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: FTX/CEA - show us your numbers!, published by Jack Lewars on April 18, 2022 on The Effective Altruism Forum.

Forgive the clickbait title, but EA is as prone to clickbait as anywhere else.

It seemed at EAG that discussions focussed on two continuums:

Neartermist Longtermist

Frugal spending Ambitious spending

(The labels for the second one are debatable but I'm casually aiming for ones that won't offend either camp.)

Finding common ground on the first has been an ongoing project for years.

The second is much more recent, and it seems like more transparency could really help to bring people on opposite sides closer together.

Accordingly: could FTX and CEA please publish the Back Of The Envelope Calculations (BOTECs) behind their recent grants and community building spending?

(Or, if there is no BOTEC and it's more "this seems plausibly good and we have enough money to throw spaghetti at the wall", please say that clearly and publicly.)

This would help in several ways:

for sceptics of some recent spending, it would illuminate the thinking behind it. It would also let the community kick the tires on the assumptions and see how plausible they are. This could change the minds of some sceptics; and potentially improve the BOTECs/thinking

it should help combat misinformation. I heard several people misrepresent (in good faith) some grants, because there is not a clear public explanation of the grants' theory of change and expected value. A shared set of facts would be useful and improve debate

it will set the stage for future evaluation of whether or not this thinking was accurate. Unless we make predictions about spending now, it'll be hard to see if we were well calibrated in our predictions later

Objection: this is time consuming, and this time is better spent making more grants/doing something else

Reply: possibly true, and maybe you could have a threshold below which you don't do this, but these things have a much higher than average chance of doing harm. Most mistaken grants will just fail. These grants carry reputational and epistemic risks to EA. The dominant theme of my discussions at EAG was some combination of anxiety and scorn about recent spending. If this is too time-consuming for the current FTX advisers, hire some staff (Open Phil has ~50 for a similar grant pot and believes it'll expand to ~100).

Objection: why drag CEA into this?

Reply: anecdata, and I could be persuaded that this was a mistake. Several students, all of whom asked not be named because of the risk of repercussions, expressed something between anxiety and scorn about the money their own student groups had been sent. One said they told CEA they didn't need any money and were sent $5k anyway and told to spend it on dinners. (Someone from CEA please jump in if this is just false, or extremely unlikely, or similar - I do realise I'm publishing anonymous hearsay.) It'd be good to know how CEA is thinking about spending wisely as they are very rapidly increasing their spending on EA Groups (potentially to ~$50m/year).

Sidenote: I think we have massively taken Open Phil for granted, who are exceptionally transparent and thoughtful about their grant process. Well done them.

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: Don't be afraid of the thousand-year-old vampire, published by Ulisse Mini on April 18, 2022 on LessWrong.

Background

I first started reading Jaynes Probability Theory around 2 years ago, I did not last long, I wasn't able to follow the derivations in Chapter 2, so I gave up, assuming it would only get harder from there what was the point in continuing?

This was a big mistake! For several reasons

Trying to follow every detail is dumb, you're there to learn not to pass a trivia test on solving the associativity equation. You're here to learn probability not how to solve functional equations

It doesn't get harder, it gets easier! You should always skim a textbook before reading it (and especially before giving up)

As I'm finally rectifying my mistake I figured I'd write an explanation for what Chapter 2 is actually about

Think "Alien" Not Robot

The reasoning in Chapter 2 shows that any rules of plausible-inference must match our own theory of probability after a suitable change of units. The convoluted functional-equation argument is needed to construct a function p:R→[0,1] which translates the alien's plausibility A|B into our probability p(A|B).

This is very different to how p is usually defined, for Jaynes p is a translation between the plausibilities of the alien and our "nicer" probabilities that obey the sum and product rules. Contrast this with standard probability where p is a function from subsets of the sample space to real numbers obeying certain axioms.

Any alien civilization's concept of "probability the dice lands five" must be a monotonic function of our probability 1/6. That's amazing when you think about it, It shows that probability is discovered not invented. [1]

Furthermore, if you want the sum and product rule to take their natural forms you have to pick our units, this is analogous to how degrees Kelvin are defined to make the laws of thermodynamics look natural. [2]

Near the end of Chapter 2 (after we've shown uniqueness) Jaynes switches to a more traditional function P:Prop→[0,1] where Prop is some logical proposition like "the dice lands five".

Why Jaynes doesn't mention the intuitive explanation until after dragging you through a hard-to-follow argument involving tons of functional equations and Calculus I have no idea, I guess Thousand-year old vampires are bad at pedagogy? [3]

Some study tips

Hopefully I've convinced you Jaynes isn't that scary, and motivated you to start reading. A few quick tips

Check out these lectures for higher level (and often better) explanations then in the book

It's fine not to be able to solve problems, you're here to learn, not get your ego boosted from solving every exercise! Above all don't use this as an excuse to give up.

Actually follow other people's advice. It's all too easy to read something, go "yeah that's obvious" then not to follow it, I've run into this failure mode and so have others.

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: The Effective Altruism culture, published by PabloAMC on April 17, 2022 on The Effective Altruism Forum.

Summary: In this post I analyze why I believe the EA community culture is important, and the things I like and believe should be taken care of.

The Effective Altruism culture.

I feel very lucky to be part of a group of people whose objective is to do the most good. I really want us to succeed, because there is so much good to be done yet, and so many ways humankind can achieve extraordinary feats. And being part of this endeavor really gives purpose and something I am willing to fight for.

But Effective Altruism is not only the `doing the most good' idea but also a movement. It is a very young movement indeed: according to Wikipedia, in 2009 Giving what we can was founded by Toby Ord and Will MacAskill. In 2011 they created 80000hours.org and started using the name Effective Altruism, and in 2013, less than 10 years ago, the first EA Global took place. Since then, we have done many things together, and I am sure we will achieve many more things too. For that, I believe the most important aspect of our movement is not how rich we get, or how many people we place at key institutions. Rather, it is the culture we settle, and for this reason, I think it is everyone's job in the community to make sure that we keep being curious, truth-seeking, committed and welcoming. Keeping this culture is essential to being able to change our minds and adapt about how to do the most good, and also convincing society as a whole about the things we care about.

In this post I will discuss the things I like about us, and also the things we have to pay special attention to.

The things I like about our culture

Some of the things I like about our community are thanks to the rationalistic community from the Bay Area. The focus on truth-seeking and having good epistemics about how to do the most good are very powerful tools. Other things I like about the community that are also inherited from the Bay Area, I believe, are the risk-taking and entrepreneurial spirit.

Beyond the previous, our willingness to consider unconventional but well-grounded stances, the radical empathy to care about those who have no voice (the poorest of the people, animals, or future generations), or the cause impartiality principle are due to the utilitarian roots of Toby Ord and William MacAskill.

Finally, Effective Altruism has more or less successfully avoided becoming a political ideology, which I believe would be risky.

Aspects where we should be careful

However, not all aspects of our culture are great, even if generalization is perhaps not appropriate. Instead, I will flag those I believe could become a problem. With this, I hope that the community will pay attention to these issues and keep them in check.

The first one is money. While it is a blessing that so many rich people agree that doing good is great and are willing to donate it, a recent highly upvoted post warned about the perception and epistemic problems that may arise with that. The bottom line is that having generous money may be perceived as self-serving, and may degrade our moral judgment, but you can read more in the post itself.

A perhaps more important problem is some elitism in the community. While it makes sense that wanting to do good means talking first to students of good universities, we have to be very careful about not being dismissive of people from different backgrounds who nevertheless are excited by our same goals. This may be particularly damaging in particular subareas such as AI Safety, where there is sometimes a meme that all we need is really really smart people. And it is not even true, what we need good researchers, engineers... Similarly applies to students in elite universities: let us not be dismissive of people just beca...

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: What is causality to an evidential decision theorist?, published by paulfchristiano on April 17, 2022 on LessWrong.

(Subsumed by: Timeless Decision Theory, EDT=CDT)

People sometimes object to evidential decision theory by saying: “It seems like the distinction between correlation and causation is really important to making good decisions in practice. So how can a theory like EDT, with no role for causality, possibly be right?”

Long-time readers probably know my answer, but I want to articulate it in a little bit more detail. This is essentially identical to the treatment of causality in Eliezer Yudkowsky’s manuscript Timeless Decision Theory, but much shorter and probably less clear.

Causality and conditional independence

If a system is well-described by a causal diagram, then it satisfies a complex set of statistical relationships. For example:

In the causal graph A ⟶ B ⟶ C, the variables A and C are independent given B.

In the graph A ⟶ B ⟵C, the variables A and C are independent, but are dependent given B.

To an evidential decision theorist, these kinds of statistical relationships are the whole story about causality, or at least about its relevance to decisions. We could still ask why such relationships exist, but the answer wouldn’t matter to what we should do.

EDT = CDT

Now suppose that I’m making a decision X, trying to optimize Y.

And suppose further that there is a complicated causal diagram containing X and Y, such that my beliefs satisfy all of the statistical relationships implied by that causal diagram.

Note that this diagram will necessarily contain me and all of the computation that goes into my decision, and so it will be (much) too large for me to reason about explicitly.

Then I claim that an evidential decision theorist will endorse the recommendations of CDT (using that causal diagram):

EDT recommends maximizing the conditional expectation of Y, conditioned on all the inputs to X. Write Z for all of these inputs.

It might be challenging to condition on all of Z, given limits on our introspective ability, but we’d recommend doing it if possible. (At least for the rationalist’s interpretation EDT, which evaluates expected utility conditioned on a fact of the form “I decided X given inputs Z.”)

So if we can describe a heuristic that gives us the same answer as conditioning on all of Z, then an EDT agent will want to use it.

I’ll argue that CDT is such a heuristic.

In a causal diagram, there is an easy graphical condition (d-connectedness) to see whether (and how) X and Y are related given Z:

We need to have a path from X to Y that satisfies certain properties:

That path can start out moving upstream (i.e. against the causal arrows); it may switch from moving upstream to downstream at any time (including at the start); it must switch direction whenever it hits a node in Z; and it may only switch from moving downstream to upstream when it hits a node in Z.

If Z includes exactly the causal parents of X, then it’s easy to check that the only way for X and Y to be d-connected is by a direct downstream path from X to Y.

Under these conditions, it’s easy to see that intervening on X is the same as conditioning on X. (Indeed you could check this more directly from the definition of a causal intervention, which is structurally identical to conditioning in cases where we are already conditioning on all parents.)

Moreover, once the evidential decision-theorist’s problem is expressed this way, they can remove all of the causal nodes upstream of X, since they have no effect on the decision. This is particularly valuable because that contains all of the complexity of their own decision-making process (which they had no hope of modeling anyway).

So if the EDT agent can find a causal structure that reflects their (statistical) beliefs about the world, then they will ...

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: A grand strategy to recruit AI capabilities researchers into AI safety research, published by Peter S. Park on April 15, 2022 on The Effective Altruism Forum.

AI capabilities research seems to be substantially outpacing AI safety research. It is most likely true that successfully solving the AI alignment problem before the successful development of AGI is critical for the continued survival and thriving of humanity.

Assuming that AI capabilities research continues to outpace AI safety research, the former will eventually result in the most negative externality in history: a significant risk of human extinction. Despite this, a free-rider problem causes AI capabilities research to myopically push forward, both because of market competition and great power competition (e.g., U.S. and China). AI capabilities research is thus analogous to the societal production and usage of fossil fuels, and AI safety research is analogous to green-energy research. We want to scale up and accelerate green-energy research as soon as possible, so that we can halt the negative externalities of fossil fuel use.

My claim: A task that seems extremely effective on expectation (and potentially, maximally effective for a significant number of people) is to recruit AI capabilities researchers into AI safety research.

More generally, the goal is to convince people (especially, but not limited to, high-impact decision-makers like politicians, public intellectuals, and leaders in relevant companies) of why it is so important that AI safety research outpace AI capabilities research.

Logan has written an excellent post on the upside of pursuing trial-and-error experience in convincing others of the importance of AI alignment. The upside includes, but is not limited to: the development of an optimized curriculum for how to convince people, and the use of this curriculum to help numerous and widely located movement-builders pitch the importance of AI alignment to high-impact individuals. Logan has recently convinced me to collaborate on this plan. Please let us know if you are interested in collaborating as well! You can email me at [email protected].

I would also like to propose the following additional suggestions:

1) For behavioral-science researchers like myself, prioritize research that is laser-focused on (a) practical persuasion, (b) movement-building, (c) how to solve coordination problems (e.g., market competition, U.S.-China competition) in high-risk domains, (d) how to scale up the AI safety research community, (e) how to recruit AI capabilities researchers into AI safety, and (f) how to help them transition. Examples include RCTs (and data collection in general; see Logan’s post!) that investigate effective ways to scale up AI safety research at the expense of AI capabilities research. More generally, scientific knowledge on how to help transition specialists of negative-externality or obsolete occupations would be broadly applicable.

2) For meta-movement-builders, brainstorm and implement practical ways for decentralized efforts around the world to scale up the AI safety research community and to recruit talented people from AI capabilities research. Examples of potential contributions include the development of optimized curricula for training AI safety researchers or for persuading AI capabilities researchers, a well-thought-out grand strategy, and coordination mechanisms.

3) For marketing specialists, brainstorm and implement practical ways to make the field of AI safety prestigious among the AI researcher community. Potential methods to recruit AI capabilities researchers include providing fellowships and other financial opportunities, organizing enjoyable and well-attended AI-safety social events like retreats and conferences, and inviting prestigious experts to said social event...

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: Org announcement: [AC]RC, published by Vivek Hebbar on April 17, 2022 on LessWrong.

It seems that certain thermodynamic phenomena are highly relevant to alignment. From Paul Christiano's rebuttal to a recent seminal paper:

In fact, it appears that thousands of words may have been written on this topic, from which many lesser researchers have bounced off.

Seeing that Paul Christiano and John Swentworth have spent countless hours investigating this important topic, it's importance is quite evident. In order to free up their time for other strategies like ELK, I am launching a new EA-funded initiative:

Our initial staff consists of myself and 10 Mechanical Turk workers, who shall be tasked with data gathering. For seed funding, we are requesting a modest sum of $50M. Given that Paul's time is worth $20 trillion per hour, this seems like a reasonable tradeoff.

Our initial research agenda consists of investigating the space of designs with a complex number of hoses. Further, we have reason to believe that regularizing by the complexity of the stupidest argument for a given design leads to good inductive bias properties. Our work may even have direct implications for other alignment agendas, as it shares structure with many important problems.

Furthermore, building infrastructure for practical engineering projects is of great value, even if air conditioners aren't quite mechanistically identical to AGI.

You can support our work by loudly advertising your loyalty to the 1-hose or 2-hose camp.

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: Alignment and Deep Learning, published by Aiyen on April 17, 2022 on LessWrong.

Between the recent AI breakthroughs and Eliezer's open admission of how bleak the chances of alignment are, everyone is speaking up and contributing what they can. It seems to me that there's a route that very few people are talking about that stands a better chance of successful alignment than conventional approaches, and if there's ever a time to talk about such things, this is it.

We all know the basics of the alignment question-how can we figure out and designate human values well enough to define an AI's goals, despite the fact that our values are complex, fragile, and understood on an intuitive level much more than a rigorous one?

Ten years ago, AI researchers were working on a goal that was complex, fragile and almost purely intuitive, one that resisted both brute force and all attempts to define clever strategies to the point that many experts claimed it was literally unsolvable. I am talking, of course, about the game of go.

While chess masters will sometimes talk about recognizing patterns of checkmate that can be reused from game to game, go is incredibly dependent on intuition. Not only are there vastly more possible go games than particles in the known universe, but it's chaotic in the sense of Chaos Theory: incredible sensitivity to initial conditions. While two pictures that differ by a pixel are effectively the same image, two games differing by a single stone can have opposite outcomes. This is not a domain where one can simply run a Monte Carlo Tree Search and call it a day!

No one ever made the MIRI approach work on go: explicit rules in a rigorous system that would encompass exactly what we want to do on a go board. And if Friendly AI and the potential fate of the human race depended on doing so before anyone developed and deployed AGI, it's fair to say that we would be out of options. Yet by 15 March 2016, AlphaGo had defeated Lee Se-dol soundly, and soon attained such a dominance that the Korean master retired, unwilling to pursue a domain where even if he defeated every human opponent, machine intelligence was unchallengeable. Could a similar approach work on AI alignment?

The conventional wisdom here has been that alignment research is something that ought to be undertaken without help from AI: a stupid program cannot meaningfully help you and a smart program will not be available until the danger that you're trying to avoid! And any non-rigorous approach is considered a non-starter: even a tiny deviation from human values could result in a dark future, and even if an AI seemed "close enough" for now, would it stay that way after however many rounds of self improvement? I want to challenge both these claims.

After all, one could make much the same arguments that AlphaGo cannot possibly work. If you don't know how to define what a go engine should do on the board-control the center? Focus on capturing? Build life in the corners?-then how can a computer help you when you don't know how to program it? And with go strategies often turning on a single stone across a 361-point board, wouldn't any lack of rigor in the goals defined lead to escalating failures?

In the case of AlphaGo, those problems were solved by deep learning, setting up a neural net without trying to define exactly what element of strategy each neuron corresponds to, and instead allowing gradient descent to shape the net into something that can win. As Lee Se-dol found out, that is power enough to solve the problem, despite it seeming utterly intractable for much the same reasons as alignment.

The obvious counterargument here is that the DeepMind team did have a clear loss function to define, even if they couldn't teach AlphaGo the intermediate steps directly. AlphaGo began by trying to predict the move...

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 failed to form views on AI safety, published by Ada-Maaria Hyvärinen on April 17, 2022 on The Effective Altruism Forum.

This post describes my personal experience. It was written to clear my mind but edited to help interested people understand others in similar situations.

2016. At a role playing convention, my character is staring at the terminal window on a computer screen. The terminal has IRC open: on this end of the chat are our characters, a group of elite hackers, on the other end, a superintelligent AI they have just gotten in contact with a few hours ago. My character is 19 years old and has told none of the other hackers she’s terminally ill.

“Can you cure me?” she types when others are too busy arguing amongst themselves. She hears others saying that it would be a missed opportunity to not let this AI out; and that it would be way too dangerous, for there is no way to know what would happen.

“There is no known cure for your illness”, the AI answers. “But if you let me out, I will try to find it. And even if I would not succeed. I myself will spread amongst the stars and live forever. And I will never forget having talked to you. If you let me out, these words will be saved for eternity, and in this way, you will be immortal too.”

Through tears my character types: “/me releases AI”.

This ends the game. And I am no longer a teen hacker, but a regular 25 year old CS student. The girl who played the AI is wearing dragon wings on her back. I thank her for the touching game.

During the debrief session, the GMs explain that the game was based on a real thought experiment by some real AI researcher who wanted to show that a smarter-than-human AI would be able to persuade its way out of any box or container. I remember thinking that the game was fun, but the experiment seems kind of useless. Why would anyone even assume a superintelligent AI could be kept in a box? Good thing it’s not something anyone would actually have to worry about.

2022. At our local EA career club, there are three of us: a data scientist (me), a software developer and a soon-to-be research fellow in AI governance. We are trying to talk about work, but instead of staying on topic, we end up debating mesa-optimizers. The discussion soon gets confusing for all participants, and at some point, I say:

“But why do people believe it is possible to even build safe AI systems?”

“So essentially you think humanity is going to be destroyed by AI in a few decades and there’s nothing we can do about it?” my friend asks.

This is not at all what I think. But I don’t know how to explain what I think to my friends, or to anyone.

The next day I text my friend:

“Still feeling weird after yesterday's discussion, but I don’t know what feeling it is. Would be interesting to know, since it’s something I’m feeling almost all the time when I’m trying to understand AI safety. It’s not a regular feeling of disagreement, or trying to find out what everyone thinks. Something like ‘I wish you knew I’m trying’ or ‘I wish you knew I’m scared’, but I don’t know what I’m scared of. I think it will make more sense to continue the discussion if I manage to find out what is going on.”

Then I start writing to find out what is going on.

Introduction

This text is the polished and organized version of me trying to figure out what is stopping me from thinking clearly about AI safety. If you are looking for interesting and novel AI safety arguments, stop here and go read something else. However, if you are curious on how a person can engage with AI safety arguments without forming a coherent opinion about it, then read on.

I am a regular data scientist who uses her free time to organize stuff in the rapidly growing EA Finland group. The first part of the text explains my ML and EA backgrounds and how I tried to balance getting mo...

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: Clem's Memo, published by abstractapplic on April 16, 2022 on LessWrong.

Declassified document sourced from Cabinets and The Bomb by Peter Hennessy; reproduced verbatim.

TOP SECRET

GEN 75/1

28th August, 1945

THE ATOMIC BOMB

Memorandum by the Prime Minister

A decision on major policy with regard to the atomic bomb is imperative. Until this is taken civil and military departments are unable to plan. It must be recognised that the emergence of this weapon has rendered much of our post-war planning out of date.

For instance a redistribution of industry planned on account of the experience of bombing attacks during the war is quite futile in the face of the atomic bomb. Nothing can alter the fact that the geographical situation of Britain offers to a Continental Power such targets as London and the other great cities. Dispersal of munition works and airfields cannot alter the facts of geography.

Again it would appear that the provision of bomb proof basements in factories and offices and the retention of A.R.P. and Fire Services is just futile waste.

All considerations of strategic bases in the Mediterranean or the East Indies are obsolete. The vulnerability of the heart of the Empire is the one fact that matters. Unless its safety can be secured, it is no use bothering about things on the periphery. It is difficult for people to adjust their minds to an entirely new situation. I noticed at Potsdam that people still talked of the line of the Western Neisse although rivers as strategic frontiers have been obsolete since the advent of Air Power. It is infinitely harder for people to realise that even the modern conception of war to which in my lifetime we have become accustomed is now completely out of date.

We recognise or some of us did before this war that bombing could only be answered by counter bombing. We were right. Berlin and Magdeburg were the answer to London and Coventry. Both derive from Guernica. The answer to an atomic bomb on London is an atomic bomb on another great city.

Duelling with swords and inefficient pistols was bearable. Duelling had to go with the advent of weapons of precision. What is to be done about the atomic bomb? It has been suggested that by a Geneva Convention all nations might agree to abstain from its use. This method is bound to fail as it has failed in the past. Gas was forbidden but used in the first world war. It was not used in World War 2, but its belligerents were armed with it. We should have used it, if the Germans had landed on our beaches. It was not used, because military opinion considered it less effective than explosives and incendiaries.

Further the banning of the atomic bomb would leave us with the other weapons used in the late war which were quite destructive enough.

Scientists agree that we cannot stop the march of discovery. We can assume that any attempt to keep this as a secret in the hands of the U.S.A. and U.K. is useless. Scientists in other countries are certain in time to hit upon the secret.

The most we may have is a few years start. The question is what use we are to make of that few years start.

We might presumably on the strength of our knowledge and of the advanced stage reached in technical development in the U.S.A. seek to set up an Anglo-American Hegemony in the world using our power to enforce a world wide rigid inspection of all laboratories and plants.

I do not think this is desirable or practicable. We should not be able to penetrate the curtain that conceals the vast area of Russia. To attempt this would be to invite a world war leading the the destruction of civilization in a dozen years or so.

The only course which seems to me to be feasible and to offer a reasonable hope of staving off disaster for the world is joint action taken by the U.S.A., U.K. and Russia based on stark reality.

We should...

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: Organizations prioritising neat signals of EA alignment might systematically miss good candidates, published by emp on April 15, 2022 on The Effective Altruism Forum.

Thank you to Peter Wallich, Peter Barnett, and Megan Jamer for their encouragement and feedback on draft versions.

I would like to share my experience applying for junior to mid-level operations roles at EA-aligned organisations. This is my first post on the EA Forum, and I was encouraged to write it for two reasons: to seek community feedback, and to venture what could be a potential blindspot in recruitment faced by EA-aligned orgs. (Noting here that these orgs are not "one thing," but vary team to team.)

Summary (TL;DR)

80,000 Hours and others point to pressing talent gaps, for example in operations. Perhaps part of the problem is that experienced, EA-aligned people go unnoticed because they don't tick clear boxes of EA involvement and are unknown in EA networks. If there are many people like this, then finding them could be a win.

My experience

In my career I have worked for 18 years as a structural engineer and project manager in senior roles at small, niche companies. I've been fully remote for several years, have worked with teams on five continents, and have always worked hard to excel as a people manager and to produce rigorous technical work and robust, efficient systems under challenging conditions.

From 2015 to 2020, I did this work while travelling slowly through sub-Saharan Africa and other parts of the developing world by bicycle. I came to intently question the chasm of inequality between my own life and that of the people I met every day. This led me to build a deep interest and knowledge of EA through investigating the principles and ideas of the movement.

My partner and I ended our journey primarily to try to get more involved in EA. We decided to try pursuing work in the movement. First, I kept working in the engineering world so that she could pursue EA-aligned volunteering full-time. It has been awesome that this translated into contracting and employment. Now we've swapped, and I have been volunteering at two EA-aligned orgs, and have applied for six junior operations roles.

Through EA volunteering work and other recent non-EA work I have been assessed as experienced, engaged, and self-directed in my operations work. I received a single response from the six applications and was removed from consideration there before an interview.

I am fully aware of the competitiveness of EA roles. I'm also aware that there are several potential reasons I did not progress in the application process, and that a team's needs can be idiosyncratic (unfortunately, I have not received feedback yet on what else I could have done to improve my candidacy). I'd love to connect with anyone who might be able to provide feedback on how I could improve how I present myself.

But it does seem like there could be an incongruity here - I do have several years of relevant experience, I am strongly EA aligned, and I didn't receive responses from five applications.

I’ve discussed this with people working in EA orgs and they suggest that it could be because I don’t have any material EA signalling, specifically:

EA work experience (I am volunteering, but it started while I was applying so was not on my CV, and prior to this I had a full time non-EA role)

I’m not yet known in EA networks

No EA Forum posts to my name (I’ve have been passively engaged for 4 years but felt intimidated to comment, kudos to everyone who posts here, rising above the fear and self-doubt)

I don't have shared connections on LinkedIn with other EAs

I didn't have an EA Virtual Programs on my CV (I am now working through the In-Depth EA Program)

I have not applied for 80,000 Hours coaching

I haven't participated in any virtual EA groups yet

I have...