Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Be wary of enacting norms you think are unethical, published by RobBensinger on January 18, 2023 on The Effective Altruism Forum.

To my eye, a lot of EAs seem to under-appreciate the extent to which your response to a crisis isn't just a reaction to an existing, fixed set of societal norms. The act of choosing a response is the act of creating a norm.

You're helping bring into existence a particular version of EA, a particular version of academia and intellectual life, and a particular world-at-large. This is particularly true insofar as EA is a widely-paid-attention-to influencer or thought leader in intellectual society, which I think it (weakly) is.

It's possible to overestimate your influence, but my impression is that most EAs are currently underestimating it. Hopefully this post can at least bring to your attention the hypothesis that your actions matter for determining the norms going forward, even if you don't currently think you have enough evidence to believe that hypothesis.

If you want the culture to be a certain way, I think it's worth taking the time to flesh out for yourself what the details would look like, and find ways to test whether there are any ways to effect that norm, or to move in the direction that seems best to you.

Anchor more to what you actually think is ethical, to what you think is kind, to what you think is honorable, to what you think is important and worth protecting. If you think your peers aren't living up to your highest principles, then don't give up on your principles.

(And maybe don't give up on your peers, or maybe do, depending on what seems right to you.)

Don't ignore the current set of norms you see; but be wary of willing bad outcomes into being via self-fulfilling prophecies.

Be dissatisfied, if the world doesn't already look like the vision of a good, wholesome, virtuous world you can picture in your head. Because as someone who feels optimistic about EAs' capacities to do what's right, I want to see more people fighting for their personal visions of what that involves.

Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Book critique of Effective Altruism, published by Manuel Del Río Rodríguez on January 17, 2023 on The Effective Altruism Forum.

Hello there! First post in the forum, so I apologize in advance for the probable mistakes and overall clumsiness. I have checked the forum writing guidelines but am pretty sure there's a high probability of my screwing up something or somewhere, so if that proves to be the case, "I am sure you have a waste basket handy".

The case is, I was just checking Amazon today for some books on Effective Altruism with which to supplement the digital EA Handbook I am reading when I found this volume which will be made available exactly a month from now: The Good It Promises, the Harm It Does: Critical Essays on Effective Altruism, by Carol J. Adams, Lori Gruen and Alice Crary. I haven't seen any post mentioning it, and I thought it might be interesting to share.

As stated, the book hasn't been published yet, but one can look inside. I have been browsing the introduction, and in line with its title, it is pretty harsh in its wording. For example, from page xxv of the introduction:

"In addition to describing how EA can harm animals and humans, the book contains critical studies of EA's philosophical assumptions and critical studies of organizations that set out to realize them. It invites readers to recognize EA as an alluring and extremely pernicious ideology, and it traces out a number of mutually reinforcing strategies for submitting this ideology for criticism".

From the tone of the introduction I can suppose the general tone will be pretty scathing and hostile, as well as its general orientation. Still, I imagine the arguments it makes will profit from some attention, discussion and counterargument when it comes out.

Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Building a transhumanist future: 15 years of hplusroadmap, now Discord, published by kanzure on January 17, 2023 on LessWrong.

What is hplusroadmap?

We are a group chat, started 15 years ago, focusing on transhumanism, writing code, and hacking on biology. We do mad science. We build. We also have some funding available for sufficiently interesting projects.

Read more background here:

We have some other items on the wiki, like the internet's largest collection of potential human genetic modifications:/

Topics

Things we tend to focus around:

do-it-yourself biology, genetic engineering, gene therapy, biohacking, synthetic biology

cryonics

molecular nanotechnology

DNA synthesis machine development

brain emulation, brain uploading, computational neuroscience

nootropics

molecular biology, polymerase, enzyme engineering, protein engineering, fusion proteins, directed evolution, gene-editing

open-source hardware development

self-replicating machines

robots, mechanical automation, machining, electronics

space colonization, habitats, mining, communication, transportation

nutrition, bodybuilding

computer architecture, supercomputing

open-source software development

transhumanism

life extension, anti-aging, longevity research

cryptography, bitcoin technology development

artificial general intelligence, software AI

breeding and selection projects

Who might be interested?

Anyone who wants to build the future by contributing to cool projects, proposing new projects, and sorting through the enormity of technologies now available.

Now on Discord

Today I am announcing the availability of a Discord bridge for hplusroadmap:

Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Calculating how much small donors funge with money that will never be spent, published by Tristan Cook on January 16, 2023 on The Effective Altruism Forum.

Epistemic status: Confident that the effect is real, though likely smaller than suggested by the toy-model.

Summary

Small donors should discount the cost effectiveness of their donations to interventions above a large funder’s bar if

they expect the large funder not to have spent all their capital by the time of AGI’s arrival

their donation to interventions above the large funder’s bar funges with the large funder.

In this post I describe a toy model to calculate how much to discount due to this effect.

I apply the model to a guess of Open Philanthropy’s spending on Global Health and Development (GHD) with Metaculus’ AGI timelines (25% by 2029, 50% by 2039). The model implies that small donors should consider interventions above OP's GHD bar, e.g. GiveWell's top charities, are only 55% as cost effective as the small donors first thought. For shorter AGI timelines (25% by 2027, 50% by 2030) this factor is around 35%.

I use OP's GHD spending as an example because of their clarity around funding rate and bar for interventions. This discount factor would be larger if one funges with 'patient' philanthropic funds (such as The Patient Philanthropy Fund).

This effect is a corollary of the result that most donor's AGI timelines (e.g. deferral to Metaculus) imply that the community spend at a greater rate. When a small donor funges with a large donor (and saves them spending themselves), the community's spending rate is effectively lowered (compared to when the small donor does not funge).

This effect occurs when a small donor has shorter timelines than a large funder, or the large funder is not spending at a sufficiently high rate. In the latter case, small donors - by donating to interventions below the large funder's bar - are effectively correcting the community's implicit bar for funding.

Toy model

Suppose you have the choice of donating to one of two interventions, A which gives a utils per $, or B, which gives b utils per $. Suppose further that the available interventions remain the same every year and that both have room for funding this year.

A large funder F will ensure that A is fully funded this year, so if you donate $1 to A, then F, effectively, has $1 more to donate in the future. I suppose that F only ever donates to (opportunities as good as) A.

I suppose that F's capital decreases by some constant amount f times their initial capital each year. This means that F will have no assets in 1/f years from now.

Supposing AGI arrives t years from now, then F will have spent fraction min(ft,1) of their current capital on A.

Accounting for this funging and assuming AGI arrives at time t, the cost effectiveness of your donation to A is then min(ft,1)a utils per $. Then if b>min(ft,1)a, marginal spending by small donors on B is more cost effective than on A.

By considering distributions of AGI's arrival time t and the large funder's funding rate f we can get a distribution of this multiplier.

Plugging in numbers

I take

The large funder F to be Open Philanthropy’s Global Health and Wellbeing spending on Global Health and Development and intervention A to be Givewell’s recommendations.

I take 1/f, the expected time until OP's funds dedicated to GHD are depleted to be distributed Normal(20,20) bounded below by 5.

I take AGI timelines to be an approximation those on Metaculus.

These distributions on AGI timelines and 1/f give the following distribution of the funging multiplier (reproducible here).

The ratio of cost effectiveness between GiveWell's recommendations and GiveDirectly, a/b, is approximately 7-8 and so small donors should give to interventions in the (5, 7)x GiveDirectly range.

For donors with shorter timeli...

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Recursive Middle Manager Hell, published by Raemon on January 17, 2023 on The Effective Altruism Forum.

Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: On AI and Interest Rates, published by Zvi on January 17, 2023 on LessWrong.

Note, To say it up front: None of this, or anything ever on this blog, is investment advice.

This post on the EA form, highlighted by Tyler Cowen, points out that

Transformative aligned AI would increase real interest rates. A lot.

Transformative unaligned AI would increase real interest rates. A lot.

Markets continue to have low interest rates not pricing this in.

Therefore, either (A) AI not only isn’t about to end the world it also is not about to do anything economically that powerful or (B) the Efficient Market Hypothesis (EMH) is false.

In conclusion:

There is no one in the market to price this in. Beyond avoiding wrong-way investments and locking in low-interest loans, the few who buy the hypothesis have better trades to make than betting on future interest rates.

Nor is the market meaningfully considering and then rejecting the hypothesis.

The pre-Covid-pandemic market did not peak until February 20, 2020.

Thank you for coming to my TED Talk.

Treasure Everywhere

The OP claims that option B implies there is treasure everywhere.

No. A lot of comments attempt to explain.

Even the OP’s proposed trade would, under favorable assumptions, only return about 10%. That’s at best a highly mediocre trade, in its full context.

At a minimum, again assuming you buy the hypothesis, buying a portfolio of companies that would profit from transformative AI is a superior play. So is shorting a portfolio of assets that would not profit from this scenario but do depend on low interest rates to justify their valuations, if you want that level of risk. If you’re so sure interest rates will go up, why settle for ~10% returns when there are Bitcoins (and many other assets including stocks) you can short?

Don’t forget to consider the marginal utility of capital in various scenarios.

Many traders at Jane Street Capital are aware of the potential of AI, or at least have considered strong arguments for it. On the time horizon of almost all their trades, this makes almost zero difference.

If We Will All Go Together When We Go

Eliezer Yudkowsky points out that taking advantage of knowing that the world will be ending is difficult. How do you collect? If you did collect, how would you spend it?

He then goes into more detail in this comment and then this second attempt, too long to quote here in full, but recommended if you remain confused or unconvinced here.

That does not mean none of your financial choices change in such a world.

While there are not amazing winning moves, there are plenty of losing moves you can avoid if you are sufficiently confident in doom. For example:

You can spend down capital and other resources.

You don’t have to save for retirement.

You can avoid making long term or illiquid investments.

You can borrow money rather than loaning out money, especially locking in low long term interest rates. This is more like a win. You still can’t that win big this way, even if you’re fully confident you never have to pay it back, because there is a limit to how much money you can borrow over a 10-year horizon.

You can also alter other life choices, including whether to have children. Unless your certainty level is far higher than I believe is justified, I think this is almost always a mistake. Children and ordinary life well lived are great joys, keep you sane and grounded, give you something to protect, improve how you are treated. Often they teach you something. The more extremely you modify your choices, the more you really will be doomed even in relatively ‘normal’ futures, and dread of this will impact your experience and life now.

Yes, there are perhaps exceptions to this where your focus really is that important, but if you are not mentioned in this post by name you probably don’t count for that.

C...

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Introducing Lafiya Nigeria, published by Klau Chmielowska on January 17, 2023 on The Effective Altruism Forum.

Reducing maternal mortality through informed family planning

In June 2021, we founded Lafiya Nigeria, a non-profit organisation that works toward ending maternal mortality in Nigeria by widening the access and information about family planning. TL; DR Introduction to our organisation in a 3-min video

This post describes (I) the challenge we aim to solve, (II) our approach, (III), our traction, (IV) our value-add, (V) our plans, (VI) how you can get involved in our initative.

(I) The challenge

In low and middle-income countries, women are dying from giving life. Nearly 300,000 women and girls are dying from pregnancy-related complications each year, according to the Guttmacher Institute. Other health complications such as obstetric fistula, postpartum anemia, and postnatal depression are also key health burdens borne by pregnant women.

Reducing the number of unintended pregnancies is an effective means of reducing health burdens for mothers and newborns. Despite a significant number of women in these countries wanting to avoid pregnancy, many are not using modern contraceptives, resulting in 85 million unintended pregnancies per year. If all women with unmet need were provided access to and used modern contraceptives, the Lancet estimated that maternal deaths globally would drop 44%. An estimated 70,000 maternal deaths could be prevented each year, with 441,000 new-born deaths also averted. Additionally, the Guttmacher Institute estimates that every dollar spent on contraceptive services beyond the current level would reduce the cost of pregnancy-related and newborn care by three dollars. The Copenhagen Consensus also found that every dollar spent on access to modern contraception leads to 120 dollars of social, economic, and environmental benefits.

Access to family planning is beyond a health issue: its dividends are seen also in positive effects on education, income generation, and children's welfare. A study in Indonesia found that providing access to family planning was three times more effective than improving school quality in keeping girls in school an extra year. Research in Colombia found that girls with access to family planning clinics were 7% more likely to participate in the formal workforce as adults. Long-term studies have also shown that providing access to family planning programs can lead to improved college completion rates of children and higher family incomes decades later. These spillover effects are difficult to measure and are often neglected in traditional cost-effectiveness analyses.

In recent years, there has been an increase in the use of modern contraceptives in countries like Nigeria but there has also been an increase in the unmet need. From 2012 to 2019, the portion of women using contraceptives in Nigeria rose from 11.2% to 14.2%, while the unmet need also rose from 22.4% to 23.7%. These figures are greatly exaggerated

Focus: Nigeria

Lafiya Nigeria focuses on rural and underserved regions of northern Nigeria

Nigeria has >45M women of child-bearing age, and 65% have unmet contraceptive needs (IHME), resulting in around 40,000 maternal deaths a year.

In Nigeria, over 83% of women had not used any contraceptive methods for family prevention in 2018. This rate reached 96% among women without any education. This staggering gap in health provision results in maternal and infant deaths. In Nigeria, over 40,000 women die each year from pregnancy-related issues. The loss of life does end with mother, either. Over one million children under the age of five also die as a result of losing their mothers to pregnancy delivery complications.

In our pilot region, Jigawa, more than 98% of women have no prior contraceptive use due to stockouts ...

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Language models can generate superior text compared to their input, published by ChristianKl on January 17, 2023 on LessWrong.

There’s a frequent misconception that assumes that a large language model will never achieve superhuman text creation ability because such models try to create texts that are maximally unsurprising. This article will explain why that assumption is wrong.

In 1906, Sir Francis Galton conducted an experiment at a fair, where he asked fair-goers to guess the weight of an ox in a weight-judging competition. The median of 787 guesses was 1,207 pounds, while the actual weight of the ox was 1,198 pounds. The error in making guesses was a result of a combination of systematic bias and random noise. The fair-goers, having knowledge of oxen, had no bias in their guesses, thus the error was entirely due to random noise. By polling the 787 guesses, Galton averaged out the random noise of each individual guess.

This phenomenon was coined wisdom of the crowd. In areas where reasoning errors are mostly random noise, crowds are smarter than individual members of the crowd. By training on large data sets, large language models can access the wisdom of the crowd. The ceiling of the ability of a large language model is the wisdom of the crowd instead of the wisdom of individual members of the crowd.

The fact that each word of a text is massively unsurprising based on preceding words in the text does not imply that the text overall would be massively unsurprising. If you have a text you can calculate for every word in the text the likelihood (Ltext) how likely it would follow the preceding words in the text. You can also calculate the likelihood (Lideal) of the most likely word that would follow the preceding text.

Lideal - Ltext is noise. If you look at a given text you can calculate the average of the noise for each word. A well-trained large language model is able to produce texts with a lot less noise than the average of the text in its training corpus.

For further reading, Kahneman wrote Noise: A Flaw in Human Judgment which goes into more detail on how a machine learning model can eliminate noise and thus make better decisions than the average of its training data.

Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Posts from 2022 you thought were valuable (and underrated), published by Lizka on January 17, 2023 on The Effective Altruism Forum.

Forum Wrapped showed you what you upvoted and strong-upvoted in 2022. We also gave you the chance to mark some posts as “most valuable.”

I’m sharing which posts were marked as “most valuable” by most people, and which posts were most underrated by their karma score (relative to the number of “most valuable” votes).

Please note that this is a very rough list; relatively few people marked posts as "most valuable," and I imagine that those who did, didn't do it very carefully or comprehensively. And there are various biases in the data (like the fact that we showed the list in order of karma).

Before you continue, consider looking back (if you haven’t done that yet).

You can look at your Forum Wrapped, consider scrolling through high-rated (adjusted) posts from 2022, or explore posts on your favorite topics.

(You can still mark posts as “most valuable” — we might revisit these numbers later, and you might find it useful to have a record of this for yourself.)

If you want to explore more content, you could look at What are the most underrated posts & comments of 2022, according to you? (thread), my comment on the Forum Wrapped announcement, curated posts from 2022, and older content like the Results from the First Decade Review and the Forum Digest Classics.

Which posts did Forum users think were most valuable?

(Note that we ordered posts in "wrapped" by karma score, meaning higher-karma posts might be artificially over-rated.)

VotesAuthor(s)TitleASB, ecaConcrete Biosecurity Projects (some of which could be big)Julia WisePower dynamics between people in EATheo HawkingBad Omens in Current Community BuildingScott Alexander"Long-Termism" vs. "Existential Risk"George RosenfeldFree-spending EA might be a big problem for optics and epistemicsHaydn BelfieldAre you really in a race? The Cautionary Tales of Szilárd and EllsbergHolden KarnofskyEA is about maximization, and maximization is perilousSimon MStrongMinds should not be a top-rated charity (yet)Fods12The FTX crisis highlights a deeper cultural problem within EA - we don't sufficiently value good governanceMathiasKBSnakebites kill 100,000 people every year, here's what you should knowWilliam MacAskillEA and the current funding situationLinchSome unfun lessons I learned as a junior grantmakerWill Bradshaw, Mike McLaren, Anjali GopalAnnouncing the Nucleic Acid Observatory project for early detection of catastrophic biothreatsThomas KwaEffectiveness is a

Conjunction of MultipliersKarolina Sarek, JoeyPresenting: 2022 Incubated Charities (Charity Entrepreneurship)Nuño SempereA Critical Review of Open Philanthropy’s Bet On Criminal Justice ReformHolden KarnofskyMy takes on the FTX situation will (mostly) be cold, not hotPeter McLaughlinGetting on a different train: can Effective Altruism avoid collapsing into absurdity?Katja GraceLet’s think about slowing down AIHolden KarnofskyImportant, actionable research questions for the most important centuryHolden KarnofskySome comments on recent FTX-related eventsMaya DI’m a 22-year-old woman involved in Effective Altruism. I’m sad, disappointed, and scared.AjeyaWithout specific countermeasures, the easiest path to transformative AI likely leads to AI takeover

VotesAuthor(s)TitleASB, ecaConcrete Biosecurity Projects (some of which could be big)Julia WisePower dynamics between people in EATheo HawkingBad Omens in Current Community BuildingScott Alexander"Long-Termism" vs. "Existential Risk"George RosenfeldFree-spending EA might be a big problem for optics and epistemicsHaydn BelfieldAre you really in a race?

The Cautionary Tales of Szilárd and EllsbergHolden KarnofskyEA is about maximization, and maximization is perilousSimon MStrongMinds should not be a top-...

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: How many people are working (directly) on reducing existential risk from AI?, published by Benjamin Hilton on January 17, 2023 on The Effective Altruism Forum.

Summary

I've updated my estimate of the number of FTE (full-time equivalent) working (directly) on reducing existential risks from AI from 300 FTE to 400 FTE.

Below I've pasted some slightly edited excepts of the relevant sections of the 80,000 Hours profile on preventing an AI-related catastrophe.

New 80,000 Hours estimate of the number of people working on reducing AI risk

Neglectedness estimate

We estimate there are around 400 people around the world working directly on reducing the chances of an AI-related existential catastrophe (with a 90% confidence interval ranging between 200 and 1,000). Of these, about three quarters are working on technical AI safety research, with the rest split between strategy (and other governance) research and advocacy. We think there are around 800 people working in complementary roles, but we’re highly uncertain about this estimate.

Footnote on methodology

It’s difficult to estimate this number.

Ideally we want to estimate the number of FTE (“full-time equivalent“) working on the problem of reducing existential risks from AI.

But there are lots of ambiguities around what counts as working on the issue. So I tried to use the following guidelines in my estimates:

I didn’t include people who might think of themselves on a career path that is building towards a role preventing an AI-related catastrophe, but who are currently skilling up rather than working directly on the problem.

I included researchers, engineers, and other staff that seem to work directly on technical AI safety research or AI strategy and governance. But there’s an uncertain boundary between these people and others who I chose not to include. For example, I didn’t include machine learning engineers whose role is building AI systems that might be used for safety research but aren’t primarily designed for that purpose.

I only included time spent on work that seems related to reducing the potentially existential risks from AI, like those discussed in this article. Lots of wider AI safety and AI ethics work focuses on reducing other risks from AI seems relevant to reducing existential risks – this ‘indirect’ work makes this estimate difficult. I decided not to include indirect work on reducing the risks of an AI-related catastrophe (see our problem framework for more).

Relatedly, I didn’t include people working on other problems that might indirectly affect the chances of an AI-related catastrophe, such as epistemics and improving institutional decision-making, reducing the chances of great power conflict, or building effective altruism.

With those decisions made, I estimated this in three different ways.

First, for each organisation in the AI Watch database, I estimated the number of FTE working directly on reducing existential risks from AI. I did this by looking at the number of staff listed at each organisation, both in total and in 2022, as well as the number of researchers listed at each organisation. Overall I estimated that there were 76 to 536 FTE working on technical AI safety (90% confidence), with a mean of 196 FTE. I estimated that there were 51 to 359 FTE working on AI governance and strategy (90% confidence), with a mean of 151 FTE. There’s a lot of subjective judgement in these estimates because of the ambiguities above. The estimates could be too low if AI Watch is missing data on some organisations, or too high if the data counts people more than once or includes people who no longer work in the area.

Second, I adapted the methodology used by Gavin Leech’s estimate of the number of people working on reducing existential risks from AI. I split the organisations in Leech’s estimate into technical sa...