The Nonlinear Library

The Nonlinear Library

By The Nonlinear FundEducation
Download on the App Store

All The Nonlinear Library episodes

  • LW - Bedbugs are a Solved Problem - DIY Bio-Weapon works. by sapphire
    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: Bedbugs are a Solved Problem - DIY Bio-Weapon works., published by sapphire on November 30, 2022 on LessWrong.
    Bed bugs are quite resistant to almost all pesticides; chemical extermination usually fails. Heat treatment works, but you need to heat infected areas to at least 120F. This can be pretty expensive. Landlords are required to pay for bedbug extermination, but they can opt for ineffective chemical treatment. Bedbugs are surprisingly hard to spread but are fairly resilient.
    Luckily the Beauveria bassiana fungus is an effective and cheap treatment. Spores spread where bedbugs will come into contact. Once they germinate, they kill the bedbugs. Spores spread from bug to bug. They also disperse into the environment and survive well as long as the environment is sufficiently humid. The spores can be made more durable by simply mixing them with mineral oil.
    The brand name Aprehend is Beauveria mixed with oil. There might be a cost-effective provider in your area you can pay to treat your home. However, DIY is very cheap. You can just buy Aprehend from diypestcontrol and spread it yourself. An even cheaper option is to buy Beauveria online and mix it with oil yourself. Unless bedbugs evolve resistance or the government restricts DIY, bed bugs are a solved problem.
    Shameless self-promotion: Please follow us on twitter.
    Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.
    2 min
  • EA - Carl Djerassi (1923–2014) by Gavin
    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: Carl Djerassi (1923–2014), published by Gavin on November 29, 2022 on The Effective Altruism Forum.
    Carl Djerassi helped invent the synthetic hormone norethindrone, one of the 500 most important medicines (actually top 50 by prescription count). A large supply is a basic requirement of every health system in the world.
    Norethindrone is important for two reasons. First, it treats menstrual disorders and endometriosis, together 0.3% of the global burden of disease. More famously, it was a component of The Pill. People mix up the timelines, which is why he is sometimes called the 'Father of the Pill'. But "neither Djerassi nor the company he works for, Syntex, had any interest in testing it as a contraceptive" and it was only used for birth control 12 years after.
    As usual in industrial chemistry, Djerassi got no royalties from the blockbuster medicine he helped develop - but, surprise ending! - he bought cheap shares in Syntex and got rich when it became one of the most important medicines in history for two reasons.
    He also synthesized the third-ever practical antihistamine, and applied new instruments in 1,200 papers on the structure of many important steroids. He also worked on one of the first AI programs to do useful work in science.
    Achievements
    Epistemic status: little better than a guess.
    Not many inventions are fully counterfactual; most simple, massively profitable things which get invented would have been invented by someone else a bit later. So the appropriate unit for lauding inventors is years saved. And if I put a number on that I'd just be making it up.
    Here are the numbers I made up:
    About 4 million US users, so maybe up to 94 million world users at present. No sense of the endometriosis / contraception split.
    Call it 600 million users, 10% endometriosis use case.
    For menstrual disorders:
    on the market 65 years and counting.
    Counterfactual: on the market 3 years before the next oral progestogen was.
    It was the first practical oral progestogen, so we should compare to the injectable alternatives
    About 1/6 of Americans hate needles so much that they refuse treatment.
    Attrition and missed doses for needle treatments is higher than pill treatments.
    Endometriosis is about 0.25 - 0.35 QALY loss. So if it's 30% effective, around $30 / QALY, an amazing deal.
    For easy contraception:
    on the market 59 years and counting.
    The big gains (besides autonomy) are averting unintended pregnancies, abortions, and pregnancy-related deaths.
    Modern cost-effectiveness in Ethiopia is $96 / QALY.
    There's probably some additive effect for endometriosis sufferers (who would want contraception anyway).
    A full account would guess the Pill's effect on the sexual revolution and cultural attitudes toward women. But I've reached my limit. (You might also consider the role of the Pill in the ongoing decline of church authority: "1980: In spite of the Pope's ruling against the Pill and birth control, almost 80% of American Catholic women use contraceptives, and only 29% of American priests believe it is intrinsically immoral.")
    How many years did he bring the invention forward? Call it 5.
    Then split the credit three ways with Luis Miramontes and George Rosenkranz.
    So (largely made-up numbers) it looks like millions of QALYs for the treatment overall, and tens of thousands counterfactually for Djerassi.
    Artist
    After surviving cancer, he decided to become a writer.
    I was very depressed, and for the first time thought about mortality. Strangely enough I had not thought about death before... I realized that who knows how long I would live? In cancer they always talk about five years: if one can survive five years then presumably the cancer had been extirpated. And I thought: gee, had I known five years earlier that I would come down with cancer, would I have led a different life during these...
    6 min
  • LW - Distinguishing test from training by So8res
    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: Distinguishing test from training, published by So8res on November 29, 2022 on LessWrong.
    Various people propose some variant of “align AGI by making it sufficiently uncertain about whether it’s in the real world versus still in training”. (The idea being that, insofar as it's in training, betraying the apparent-operators predictably gets high loss; so if the AGI is uncertain about which situation it's in, and assigns sufficiently high probability to "I'm still in training", it will always play it safe and avoid betraying its operators.)
    This approach seems extremely unpromising to me. Setting aside the various reasons why the AGI kills you even if you manage to confuse it (such as an inner alignment failure, where the AGI turns out not to care about the predictably-high outer-loss from betraying the apparent-operators), let's consider why it's tricky to dupe a superintelligence into being uncertain about a fact that you yourself consider obvious.
    Before reading the short list below, I recommend the following exercise. Suppose that some slow and individually-dumb aliens had a copy of your mind, and were putting you through a series of nested simulations. How would you tell when you finally made it to the outside?
    A short list
    (Note: As usual, treating this list as exhaustive is suicidal.)
    1. Reality is the first place that's seriously large.
    1.1. It's the first place large enough to contain a plausible explanation for how the AGI itself actually came to be.
    1.2. It's the first place large enough to contain a plausible explanation for how the operators themselves came to be.
    1.3. It's the first place with far more computing power than the AI itself, as can be readily verified by a variety of experiments (involving, say, fluid dynamics).
    2. Reality is the first place where the AI's history may contain the signature of superintelligent optimization interacting with complex systems.
    For example, suppose the AGI makes superintelligently-motivated trades in a stock market. 24 hours later, there are a billion additional human-intelligence trades made in response to it. The signature of a massive human-intelligence response to superintelligent-trades is implausibly hard for the operators to fake, even if generating a billion fake dumb trades is easy.
    3. Reality is the first place that embeds the AGI and has laws of physics so simple that the hypothesis "this is being simulated somewhere else" is plausibly significantly more complex than "physics is reality".
    Like, various simulated environments that we can concoct have all sorts of weird tics and hiccups and arbitrary choices in them. The source code of minecraft is probably significantly more complicated than the laws of physics. An AI that finds itself in minecraft is plausibly just able to recognize that it's living in the sort of place that, across the multiverse, occurs moreso inside simulations than it occurs "on its own".
    Whereas, while our universe’s mechanics is surely being simulated in other places across the multiverse, it’s simple enough that it perhaps "mostly" occurs in its own right (whatever the heck that means).
    In particular, worlds like ours are (plausibly) resource-rich and base-reality-ish enough to be worth optimizing in their own right, on the grounds that "are we being simulated?" is a wrong question, and "to what degree are we being simulated?" is a better question, to which the answer is "a low degree", in which case the AI might well just notice this and then not care about your insistence that perhaps it's being simulated.
    (If the idea of optimizing all the instances of our universe at once – both the simulated instances and the non-simulated ones – seems odd to you, see Notes on “Can you control the past?” for a fuller discussion.)
    And perhaps that's all confused metaphysical bullshit, but it ha...
    9 min
  • EA - SBF interview with Tiffany Fong by Timothy Chan
    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: SBF interview with Tiffany Fong, published by Timothy Chan on November 29, 2022 on The Effective Altruism Forum.
    Covered here:
    And here:/
    Disclaimer: My transcriptions might contain some inaccuracies.
    The section beginning 11:31 in the second interview might be especially interesting for the readers of the forum. From this and other publicly available information, it seems likely to me that Sam acted as a naive consequentialist. This, of course, isn't mutually exclusive with having elevated dark triad traits (as mentioned here and here).
    Quotes from that section:
    "Honestly, I like, right now, I'm mostly focusing on, what I can do, and like, where I can be helpful, and like, it's... there will be a time and a place for (...) ruminating on my future, but [sighs] I, right now it's more one foot in front of the other, and you know, trying to be as helpful and constructive as I can, and uh, that's all I can do for now, and there's no (...) I don't know what the future will hold for me - it's pretty unclear - it's certainly not the future I once thought it was (...) my future is not the thing that [inaudible] not the thing that matters here - what matters is the world's future -I'm much more worried about the damage that I did to that than whatever happens to me personally."
    "I made a decision a while ago that like, I was gonna, like you know, spend my life trying to do what I could for the world, and like obviously it hasn't turned out like how I had hoped."
    "I feel really really bad for the people who trusted me and believed in me, and, then, you know, we're trying to do great things for the world and tied it to me - and that got, you know, undermined so I fucked up. And that's like... I don't know, that's the shittiest part of it. If it were just myself that it hurt, like, then whatever, but it wasn't."
    Other highlights:
    Sam claims that he donated to Republicans: "I donated to both parties. I donated about the same amount to both parties (...) That was not generally known (...) All my Republican donations were dark (...) and the reason was not for regulatory reasons - it's just that reporters freak the fuck out if you donate to Republicans [inaudible] they're all liberal, and I didn't want to have that fight".
    In response to to his lawyers' advice regarding his public apologies, Sam says he told his lawyers "to go fuck [themselves]" and claims that they "know what they talk about in extremely narrow domain of litigation - they don't understand the broader context of the world, like, if you're a complete dick about everything, even if it narrowly avoids maybe moderately embarrassing statements, it's not helping [mostly inaudible - but maybe he said 'any of them'?]".
    Sam describes the collapse as a "risk management failure". Fong asks: "I mean, you can't be the only person that was like, aware - in charge of all of this." Sam replies: "I think the bigger problem was that there was no [...] person who was chiefly in charge of monitoring the risk of margin positions on FTX. Like there should have been but there wasn't". Later, Sam also says "at the same time, I think, you know, we stretched ourselves too thin. And we're doing a lot of things at the company - and, you know, I think we should have cut a few of them out and focus more on making sure that the fundamental, like, the most important things we were doing well at". Malice and incompetence can mix and match, so that might be the case (in addition to a lack of moral qualms).
    Other relevant videos from Fong:
    Why is Sam Bankman-Fried Talking To Me? (AUDIO CLIP) Phone Call with SBF - Former FTX CEO / Founder
    CONTEXT: My Phone Call with SBF / Sam Bankman-Fried- Former CEO of FTX - Phone Calls About Ch 11
    Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.
    4 min
  • LW - Why Would AI "Aim" To Defeat Humanity? by HoldenKarnofsky
    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: Why Would AI "Aim" To Defeat Humanity?, published by HoldenKarnofsky on November 29, 2022 on LessWrong.
    I’ve argued that AI systems could defeat all of humanity combined, if (for whatever reason) they were directed toward that goal.
    Here I’ll explain why I think they might - in fact - end up directed toward that goal. Even if they’re built and deployed with good intentions.
    In fact, I’ll argue something a bit stronger than that they might end up aimed toward that goal. I’ll argue that if today’s AI development methods lead directly to powerful enough AI systems, disaster is likely1 by default (in the absence of specific countermeasures).
    Unlike other discussions of the AI alignment problem,3 this post will discuss the likelihood4 of AI systems defeating all of humanity (not more general concerns about AIs being misaligned with human intentions), while aiming for plain language, conciseness, and accessibility to laypeople, and focusing on modern AI development paradigms. I make no claims to originality, and list some key sources and inspirations in a footnote.5
    Summary of the piece:
    My basic assumptions. I assume the world could develop extraordinarily powerful AI systems in the coming decades. I previously examined this idea at length in the most important century series.
    Furthermore, in order to simplify the analysis:
    I assume that such systems will be developed using methods similar to today’s leading AI development methods, and in a world that’s otherwise similar to today’s. (I call this nearcasting.)
    I assume that AI companies/projects race forward to build powerful AI systems, without specific attempts to prevent the problems I discuss in this piece. Future pieces will relax this assumption, but I think it is an important starting point to get clarity on what the default looks like.
    AI “aims.” I talk a fair amount about why we might think of AI systems as “aiming” toward certain states of the world. I think this topic causes a lot of confusion, because:
    Often, when people talk about AIs having goals and making plans, it sounds like they’re overly anthropomorphizing AI systems - as if they expect them to have human-like motivations and perhaps evil grins. This can make the whole topic sound wacky and out-of-nowhere.
    But I think there are good reasons to expect that AI systems will “aim” for particular states of the world, much like a chess-playing AI “aims” for a checkmate position - making choices, calculations and even plans to get particular types of outcomes. For example, people might want AI assistants that can creatively come up with unexpected ways of accomplishing whatever goal they’re given (e.g., “Get me a great TV for a great price”), even in some cases manipulating other humans (e.g., by negotiating) to get there. This dynamic is core to the risks I’m most concerned about: I think something that aims for the wrong states of the world is much more dangerous than something that just does incidental or accidental damage.
    Dangerous, unintended aims. I’ll examine what sorts of aims AI systems might end up with, if we use AI development methods like today’s - essentially, “training” them via trial-and-error to accomplish ambitious things humans want.
    Because we ourselves will often be misinformed or confused, we will sometimes give negative reinforcement to AI systems that are actually acting in our best interests and/or giving accurate information, and positive reinforcement to AI systems whose behavior deceives us into thinking things are going well. This means we will be, unwittingly, training AI systems to deceive and manipulate us.
    The idea that AI systems could “deceive” humans - systematically making choices and taking actions that cause them to misunderstand what’s happening in the world - is core to the risk, so I’ll elaborate on this.
    For this and...
    50 min
  • LW - Alignment allows "nonrobust" decision-influences and doesn't require robust grading by TurnTrout
    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 allows "nonrobust" decision-influences and doesn't require robust grading, published by TurnTrout on November 29, 2022 on LessWrong.
    Definition. On how I use words, values are decision-influences (also known as shards). “I value doing well at school” is a short sentence for “in a range of contexts, there exists an influence on my decision-making which upweights actions and plans that lead to e.g. learning and good grades and honor among my classmates.”
    Summaries of key points:
    Nonrobust decision-influences can be OK. A candy-shard contextually influences decision-making. Many policies lead to acquiring lots of candy; the decision-influences don't have to be "globally robust."
    Values steer optimization; they are not optimized against. The value shards aren't getting optimized hard. The value shards are the things which optimize hard, by wielding the rest of the agent's cognition (e.g. the world model, the general-purpose planning API). Since values are not the optimization target of the agent with those values, the values don't have to be adversarially robust.
    Since values steer cognition, reflective agents try to avoid adversarial inputs to their own values. In self-reflective agents which can think about their own thinking, values steer e.g. what plans get considered next. Therefore, these agents convergently avoid adversarial inputs to their currently activated values (e.g. learning), because adversarial inputs would impede fulfillment of those values (e.g. lead to less learning).
    Follow-up to: Don't design agents which exploit adversarial inputs, Don't align agents to evaluations of plans
    I: Nonrobust decision-influences can be OK
    Decision-making influences don'thave to be “robust” in order for a person to value doing well at school. Consider two people with slightly different values:
    One person is slightly more motivated by good grades. They might study for a physics test and focus slightly more on test-taking tricks.
    Another person is slightly more motivated by learning. They might forget about some quizzes because they were too busy reading extracurricular physics books.
    But they might both care about school, in the sense of reliably making decisions on the basis of their school performance, and valuing being a person who gets good grades. Both people are motivated to do well at school, albeit in somewhat different ways. They probably will both get good grades, and they probably will both learn a lot. Different values simply mean that the two people locally make decisions differently.
    If I value candy, that means that my decision-making contains a subroutine which makes me pursue candy in certain situations. Perhaps I eat candy, perhaps I collect candy, perhaps I let children tour my grandiose candy factory. The point is that candy influences my decisions. I am pulled by my choices from pasts without candy to futures with candy.
    So, let C be the set of mental contexts relevant for decision-making, and let A be my action set. My policy has type signature π:CA, and e.g. contains a bunch of shards of value which influence its outputs. The values are subcircuits of my policy network (i.e. my brain). For example, consider a candy shard consisting of the following subshards:
    If center-of-visual-field activates candy’s visual abstraction, then grab the inferred latent object which activated the abstraction.
    If hunger>50 and sugar-level<6, and if current-plan-stub activates candy-obtainable, then tell planning API to set subgoal to obtain candy.
    If heard 'candy' and hunger>20, then salivate.
    Suppose this is the way I value candy. A few thousand subshards which chain into the rest of my cognition and concepts. A few thousand subshards of value which were hammered into place by tens of thousands of reinforcement events across a lifetime of experience.
    This shar...
    32 min
  • EA - Come get malaria with me? by jeberts
    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: Come get malaria with me?, published by jeberts on November 29, 2022 on The Effective Altruism Forum.
    I promise I'm not going to start spamming the forums every week to badger you about getting a different exotic disease (see the post about Zika). But I was accepted into this study in Baltimore, which you can sort of think of as a part-time job fighting malaria from January to early March-ish. I figured some of you may be interested as well. (Please let me know if you sign up for pre-screening! I am always happy to talk about things in more detail if you have questions.)
    If you're not in the DMV, 1Day Sooner keeps track of potentially high-impact studies on our website and via our newsletter; sign up if you want to hear about hot, single challenge trials recruiting in YOUR area in the future.
    The following is a condensed version of what's outlined in this informal document. Obligatory disclosure: Neither I nor my org, 1Day Sooner, represent the study. I'm just excited about it :)
    Why would an EA consider this? Malaria is bad and the vaccine in this trial could result in nontrivial decreases in malaria mortality. What it comes down to is whether the costs — especially time commitment — make sense for you specifically. As the expanded document discusses, the risks of serious complications are very, very low.
    You'll be screened to make sure you're not at any elevated risk, and treatment will be initiated very quickly after you contract malaria (if you do at all). So if you do feel like you have the bandwidth for a part-time malaria-fighting job (remote work capability pretty much necessary), this is a good tangible way to make a difference. Also, you'll get paid. And you'll become friends with me, and I am very fun to be around, in my opinion.
    30-second trial summary: A malaria vaccine candidate with solid chance of eventual deployment in pregnant women needs to undergo this challenge trial being held by U Maryland — Baltimore's Center for Vaccine Development. It's outpatient. The burden of time will very likely be more than the burden of actually being sick. I will be in it, and other EAs and 1Day Sooner volunteers have expressed interest. Compensation runs up to $3,845. 1Day Sooner can help ease burdens of participation, especially transport from DC.
    Slightly expanded summary:
    A promising malaria vaccine, PfSPZ, needs to be tested as a key step in licensure as a traveler's vaccine, which in turn will support eventual deployment among pregnant women.
    This vaccine will not be the final cure humanity has been waiting for. It is most promising for use during pregnancy.
    There is never a guarantee a vaccine candidate will be successful, nor successfully deployed. It may well be that PfSPZ works well, but in a few years another vaccine turns out to be even better.
    The primary burden is time rather than discomfort, i.e., being sick with malaria for at most a few days, but also lots of blood draws. I have a rough estimate of hours spent in this spreadsheet.
    Low estimate for time spent if you are from DC (transit, visits, other lost productive time): 60-70 hours. High estimate: 90-100.
    For Baltimoreans, probably more like 45-65 hours.
    A few of these hours can be productive, like on a train or waiting around at some of the longer visits.
    Vaccination begins in January (specific dates to be announced shortly). There are three doses spaced out across one month. The vaccine has already been tested rather extensively for safety, it is very well tolerated. Three weeks after the final dose, we will be challenged with malaria. One-fourth of the participants will get a placebo injection.
    One week after the malaria challenge, we will begin going into the clinic in Baltimore for short, daily blood tests in the morning. Treatment will be administered ASAP after malaria is detected, definite...
    5 min
  • EA - A Barebones Guide to Mechanistic Interpretability Prerequisites by Neel Nanda
    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 Barebones Guide to Mechanistic Interpretability Prerequisites, published by Neel Nanda on November 29, 2022 on The Effective Altruism Forum.
    Co-authored by Neel Nanda and Jess Smith
    Crossposted on the suggestion of Vasco Grilo
    Why does this exist?
    People often get intimidated when trying to get into AI or AI Alignment research. People often think that the gulf between where they are and where they need to be is huge. This presents practical concerns for people trying to change fields: we all have limited time and energy. And for the most part, people wildly overestimate the actual core skills required.
    This guide is our take on the essential skills required to understand, write code and ideally contribute useful research to mechanistic interpretability. We hope that it’s useful and unintimidating. :)
    Core Skills:
    Maths:
    Linear Algebra: 3Blue1Brown or Linear Algebra Done Right
    Core goals - to deeply & intuitively understand these concepts:
    Basis
    Change of basis
    That a vector space is a geometric object that doesn’t necessarily have a canonical basis
    That a matrix is a linear map between two vector spaces (or from a vector space to itself)
    Bonus things that it’s useful to understand:
    What’s singular value decomposition? Why is it useful?
    What are orthogonal/orthonormal matrices, and how is changing to an orthonormal basis importantly different from just any change of basis?
    What are eigenvalues and eigenvectors, and what do these tell you about a linear map?
    Probability basics
    Basics of distributions: expected value, standard deviation, normal distributions
    Log likelihood
    Maximum value estimators
    Random variables
    Central limit theorem
    Calculus basics
    Gradients
    The chain rule
    The intuition for what backprop is - in particular, grokking the idea that backprop is just the chain rule on multivariate functions
    Coding:
    Python Basics
    The “how to learn coding” market is pretty saturated - there’s a lot of good stuff out there! And not really a clear best one.
    Zac Hatfield-Dodds recommends Al Sweigart's Automate the Boring Stuff and then Beyond the Basic Stuff (both readable for free on inventwithpython.com, or purchasable in books); he's also written some books of exercises. If you prefer a more traditional textbook, Think Python 2e is excellent and also available freely online.
    NumPy Basics
    Try to do the first ~third of these:. Bonus points for doing them in pytorch on tensors :)
    ML:
    Rough grounding in ML.
    fast.ai is a good intro, but a fair bit more effort than is necessary. For an 80/20, focus on Andrej Karpathy’s new video explaining neural nets:
    PyTorch basics
    Don’t go overboard here. You’ll pick up what you need over time - learning to google things when you get confused or stuck is most of the real skill in programming.
    One goal: build linear regression that runs in Google Colab on a GPU.
    Transformers - probably the biggest way mechanistic interpretability differs from normal ML is that it’s really important to deeply understand the architectures of the models you use, all of the moving parts inside of them, and how they fit together. In this case, the main architecture that matters is a transformer! (This is useful in normal ML too, but you can often get away with treating the model as a black box)
    Check out the illustrated transformer
    Note that you can pretty much ignore the stuff on encoder vs decoder transformers - we mostly care about autoregressive decoder-only transformers like GPT-2, which means that each token can only see tokens before it, and they learn to predict the next token
    Good (but hard) exercise: Code your own tiny GPT-2 and train it. If you can do this, I’d say that you basically fully understand the transformer architecture.
    Example of basic training boilerplate and train script
    The EasyTransformer codebase is probably good to riff off of here
    An ...
    6 min
  • EA - Apply to join Rethink Priorities’ board of directors. by abrahamrowe
    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: Apply to join Rethink Priorities’ board of directors., published by abrahamrowe on November 29, 2022 on The Effective Altruism Forum.
    Right now (through January 13th), you can apply to join RP’s board of directors in an unpaid (3-10 hours per month) or paid (10-15 hours per week) capacity.
    Rethink Priorities (RP) has grown quickly, is now large, and remains ambitious (see our recent post for more details). We're looking for people to join our board who can help RP really secure its foundations, and scale in the next several years. While we had been planning on opening these roles prior to FTX’s collapse because we recognized governance as an area of growth for our organization, the recent events help in highlighting why these roles are important. We want to ensure that RP is healthy and sustainable, and thinking about risk and success well in the long-run.
    Our board of directors plays an active role in ensuring that our senior management is making responsible, legal, and risk-aware decisions for the organization in the long-run. They evaluate things like our financial controls, the performance of our Co-CEOs, and budgets and fundraising to help ensure the organization is acting legally and ethically. They also advise our senior management to help ensure the organization stays on track, and continues to target high-level goals for itself.
    If you have any questions about these positions, please contact [email protected]. If you have questions about RP’s governance generally, contact [email protected].
    What does the board of directors do?
    Our board’s primary functions are:
    Providing long-term financial oversight to the organization including:
    Reviewing and approving the annual budget, and spending controls for the Co-CEOs
    Reviewing annual audits of financial statements and financial controls
    Providing oversight for the Co-CEOs including:
    Performance evaluations of senior management
    Serving as contacts for staff outside the chain of commands
    Providing feedback on Co-CEOs strategic plans
    Providing legal oversight for the organization, such as:
    Helping assess risky and complicated situations, and providing feedback on plans to navigate those situations
    Ensuring that RP is compliant with its charitable purposes
    Advising on RP’s long-term strategy and direction
    What qualifications are you looking for in board members?
    We are particularly interested in adding individuals who have knowledge/experience within longtermism, launching/supporting new ventures, and/or scaling organizations. We’d also be excited for candidates with professional legal or nonprofit finance experience.
    Do I have to be an American to join the board?
    No! Though we are a US-based organization, these roles do not require US residency. However, we’d like the majority of our board to be made up of US residents (including non-US citizens), and some board functions may require US residency, so while location wouldn’t be disqualifying, it may be a consideration.
    What’s the difference between paid and unpaid roles?
    The majority of our board is required by our bylaws to be unpaid. However, we think that there is significant value in our board being more engaged than many members are able to be in a voluntary capacity, so we’d like to pay up to 2 members of the board to provide administrative assistance to the other members, and to tackle some of the more work intensive tasks (such as performance evaluations of senior management). In our view, a failure for many nonprofit boards is they select for skills but not time, and that contributes towards a tendency for boards to not do a very thorough job. We're excited to experiment with one to two people who are designated to spend 10-15 hrs/week on board duties. Right now we have some idea of how this will work, but it will be a first for us. We ...
    5 min
  • EA - Double your donation with matching opportunities by BarryGrimes
    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: Double your donation with matching opportunities, published by BarryGrimes on November 28, 2022 on The Effective Altruism Forum.
    There are several matching opportunities running this Giving Season. If you know of any that I've missed, please mention them in the comments.
    Double Up Drive
    The 2022 Double Up Drive will start at 8:00 am PT / 11:00 am ET / 4:00 pm GMT on Giving Tuesday (29 November).
    I don’t know how big the matched pool will be or how long it will last so it's best to donate as soon as it opens.
    2022 charities
    Against Malaria Foundation
    Animal Charity Evaluators
    Clean Air Task Force
    Evidence Action
    Founders Pledge
    International Refugee Assistance Project
    New Incentives
    StrongMinds
    The Good Food Institute
    The Life You Can Save
    UBS Optimus Foundation (StrongMinds)
    All donations to StrongMinds will be matched up to $800,000 or until the end of 2022 (whichever comes first)
    Animal Charity Evaluators
    Until December 31, any donation you make to ACE’s Recommended Charity Fund will be matched up to $300,000, thanks to the estate of a generous legacy donor.
    Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.
    2 min