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: EA will likely get more attention soon, published by Julia Wise on May 12, 2022 on The Effective Altruism Forum.

As EA-aligned foundations and projects direct more money, EA ideas continue to gain traction, What We Owe the Future comes out, etc, there’s naturally going to be more attention on EA soon. That attention will likely range from enthusiasm to thoughtful criticism to . . . less thoughtful criticism. If you’ve been involved in EA for a while, this transition might be a bit disorienting.

I’m writing this post on behalf of some staff (at CEA, Forethought Foundation, and Open Philanthropy) who are working on communications for EA as a movement. We’re trying to prepare for increased attention, plan the best ways to communicate complex ideas succinctly, and increase the chance that EA will be portrayed accurately and thoughtfully.

Doing more proactive communications work

For the last several years, most EA organizations did little or no pursuit of media coverage. CEA’s advice on talking to journalists was (and is) mostly cautionary. I think there have been good reasons for that — engaging with media is only worth doing if you’re going to do it well, and a lot of EA projects don’t have this as their top priority.

While this may have made sense for each individual organization, as a result, we’ve missed out on opportunities to convey the good ideas and work coming from EA. There’s also confusion out there about what EA is even about. Ideally more people would have a clearer sense of what EA is, so they can agree or disagree with an accurate representation of EA and not with a misconception.

Several EA organizations are working together with a communications advising firm to answer questions like

Who are key audiences we especially want to reach?

How do these audiences currently see EA?

What are the best ways to reach these audiences?

What EA ideas are especially important to convey?

Connecting EA projects with journalists

I’ve been writing to EA organizations and projects to see if they have recent success stories that journalists might be interested in covering. If I’ve missed your project and you’d like some help connecting with journalists who might want to cover your work, please do get in touch! [email protected]

As before, if a journalist reaches out to you, we suggest you look through our guide on responding to journalists.

FAQs

If I see a new media piece about EA, who should I flag it to?

Feel free to flag things to [email protected] and we’ll talk with our advisors about whether some kind of response makes sense. We’ll likely have heard about pieces in large-scale publications, but might miss coverage of EA in publications in languages other than English, or targeted readership that might be of interest (e.g. university student newspapers, professional sub-communities).

Why don’t CEA or other EA orgs push back more publicly on misconceptions?

The advice we’ve gotten so far is to not repeat misconceptions. You’re unlikely to see an EA organization say “No, X isn’t true; actually Y is true.” Instead they’re more likely to talk about “Here’s why Y is important.”

Some criticisms will be unfair or uninformed. Typically we expect to respond by writing pieces explaining our own views rather than responding directly to critical pieces.

Should I reach out to celebrities, HNWIs, etc about getting involved with EA?

Very unlikely. There are existing projects doing this, and it’s better that outreach happen in a coordinated way than a bunch of people contacting them. Simran Dhaliwal of Longview Philanthropy writes: “Please do reach out if you have a connection to an UHNW individual/family; we'd be more than happy to invest the many hours it takes to build a relationship and discuss EA concepts in-depth / assist with coordination." s...

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 to apply for a PhD, published by Abby Hoskin on May 11, 2022 on The Effective Altruism Forum.

Many people in the EA community are intelligent truth seekers. Perhaps unsurprisingly, many of us are attracted to academia/graduate school. Now is actually a pretty good time to start prepping to apply for PhD programs, so here's the advice I give to all the undergrads who contact me. People have different follow up questions, so I'll monitor the comments section of this post if you have specific questions.

Credentials: I'm currently completing a PhD in Psychology and Neuroscience at Princeton University. I also got PhD offers from Columbia and Boston University. I was shortlisted at Harvard and Stanford, but I told them I was going to take my Princeton offer so the advisors I applied to didn't "waste" their admission slot on me.

How to get in

Get as much research experience as possible before you apply. Summer internships are the perfect time to work in a lab. If your university offers class credit for working in a lab during the semester, do that as well. You are applying to be a full time researcher, so research experience and publications are the most impressive things to put on your resume.

Being a research assistant/lab manager in a lab will look great on your resume; it will also be a VIP ticket to that lab when it comes to application season. It is very rare that RAs who want to stay to be graduate students are not able to stay in their labs. This is a thing you can do for a year or two after graduating university.

You can also apply for post-baccalaureate research fellowships for the year or two after you graduate university. The NIH runs a good post-bacc program you can look into.

Good grades are important, but research experience and publications can make professors look past mediocre grades and/or standardized test scores.

For STEM programs, try to learn as much stats and programming as possible, both before and during graduate school. For psychology and neuroscience, linear algebra, Bayesian statistics, and mixed effects regression models are all super helpful to learn. Python is the most useful and versatile language to learn. R is popular for statistics, but you can do most stats in Python too. Matlab is popular in academic environments but is declining in use. A lot of legacy code is written in Matlab, so chances are you will encounter it if you inherit legacy code from a senior student.

For your personal statement, avoid talking about your personal dreams like, "It has always been your dream to study the human condition at a beautiful Ivy League university". Professors don't care about this. Professors want to hire somebody who will immediately start producing publishable research. Instead, structure your personal statement to explain your research skills (e.g., "I developed my data analysis skills by analyzing economic trends in R for my senior thesis.") and your specific research interests, like: "I want to explore the psychological impact of growing up in poverty by studying the development of schizophrenia in different income brackets."

For your personal statement, identify at least one professor at the school you're applying to and explain what interests you about their work. For example, "I am excited to apply to Brown University because of the groundbreaking work Professor Shenhav has done on cognitive control. I would like to apply his expected value of control theory to better understand decision making in different clinical populations."

Understand the structure of the program you're applying to. Are you applying to work with a single professor, as you do in Princeton's Psychology program? Or are you applying to a committee and will you be spending your first year rotating through labs, as you do in Princeton's Neuroscience program? If the for...

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: [Intro to brain-like-AGI safety] 14. Controlled AGI, published by Steve Byrnes on May 11, 2022 on The AI Alignment Forum.

Part of the “Intro to brain-like-AGI safety” post series.

14.1 Post summary / Table of contents

Post #12 suggested two paths forward for solving “the alignment problem” for brain-like AGI, which I called “Social-instinct AGI” and “Controlled AGI”. Then Post #13 went into more detail about (one aspect of) “Social-instinct AGI”. And now, in this post, we’re switching over to “Controlled AGI”.

If you haven’t read Post #12, don’t worry, the “Controlled AGI” research path is nothing fancy—it’s merely the idea of solving the alignment problem in the most obvious way possible:

The “Controlled AGI” research path:

Step 1 (out-of-scope for this series): We decide what we want our AGI’s motivation to be. For example, that might be:

“Invent a better solar cell without causing catastrophe” (task-directed AGI),

“Be a helpful assistant to the human supervisor” (corrigible AGI assistants),

“Fulfill the human supervisor’s deepest life goals” (ambitious value learning),

“Maximize coherent extrapolated volition”,

or whatever else we choose.

Step 2 (subject of this post): We make an AGI with that motivation.

This post is about Step 2, whereas Step 1 is out-of-scope for this series. Honestly, I’d be ecstatic if we figured out how to reliably set the AGI’s motivation to any of those things I mentioned under Step 1.

Unfortunately, I don’t know any good plan for Step 2, and (I claim) nobody else does either. But I do have some vague thoughts and ideas, and I will share them here, in the spirit of brainstorming. This post is not meant to be a comprehensive overview of the whole problem, just what I see as the most urgent missing ingredients.

Out of all the posts in the series, this post is the hands-down winner for “most lightly-held opinions”. For almost anything I say in this post, I can easily imagine someone changing my mind within an hour of conversation. Let that ‘someone’ be you—the comment section is below!

Table of contents:

Section 14.2 discusses what we might use as “Thought Assessors” in an AGI. If you’re just tuning in, Thought Assessors were defined in Posts #5–#6 and have been discussed throughout the series. If you have a Reinforcement Learning background, think of Thought Assessors as the components of a multi-dimensional value function. If you have a “being a human” background, think of Thought Assessors as learned functions that trigger visceral reactions (aversion, cortisol-release, etc.) based on the thought that you’re consciously thinking right now.

In the case of brain-like AGIs, we get to pick whatever Thought Assessors we want, and I propose three categories for consideration: Thought Assessors oriented towards safety (e.g. “this thought / plan involves me being honest”), Thought Assessors oriented towards accomplishing a task (e.g. “this thought / plan will lead to better solar cell designs”), and Thought Assessors oriented purely towards interpretability (e.g. “this thought / plan has something to do with dogs”).

Section 14.3 discusses how we might generate supervisory signals to train those Thought Assessors. Part of this topic is what I call the “first-person problem”, namely the open question of whether it’s possible to take third-person labeled data (e.g. a YouTube video where Alice deceives Bob), and transmute it into a first-person preference (an AGI’s desire to not, itself, be deceptive).

Section 14.4 discusses the problem that the AGI will encounter “edge cases” in its preferences—plans or places where its preferences become ill-defined or self-contradictory. I’m cautiously optimistic that we can build a system that monitors the AGI’s thoughts and detects when it encounters an edge case. However, I don’t have any good idea about what to...

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: Have You Ever Doubted Whether You're Good Enough To Pursue Your Career?, published by lynettebye on May 11, 2022 on The Effective Altruism Forum.

This post is cross posted on my blog.

People often look to others that they deem particularly productive and successful and come up with (often fairly ungrounded) guesses for how these people accomplish so much. Instead of guessing, I want to give a peek behind the curtain.

I interviewed eleven people I thought were particularly successful, relatable, or productive. We discussed topics ranging from productivity to career exploration to self-care.

The Peak behind the Curtain interview series is meant to help dispel common myths and provide a variety of takes on success and productivity from real people. To that end, I’ve grouped responses on common themes to showcase a diversity of opinions on these topics.

This first post covers “Have you ever doubted whether you're good enough to pursue your career?” and other personal struggles.

My guests include:

Abigail Olvera was a U.S. diplomat last working at the China Desk. Abi was formerly stationed at the US Embassies in Egypt and Senegal and holds a Master's of Global Affairs from Yale University. Full interview.

Ajeya Cotra is a Senior Research Analyst at Open Philanthropy where she worked on a framework for estimating when transformative AI may be developed, as well as various cause prioritization and worldview diversification projects. Ajeya received a B.S. in Electrical Engineering and Computer Science from UC Berkeley. Full interview.

Ben Garfinkel was a research fellow at the Future of Humanity Institute at the time of the interview. He is now the Acting Director of the Centre for the Governance of AI. Ben earned a degree in Physics and in Mathematics and Philosophy from Yale University, before deciding to study for a DPhil in International Relations at the University of Oxford. Full interview.

Daniel Ziegler researched AI safety at OpenAI. He has since left to do AI safety research at Redwood Research. Full interview.

Eva Vivalt did an Economics Ph.D. and Mathematics M.A. at the University of California, Berkeley after a master’s in Development Studies at Oxford University. She then worked at the World Bank for two years and founded AidGrade before finding her way back to academia. Full interview.

Gregory Lewis is a DPhil Scholar at the Future of Humanity Institute, where he investigates long-run impacts and potential catastrophic risk from advancing biotechnology. Previously, he was an academic clinical fellow in public health medicine and before that a junior doctor. He holds a master’s in public health and a medical degree, both from Cambridge University. Full interview.

Helen Toner is Director of Strategy at Georgetown’s Center for Security and Emerging Technology (CSET). She previously worked as a Senior Research Analyst at Open Philanthropy. She is a member of the board of directors for OpenAI. Helen holds an MA in Security Studies from Georgetown. Full interview not available.

Jade Leung is Governance Lead at OpenAI. She was the inaugural Head of Research & Partnerships with the Centre for the Governance of Artificial Intelligence (GovAI), housed at Oxford’s Future of Humanity Institute. She completed her DPhil in AI Governance at the University of Oxford and is a Rhodes scholar. Full interview.

Julia Wise serves as a contact person for the effective altruism community and helps local and online groups support their members. She serves on the board of GiveWell and writes about effective altruism at Giving Gladly. She was president of Giving What We Can from 2017-2020. Before joining CEA, Julia was a social worker, and studied sociology at Bryn Mawr College. Full interview.

Michelle Hutchinson holds a PhD in Philosophy from the University of Oxford, where her thesis ...

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 I torched my biggest career opportunity so far, published by Bluefalcon on May 11, 2022 on The Effective Altruism Forum.

I probably could have worked for a US Cabinet-level appointee, with an annual budget in the $10 billion order of magnitude, in a capacity where I would have had routine daily access to this person and influence on this person's decisions, if I had made a better decision at a key time several years prior.

There's been some discussion of what one's chances are if one takes more career risks to maximize expected value, using Sam Bankman-Fried as an example. So I want to tackle this from a different angle, looking at my own failure to make good on the most impactful career opportunity I've had thus far, in politics rather than earning to give. I also want to discuss other ways I could easily have screwed it up, that would cause me to be unaware of how close I came, because I think if you've never had a highly visible near miss like mine, you may not even realize what kinds of opportunities would be available to you if you tried.

I don't want to dox myself so I have to keep some details vague.

The time from when I started going all out maximizing my efforts, to when the highest-impact opportunity would have occurred, is about 13 years. I did not come from any kind of money, political connections, or other elite background prior to going to college. Where and when I went to college played a role here, but mostly by putting me in the right place at the right time.

I had the opportunity to work for someone who later went on to serve in a Cabinet-level position, where I would have had significant daily one on one time with this person, and I turned it down. This wasn't a totally crazy decision in light of the information known to me at the time I made it, but I think a person with better judgment would have seen the job offer as having higher expected value than I recognized at the time.

Why did I turn it down? Because it was an entry level job, and I had enough political experience to find something higher up elsewhere. And ultimately, the job I got a few months later was better on many metrics. It was a more senior role, providing better preparation for my goal at the time to hop on a gubernatorial or senate campaign in an even more senior role in the following election cycle, with the goal of then going to work on that person's staff if they were elected. I succeeded in every step of this plan, except that my candidate the following cycle lost, so the final step was unachievable.

What factors of the situation could I have given more weight to, that might have changed my decision?

1) The job I was offered would have given me a lot of time with the candidate, which lends itself well to upward mobility within the campaign and also provides opportunities to influence the candidate's views on policy informally.

2) The candidate sought me out, rather than me applying for the job, suggesting they thought I had a lot of potential, either for promotion within the campaign or as someone to bring into government with them if elected. Additionally, I had extremely strong references this candidate was close to, who were not consulted before the decision to offer me the job, so I perhaps could have leveraged them in seeking a promotion later.

3) I ultimately went to work for an incumbent's campaign, while the job I turned down was on an open seat race. Incumbents have fewer jobs in government to give out if they win, because they've already got their team in place. Fine if you are dead set on jumping on another campaign, but obviously something to consider since at some point you have to go into government to make this all worthwhile.

4) The candidate I turned down was one of my own elected officials, meaning I could potentially have run directly for their seat wh...

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: ProjectLawful.com: Eliezer's latest story, past 1M words, published by Eliezer Yudkowsky on May 11, 2022 on LessWrong.

So if you read Harry Potter and the Methods of Rationality, and thought...

"You know, HPMOR is pretty good so far as it goes; but Harry is much too cautious and doesn't have nearly enough manic momentum, his rationality lectures aren't long enough, and all of his personal relationships are way way way too healthy."

...then have I got the story for you! Planecrash aka Project Lawful aka Mad Investor Chaos and the Woman of Asmodeus, is a story in roleplay-format that I as "Iarwain" am cowriting with Lintamande, now past 1,000,000 words.

It's the story of Keltham, from the world of dath ilan; a place of high scientific achievement but rather innocent in some ways. For mysterious reasons they've screened off their own past, and very few now know what their prescientific history was like.

Keltham dies in a plane crash and ends up in the country of Cheliax, whose god is "Asmodeus", whose alignment is "Lawful Evil" and whose people usually go to the afterlife of "Hell".

And so, like most dath ilani would, in that position, Keltham sets out to bring the industrial and scientific revolutions to his new planet! Starting with Cheliax!

(Keltham's new friends may not have been entirely frank with him about exactly what Asmodeus wants, what Evil really is, or what sort of place Hell is.)

This is not a story for kids, even less so than HPMOR. There is romance, there is sex, there are deliberately bad kink practices whose explicit purpose is to get people to actually hurt somebody else so that they'll end up damned to Hell, and also there's math.

The starting point is Book 1, Mad Investor Chaos and the Woman of Asmodeus. I suggest logging into ProjectLawful.com with Google, or creating an email login, in order to track where you are inside the story.

Please avoid story spoilers in the comments, especially ones without spoiler protection; this is not meant as an "ask Eliezer things about MICWOA" thread.

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: Crises Don't Need Your Software, published by GabrielExists on May 10, 2022 on LessWrong.

About a month ago, I was invited to contribute to a group looking to help Ukrainian refugees in Poland. The group consisted of volunteers, including some people from the rationalist community, who knew they wanted to help, and were searching for high impact ways to do so, likely through software. I had been intending to find a way to help Ukraine, so when this opportunity to use my programming background to help arose, I felt elated to finally be able to do something.' I took two weeks off work the next morning (thank you, my employer, for letting me do that) and optimized my life to spend as many hours as I could on the project without harming myself.

The group consisted of some really amazing people. There were two students from the US, Soren and Luke, who had more or less dropped everything, sold their cars, and just moved to Poland to help. They were our eyes on the ground and the ones starting the project. My close friend, Leonardo, was the one who invited me, and he was the one who inspired me to take time off work by doing it himself first. And a lot of other brilliant people with very valuable contributions, who I won't list to keep the cast small for the reader. Just know I'm very thankful for everyone who helped out.

We knew we wanted to help somehow, but we didn't know how. We focused on looking for problems we could solve with software since the group had a high density of software people. Over the course of the first few days we went through a concept of tracking the movements of people through the country (dismissed because it would be a GDPR nightmare and also very abusable), a concept of assigning refugees entering the country to different destinations so not everyone ended up in Krakow (depended on a model of reality that was much too simplified, and also had problems with how to gather data to base algorithmic decisions on) and we finally arrived at a concept around a certain kind of buses.

Somehow I ended up in a leadership position along the way, a de facto product manager so to say, and I created the bus concept from something a volunteer hosting refugees in his office space had said. They had been in contact with a municipality from Norway who had space for refugees and was sending buses to directly transport refugees. This seemed to be a regular occurrence, where individual people, small aid organisations or municipalities were sending buses and cars to Poland. Then they would show up and nobody would want to get on, because no one knew they'd be there, what it's like where they would be going or if they could be trusted.

So the concept was to create a website where the bus drivers could see a heatmap of where refugees willing to go to their country were and plan their trips after where there were too few buses compared to refugees. And the refugees (or rather, volunteers helping refugees) could sign up to add to the data set the heat map comes from, and then see which planned trips would be nearby, and then "follow" them for updates and mark interest for which ones they'd want to go with. Having access to contact information in advance would allow them to get an understanding of where the trip would be going, have time to look up the country, maybe see a "day in the life" video to understand what the country feels like and to some degree vet the bus driver in advance.

By not addressing the things that worked, such as the trains, flights and regularly scheduled buses, and focusing on connecting small scale actors with other small scale actors that need each other, we wanted to make a difference at scale by enabling many small interactions.

After we'd worked on refining the concept from Wednesday to Saturday, as well as starting development and looking for ...

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: EA Tours of Service, published by Ben West on May 10, 2022 on The Effective Altruism Forum.

Summary: jobs in EA organizations are often unusual in ways which makes candidates uncertain if they want to have that job indefinitely (e.g. people may have a less prestigious title because EA's tend to be “overqualified”). Hiring people through a "Tour of Service" which is time-bound and has specific outcomes can address these concerns. An example summary of a Tour is: join CEA for two years with the goal of doubling the percentage of respondents to the EA survey who say they made an important connection on the EA Forum.

Background

About a decade ago, Tours of Service were introduced as a replacement for the traditional "at will" employment structure. The authors summarized:

If you think all your people will give you lifetime loyalty, think again: Sooner or later, most employees will pivot into a new opportunity. Recognizing this fact, companies can strike incremental alliances. When Reid founded LinkedIn, he set the initial employee compact as a four-year tour of duty, with a discussion at two years. If an employee moved the needle on the business during the four years, the company would help advance his career. Ideally this would entail another tour of duty at the company, but it could also mean a position elsewhere.

The central aspects of the Tour of Service model are, to my mind:

Time limited (usually 2-4 years).

Target concrete outcomes in which the employee’s work will benefit the employer.

Target concrete outcomes in which the employee will be benefited by working with the employer.

Importantly, the final point may be oriented around ways in which the employee would be more employable at other organizations.

Note that the Tour of Service (both in the original version and CEA’s) is an informal and non-legally binding agreement. The legal structure of employment is unchanged (which, in the US, usually means "at will" employment).

Examples

Probably the easiest way to understand Tours of Service is to look at examples. These examples are all from actual Tours of Service of CEA staff.

Service Objectives (benefits to employer)

We the company expect this current Tour of Service to encompass the time it takes for you to execute the following objectives:

100% increase in the percentage of respondents to the EA survey who say they made an important connection on the EA Forum.

>5,000 “hot leads”/year are handed off to group leaders or other CEA teams through products you create.

The EA Forum improves event discoverability, such that 30 event attendees per month are attributable to the Forum.

Service Results (benefits to employee)

Here is what a successful tour of duty look like for you (knowledge, skills, accomplishments, recognition, etc.):

Recognition from CEA leadership and group leaders as someone who can successfully complete large, full-stack, technical projects with diverse stakeholders.

The impact of having built a pipeline that >20% of new, highly-engaged EAs will have gone through at least part of.

A portfolio of open source contributions to successful projects which can be shown to prospective employers.

Full Example

Here is a full Tour of Service that I wrote for myself when I first joined CEA in 2018:

We the company expect this current tour of duty to encompass the time it takes for you to execute the following mission objectives:

Each team managed by you (events, community health, tech, grants) has a clear strategy with success metrics

Top performers have high morale and are retained; substandard performance is addressed.

New CEO can be brought up to speed easily.

I expect this tour of duty will last approximately the following amount of time:

Until a new CEO starts (approximately end of 2019).

Here is what the results of a successful tour of duty look like for the...

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: Forecasting Newsletter: April 2022, published by NunoSempere on May 10, 2022 on The Effective Altruism Forum.

Highlights

The US CDC creates a pandemic forecasting center

Tetlock's forecasting team is recruiting for a new tournament on existential risk

Manifold Markets makes their webpage open-source

Index

Notable news

Platform by platform

Correction on Polymarket incentivizing wash-trading.

Shortform research

Longform research

You can sign up for this newsletter on substack, or browse past newsletters here.

The state of forecasting: Where next?

At the high level, various new startups have been exploring different parts of the forecasting and prediction market space. Most recently, Manifold Markets allows anyone to easily instantly create a market, which comes at the cost of these markets having to use play money. And Polymarket or Insight Prediction have set up real-money prediction markets on some topics of real interest.

Our hope is that with time, a new project might stumble upon a winning combination, one that is able to produce large amounts of value by allowing people to pay for more accurate probabilities, even when these are unflattering.

Putting on my Scott Alexander hat, prediction markets or forecasting systems have the potential to be a powerful asymmetric weapon (a): a sword of good that systematically favours the side which seeks the truth. A channel which is able to transmit accurate information even when the evil forces of Moloch (a) would seek to prevent this.

In Linux circles, there is the meme of "the year of the Linux desktop": a magical year in which people realize that Windows adds little value, and that Linux is based. A year in which people would together switch to Linux (a) and sing John Lennon. Instead, Linux, much like prediction markets, remains a powerful yet obscure tool which only a few cognoscenti can wield. So although I may wax poetically about the potential of prediction markets, "the year of the prediction market" has yet to come, much like "the year of the Linux desktop".

To get there, there needs to be investment in accessibility and popularization. And so I am happy to see things like the Forecasting Wiki (a), simpler apps (a), popular writers introducing prediction markets to their audiences, or platforms experimenting with simplifying some of the complexity of prediction markets, or with combining the ideas behind prediction markets in new ways and exploring the space of prediction market-like things. It might also be wise for prediction markets to take a more confrontational stance (a), by challenging powerful people who spout bullshit.

Notable news

The CDC has a new (a), $200M center (a) for pandemic forecasting.

When Walensky tapped outside experts to head the new outfit, the move was widely viewed as an acknowledgment of long-standing and systemic failures regarding surveillance, data collection and preparedness that were put into high relief by the pandemic.

Scientists will also look at who is infecting whom, how well vaccines protect against infection and severe illness, and how that depends on the vaccine, variants and the time since vaccination, said Marc Lipsitch, an epidemiologist and the center’s science director.

The center will be based in D.C. and will eventually have about 100 staff members, including some at CDC’s Atlanta headquarters. It will report to Walensky.

This is, broadly speaking, good. But it also seems like too little, too late. It seems suboptimal to have this center report to the CDC director, given that the CDC's leadership wasn't particularly shining during the pandemic. And the center is playing defence against the last black swan, whereas I would prefer to see measures which could defend against unknown unknowns, such as this one (a).

I recently stumbled upon a few prediction markets previously...

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: Potatoes: A Critical Review, published by Pablo Villalobos on May 10, 2022 on The Effective Altruism Forum.

Executive summary

Nunn and Qian study the effect of the introduction of potatoes in the Old World on population growth between 1700 and 1900.

We think the paper credibly establishes that between one-sixth and one-quarter of the growth is a consequence of the introduction of potatoes.

The main reason for doubt is the possibility of spurious correlation due to spatiotemporal autocorrelation and the fact that potatoes were mainly grown in Europe, which at the time was experiencing growth due to unrelated factors.

After performing several tests to account for these concerns, we conclude they are not strong enough to reject the conclusion of the paper.

Introduction

During the 18th and 19th centuries we saw rapid urbanisation and a drastic increase of population throughout the world. Some attribute this to industrialization; Nunn and Qian argue that we owe that to potatoes.

They are not coy about it either. “According to our most conservative estimates, the introduction of the potato accounts for approximately one-quarter of the growth in Old World population and urbanisation between 1700 and 1900”.

They back their hypothesis with data from several natural experiments, studying variability of city populations and adult heights between and within countries, and relate it to the timing when potatoes were introduced in different places.

In this article we will summarise, replicate and critically review Nunn and Qian’s paper. We follow the methods established in (Sevilla, 2021) to study whether their findings are robust and whether they successfully establish a causal relation.

Paper Summary

The study purports to evaluate the causal impact of the introduction of potatoes in the Old World in population growth and urbanisation rates.

The theory is that potatoes are more nutritious and provide 3x more calories per acre, so when they started being cultivated in the Old World they produced a positive shock in agricultural productivity, enabling higher populations and wealth per capita.

A previous study by (Mokyr, 1981) estimated the causal effect of potatoes on population growth in Irish counties in 1845, finding an effect size of 0.7. This implies counties with high potato cultivation grew an extra 0.15% that year, compared to counties with low cultivation. In comparison, Nunn and Qian’s study uses data of the whole Old World, from 1300 to 1900, instead of just Ireland in a single time slice, and looks at urbanisation rates in addition to population.

To estimate the causal impact of potatoes, the authors exploit two sources of variation on each country’s ability to grow potatoes: the time of introduction and the suitability of the land for potato cultivation. In their baseline analysis they find that around a quarter of the growth in population and urbanisation rates in that time period can be attributed to potatoes.

The main pitfall of this strategy is that it relies on there being no other shocks during that time which are correlated with suitability for cultivation. Unfortunately, Europe is much more suitable for potato cultivation than other Old World regions, and the potato was introduced at the same time when Europe was diverging from other countries due to several unrelated factors.

To check that the effect is indeed causal, the authors use several strategies:

Adding several controls for alternative drivers of population and economic growth.

Comparing only countries within the same continent

Comparing cities within the same country

Comparing the heights of soldiers in France, using their town of birth to estimate the importance of potatoes in their childhood diet.

In all of those cases, they find a significant effect of potatoes on the outcome. For the within-continen...