradiction to the idea that we should have fewer to protect the environment. But it would naively seem that being concerned about the environment would make you more amenable to longtermist views (as distinct from totalism), because if you don't care about future people then most of the damage from climate change can be ignored.

Comment by Larks on Democracy and Development, a Simple Model · 2021-03-12T03:18:35.878Z · EA · GW

Thanks for writing this, I found it very interesting.

You make it sound like democracies are all relatively quite similar - e.g. basically copies of either the UK or the US political systems, with few refinements. In contrast, the autocracies vary quite broadly, and the autocrats have potentially more room to change their institutional setup. So maybe our binary/linear categorisation is misleading - really there is a narrow cluster in polity-space of democracies, and the rest is all 'autocracy'. Democracy spread very easily during a narrow window of history that favoured this cluster, but our prior should be against this persisting. 

Comment by Larks on A brief explanation of the Myanmar coup · 2021-03-08T05:15:34.402Z · EA · GW

I was under the impression that 'Ejectorate' refereed to people who had some ability to remove leaders, e.g. through coups?

Comment by Larks on [deleted post] 2021-03-07T02:14:10.882Z

I think the answer to your question depends on what your opportunity cost is. Without knowing your situation I would guess this role is better than most others and so would encourage you to take it!

Comment by Larks on Resources On Mental Health And Finding A Therapist · 2021-03-01T16:50:09.464Z · EA · GW

I estimate the true number who would benefit from therapy or mental health self-help resources at some point in their life is closer to two thirds (see footnote)

I was wondering how you estimated this? The footnote provides some data about the supposed prevalence of mental illness, but doesn't provide any evidence that therapy would help these people.

I bring this up because there is a recent paper I saw on twitter which suggests that increasing the diagnosis of mental illness may actually hurt people :

More than one in ten adults in the U.S. and Europe are, at any moment in time, diagnosed with a mental illness. This paper asks whether mental illness is over (or under) diagnosed, by looking at its causal effect on individuals at the margin of diagnosis. We follow all Swedish men born between 1971 and 1983 matched to administrative panel data on health, labor market, wealth and family outcomes to estimate the impact of a mental illness diagnosis on subsequent outcomes. Exploiting the random assignment of 18-year-old men to doctors during military conscription, we find that a mental illness diagnosis for people at the margin increases the future likelihood of death, hospital admittance, being sick from work, and unemployment while lowering the probability of being married. Using a separate identification strategy, we measure the effect of military service on the same set of outcomes to rule out that the effect of diagnosis in our setting is primarily mediated by altering the probability of serving. Our findings are consistent with the potential over-diagnosis of mental illness. [emphasis added]

It would clearly be quite bad for the EA movement to expend resources on incremental mental illness diagnosis if this is actually harmful.

Comment by Larks on Propose and vote on potential tags · 2021-03-01T14:44:13.525Z · EA · GW

One consideration I just thought of, which I do not recall seeing mentioned elsewhere, is that the optional number of tags depends somewhat on the typical tag use case.

  • Clicking on an article's tags to find other related articles
    • As only a small % of tags apply to any given article, and this % will fall as the number of tags increases, article tag spaces will not become too 'busy'. 
    • Hence there should be many tags, so that each article can be tagged as usefully as possible.
  • Clicking on the tag list to find a specific topic
    • There are already so many tags it is hard to find the one you want.
    • This is especially an issue because any given concept often has multiple associated words, so you can't always cntrl-f.
Comment by Larks on Alice Crary's philosophical-institutional critique of EA: "Why one should not be an effective altruist" · 2021-02-25T20:32:56.553Z · EA · GW

This has been previously discussed at some length here.

Comment by Larks on Charges against BitMEX and cofounders · 2021-02-25T17:37:36.124Z · EA · GW

I don’t want this all to read as commentary on Delo or BitMEX specifically. 

Well you did announce the policy change as a comment on an article about Delo!


EA has long included the idea that some ways of making money could create net negative impact even if you donate your earnings, for example 80,000 Hours’ post on Why you should avoid harmful jobs even if you’ll do more good.

I think (?) I may have pointed this out previously, but there are some significant issues with this article. For example, it suggests a $42,000 average social cost of jobs in finance:

We also made a rough estimate of the damage caused by jobs in the financial sector that increase the chance of a financial crisis, and found a figure of $42,000 per year.


First, the estimates are highly uncertain, and only apply to finance jobs on average. If you picked the most harmful jobs in finance, they could be much worse.

But if you follow the source link, you can see that this estimate is actually for only the 10% most harmful jobs:

The US financial sector employs some 6,000,000 people. I am going to guess that the share of people that are involved in activities that could predictably increase systemic financial risk is around 10%. ... 

That would suggest through financial instability each of these 600,000 people lowered other people’s income in the US by $42,000 on average, for each year they worked.

So the average harm is 10x less, i.e. $4,200.

Even then, I think this is quite a poor estimate. It relies on ascribing all the expected costs of financial crisis to financial workers. However, a huge deal of the responsibility should surely be borne by other actors. Depending on your views of the causes of the crisis, some collection of these groups are quite responsible:

  • Politicians who passed regulations like the community reinvestment act.
  • Regulators who focused on solvency over liquidity.
  • The Federal Reserve for excessively tight monetary policy.
  • Individual borrowers who knowingly mis-stated their ability to repay.
  • Investors in ABS CLOs or Prime MMFs who did not do their due diligence.
  • Academics who over-emphasised normal distributions and ignored skew/kurtosis.

Furthermore, it does not assign any monetary value to the positive aspects of finance, even though these are probably very large:

  • Banks allow people to save money without being afraid it will be stolen by burglars.
  • Credit card networks allow us to purchase things without needing to carry cash.
  • Paypal allows us to buy things from merchants who are not physically nearby.
  • The stock market provides signals to investors about where would be useful to invest their money.
  • Companies can raise money in order to grow much more quickly than they would otherwise.
  • The bond market allows governments to borrow money to finance additional spending during recessions and pandemics.
  • (etc.)

Similarly, the article suggests that being a Tobacco CEO is unacceptable, linking to this analysis. However, I think the fermi calculation involved in this estimate was quite far off, as I explained here:

However, I think that while 80,000 Hours substantially improved on the analysis in the original paper, they omit a number of factors. Unfortunately, almost all of these factors seem to pull in the same direction, causing them to over-estimate the amount of harm done by a tobacco CEO. These include both over-estimating the direct harm done and under-estimating the benefits donating your income would cause.

At the time  Rob suggested he would think more on the issue, but to my knowledge the analysis was never updated.

Comment by Larks on Google's ethics is alarming · 2021-02-25T15:29:51.517Z · EA · GW

This video seems quite sensationalist, and in many places the argument seems like a stretch. For example, you say that Timnit was fired, but the only evidence of this seems to be that she claims this is the case - in contrast, Google says she offered to resign:

Timnit responded with an email requiring that a number of conditions be met in order for her to continue working at Google, including revealing the identities of every person who Megan and I had spoken to and consulted as part of the review of the paper and the exact feedback. Timnit wrote that if we didn’t meet these demands, she would leave Google and work on an end date. We accept and respect her decision to resign from Google.

Even if you thought Google is mistaken/lying I think you should at least mention this.

I would also encourage you to submit text rather than video to the forum in the future. In many cases you mentioned things that I would like to respond to - for example, the idea that Google's responsibility for searches linking to bad medical advice is similar to Boeing's responsibility for plane crashes - but it is very hard to do so without text to easily search, analyse and quote.

Comment by Larks on Why EA groups should not use “Effective Altruism” in their name. · 2021-02-22T16:07:42.260Z · EA · GW

This was one of the reasons why we picked GWWC and 80k as names in the early days, rather than leading with effective altruism.

I was curious about what you meant by this? I recall the term 'Effective Altruism' not being determined until 2011 or 2012, long after the names 'GWWC' and '80k' were chosen.

Comment by Larks on Why EA groups should not use “Effective Altruism” in their name. · 2021-02-20T17:56:50.534Z · EA · GW

Thanks for writing this, I thought it was an interesting post - the point about translation awkwardness was especially new to me and seems pretty credible.

One consideration against I might consider would be that of organisational value drift. If you have invested a lot of effort into building an organisation, it makes sense to want to ensure it stays closely affiliated with the EA movement. If it has 'Effective Altruism' in the title, that is a powerful anchor. In contrast, if you're called the 'Positive Impact Club' or somesuch, I can easily imagine someone  arguing, "Well might not be EA but recycling awareness still has positive impact, so we should still promote it!" In particular, I think a strong name anchor makes it easier to resist social pressure from other activist groups on campus to promote their thing.

Comment by Larks on Effective Altruism Funds Project Updates · 2021-02-20T05:58:34.649Z · EA · GW

Obviously, I do still think there is a place for considering something more like "variance of impact", but I don't actually think that that dimension has played a large role in people's historical reactions to grants we have made, and I don't expect it to matter too much in the future.


Relatedly, I don't recall anyone pointing out that funding a large number of 'risky' individuals, instead of a small number of 'safe' organisations, might be less risky (in the sense of lower variance), because the individual risks are largely independent, so you get a lot of portfolio diversification.

Comment by Larks on Alternatives to donor lotteries · 2021-02-19T17:28:09.848Z · EA · GW

I don't think the research is much evidence here. The whole point of the donor lottery is that the winner can justify doing a lot more research. This would be the case even if they hated the other entrants. 

You're right that they wouldn't necessarily have to share that research, but many people enjoy posting on the forum anyway. Previously Jonas has been at pains to clarify that such reports are not required.

Comment by Larks on Charges against BitMEX and cofounders · 2021-02-18T17:26:44.314Z · EA · GW

Is the idea that such controls, had they been implemented in the past, would have prevented you from accepting Delo's donations? 

Also, I am curious to see CEA's cost-benefit analysis behind this decision. Naively this seems like incurring a cost (staff time, consultant fees, lawyer fees, annoy donors) in order to reduce a benefit (donations). Based on my cursory research (talking to a lawyer and reading this) I couldn't work out if this was actually legally required given CEA's situation, though it does seem to be reasonably common.

Comment by Larks on Alternatives to donor lotteries · 2021-02-16T19:02:43.289Z · EA · GW

At the moment we offer several options for donors who wish to take advantage of more research than they can justify doing themselves:

  1. EA Funds, where they money will be allocated basically on semi-professional allocators who have been chosen (ideally) for their knowledge and competence.
  2. Donor lotteries, where the expected money you donate is conserved, but your counterfactual research time is concentrated.
  3. Give Directly, to give power to the poorest people in the world and take advantage of their local knowledge of their needs and opportunities.

Many of your suggestions seems essentially like a mixture of these three, but without really the advantages of either. For example, Random Donor Pooling (2) or Reverse-Donation-Weighted is basically like the EA funds except with a less rigourous hiring procedure. Unless you think the EA Funds search for qualified candidates adds negative value I'm not sure why this would be desirable. Alternatively, it's basically like the donor lottery, except it punishes you for donating larger amounts of money; I don't think many people would want to donate more than the minimum if it means reducing their expected influence in favour of not-obviously-more-qualified people. Or it's like GiveDirectly, except you're empowering smaller donors who are still relatively quite well off. Taken all together, I'm not sure why someone would prefer one of these alternatively-weighted lotteries over some combination of these already-existing options.

I also disagree with your suggestion that biasing donor lotteries towards smaller donors would necessarily widen the search space:

Randomness expands the search space by increasing the chances of people not typically selected to allocate large pots of money

The search space covers the various possibilities that you are examining. For a typical well-informed EA this might include a few dozen charities in global health, animal rights and existential risk reduction. If we select some unusual person, who spends all their time researching charities to promote local museums in rural Ireland, this doesn't really expand our search space unless that person also researchers the standard charities in as much as depth as a typical EA donor would. Otherwise we've just replaced it with a different search space, one I suspect most EAs would consider inferior.

Worse, it seems quite plausible that if EAs donated a lot of money to a pot whose direction would be determined on a per-capita basis (rather than proportionally) that some other community would decide to 'raid' it by having a lot of people donate the minimum amount and then divert the money to their own ends.

Comment by Larks on Population Size/Growth & Reproductive Choice: Highly effective, synergetic & neglected · 2021-02-14T18:32:09.046Z · EA · GW

At least in the US women have been having fewer children than they want for many decades:

As a result, the gap between the number of children that women say they want to have (2.7) and the number of children they will probably actually have (1.8) has risen to the highest level in 40 years.

Comment by Larks on Population Size/Growth & Reproductive Choice: Highly effective, synergetic & neglected · 2021-02-14T18:29:08.608Z · EA · GW

I think this post is very mistaken.

There are some relatively minor issues. For example, the methodology used in Wynes & Nicholas (from Murtaugh & Schlax) to compare the climate change impacts of less driving with fewer children involves comparing temporary changes in automobile usage with an infinite duration population change, which seems like an obviously unfair comparison. Indeed, there is nothing in this methodology to prevent each individual child from having an infinite impact on climate change... except that the authors arbitrarily prevented this, probably realising it would show their results were absurd:

Some lineages persist indefinitely, in which case a pre-specified time limit terminates the simulation.

Similarly, you suggest that 'unintended' pregnancy means women are 'forced to bear children against their will'. But this is incorrect. Many couples are not planning on conceiving but are nonetheless happy to be blessed with a child! Indeed, the Wikipedia article you quote makes this clear in the sentence  immediately preceding your quote:

Terming a pregnancy "unintended" does not indicate whether or not a pregnancy is welcomed

More importantly, when doing cost-benefit analysis, it is important to analyse both the costs and the benefits. Here, you analyse in detail the costs of population growth - but with no mention at all of the benefits. Yet these benefits are very large:

  • Population growth means more people will get to experience the joys of living. They will get to learn, play, make meaningful relationships and fall in love. They can play with lego, read books and enjoy the sunset. I think for most people life is good, so this is a huge deal!
  • Children benefit existing generations who will enjoy being parents and grandparents. Anecdotally grandchildren seem to be one of the primary joys in most grandparents' lives, and people generally rank their relationship with their children as one of the most meaningful parts of their life.
  • Larger populations allow more specialisation and division of labour, which are key components of economic growth, as well as other economies of scale.
  • Larger populations mean more artists and scientists, whose inventions and discoveries can then be enjoyed by everyone at essentially zero marginal cost.
  • More prosaically, we need future generations to keep social security and other retirement systems working.

Given that the post has ignored all positives of population growth, it is unsurprising it found that population growth to be negative. Indeed, according to the analysis here - which treats people as entirely a burden, with no benefits whatsoever, whether to themselves or others - encouraging the deaths of existing people, by spreading the pandemic for example, could also appear a very positive policy! 

Comment by Larks on Blameworthiness for Avoidable Psychological Harms · 2021-02-10T04:33:14.544Z · EA · GW

Great article, very logical approach.

We might want to consider the idea that the 'victims' of ideology-driven psychological harms might be blameworthy, even if they are not the least-cost avoider any more. It might be the case that the cheapest way to avoid the harm is to not adopt the ideology in the first place but, having adopted it, it is very hard to avoid subsequent harm, and it cannot easily be un-adopted. In this case I think we would not want to encourage people to adopt such an ideology, so we might want to hold them responsible after the fact. (This is implicitly covered in your piece but I thought I'd make it explicit).

Comment by Larks on In diversity lies epistemic strength · 2021-02-08T02:36:24.927Z · EA · GW

As we cannot measure the diversity of perspectives of a person directly, our best proxy for it is demographic diversity, as our life shapes our assumptions.

This seems like quite a key premise to your argument, but you don't seem to spend much time arguing for its plausibility; indeed, it seems quite likely false. Could we not simply ask someone their views on a variety of subjects? If we know what their views are on Sports, Reincarnation, the Holy Roman Empire, Two Dogmas of Empiricism, Abortion, Pokemon, the Axiom of Choice, MIRI, and Socks with Sandles, I feel like we could have a pretty good sense of whether they add a new perspective. 

Similarly with education and career; I would in general expect more perspective diversity from a group consisting of an economist, a biologist, a nurse, a poet, a cop and a prostitute, even if they were all the same race, age and sex, than I would from a demographically diverse group of harvard-educated lawyers.

Comment by Larks on Retention in EA - Part II: Possible Projects · 2021-02-06T03:39:54.879Z · EA · GW

There are a lot of Facebook groups exist because there is an obvious niche but don't really have enough quality content to justify themselves, but the Parents group is actually pretty good; I definitely recommend if you're a parent and want to chat with similar other people.

Comment by Larks on Policy question: basic income after public service? · 2021-02-03T01:01:40.046Z · EA · GW

The volunteer labor could be applied to domestic infrastructure as in the CCC

Unfortunately the limiting factor on US domestic infrastructure isn't typically labour but regulatory. Major infrastructure projects have to go through literally years of applications and hearings to get planning permission, with expensive concessions required and no guarantee of success. This occurs both for very small projects  - e.g. turning a disused parking lot into a house in the bay area - and large ones - the KeyStone XL pipeline, for example, was originally proposed in 2008... and 12 years later, after exhaustive environmental review, was blocked by the government.

Comment by Larks on DanielFilan's Shortform · 2021-01-22T18:31:44.599Z · EA · GW

Seems plausible. Presumably if some crime is deterred by these rules, which would leave the $3bn an under-estimate of the benefit. On the other hand, without the rules we might see more innovation in financial services, which would suggest the $300bn an under-estimate of the costs.

Unfortunately I think it is very unlikely we could make any progress in this regard, as governments do not like giving up power, and the proximate victims are not viewed sympathetically, even if the true incidence of the costs is broad.

There have been attempts in the past to reform, as they particular harm poor immigrants trying to send cash home, but as far as I am aware these attempts have been almost entirely unsuccessful.

Comment by Larks on The Folly of "EAs Should" · 2021-01-10T18:10:46.999Z · EA · GW

we need to stop saying "don't donate to your local theatre" ... because actually [that is a] bad advice a lot of the time

I'm surprised you would say this - I would expect that not donating to a local theatre would have basically no negative effects for most people. I can see an argument for phrasing it more delicately - e.g. "I wouldn't donate to a local theatre because I don't think it will really help make the world a better place" - but I would be very surprised if it was actually bad advice. Most people who stop donating to a charity suffer essentially no negative consequences from doing so.

Comment by Larks on AMA: Elizabeth Edwards-Appell, former State Representative · 2021-01-10T17:59:34.300Z · EA · GW

Do you think being an EA, believing EA things, or being identified as one, represents any disadvantage (or advantage) in running for office?

Comment by Larks on EA and the Possible Decline of the US: Very Rough Thoughts · 2021-01-08T16:07:39.302Z · EA · GW

Peaceful Scenarios

Collapse need not be violent or tumultuous. For example, there could be a legal agreement to split the country into different independent countries. Although difficult to imagine, the US could also enter into a treaty with an independent country that would integrate the two, fundamentally altering each.

These seem like quite different scenarios to the others discussed. If the US agreed to let California become independent, or annexed Canada, I would not expect any threat to nuclear security, or AI lab integrity, or drastic loss of life in the process. Annexing Canada could even potentially help continue US international hegemony through increased population and GDP, though it might be bad in other ways.