The Nonlinear Library

The Nonlinear Library

By The Nonlinear FundEducation
Download on the App Store

All The Nonlinear Library episodes

  • AF - Provisionality by Tsvi Benson-Tilsen
    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: Provisionality, published by Tsvi Benson-Tilsen on June 19, 2023 on The AI Alignment Forum.
    [Metadata: crossposted from. First completed March 7, 2023.]
    A mental element has to be open to revision, and so it has to be treated as though it might be revised.
    Thanks to Sam Eisenstat for relevant conversations and for the basic notion of provisionality.
    Terms
    Provisionality
    A mental element E is provisional when it is suitable to treat E as open to revision, i.e. as though it might be suitable to revise E in the future. Provisionality is the state of being provisional.
    Many elements are provisional: they have yet to be fully grasped, made explicit, connected to what they should be connected to, had all their components implemented, carved at the joints, made available for use, indexed, suitably redescribed in other terms.
    An element is essentially provisional when it is, in its essence, provisional. That is, an essentially provisional element is, by its nature and by its role in the mind, provisional: in all (or a great range of) possible worlds the element is always provisional, even after more explicitizing and correcting has happened.
    Provisionality is normative
    Provisionality isn't how a mind actually treats the element. A mind might treat an element as-if-provisionally even though really the element is not provisional (i.e., really it ought to be treated as finalized), and a mind might treat an element not as-if-provisionally even though the element really is provisional.
    Provisionality is also not whether the element will actually be revised.
    So, provisionality is normative (it describes "right" behavior, and so is an imperative for an agent behaving "rightly"), while as-if-provisionality is descriptive (describes actual behavior). Revision of an element is a kind of witness to its prior provisionality, since probably if the element actually gets revised, it should have been treated as though it will be revised.
    Openness
    An element E is open when it is suitable to treat E as non-closed, non-circumscribed, open to novel relations; that is, as though it might in the future be suitable to relate E to more elements than it currently relates to. Openness is the state of being open. Openness could also be called noncontainedness.
    As with provisionality, openness is normative.
    Openness is a subclass of provisionality. If an element E is open then E is also provisional in that way, because a novel relation with E is a kind of revision of E. Updating a probability distribution can be a substantial revision that's not (much of) a novel relation, witnessing provisionality but not openness. (On the other hand, to deduce consequences of a hypothesis witnesses openness, because it brings more kinds of evidence to bear on the hypothesis.)
    Essentiality
    Essentiality is when a property holds of a thing by the nature of that thing: in all (or a great range of) possible worlds, the property holds of the thing.
    An example is essentially undecidable logical theories: theories which can't be consistently extended to a decidable theory.
    A class of examples comes from logical implication, i.e. analyticity, as in "Pediatricians are essentially doctors.". (The concept of analytic is problematic because it relies somewhat circularly on notions of logical rules and implication, see Quine's "Two Dogmas of Empiricism", but relative analyticity (or enthymemic analyticity, understood to always be relative to unstated premises) seems less problematic.)
    In terms of a thing as an inductive nexus of reference, essentiality means centrality or eventual permanence. A thing is essentially P if, following the reference of the thing however deep into its nexus, whatever structure is there will satisfy property P.
    Quoting from "Rootedness" about essential provisionality and essential openness:
    The novelty c...
    11 min
  • LW - Guide to rationalist interior decorating by mingyuan
    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: Guide to rationalist interior decorating, published by mingyuan on June 19, 2023 on LessWrong.
    Recently someone asked me to write a guide to rationalist interior decorating, since there’s a set of products and best practices (originating with CFAR and Lightcone) that have gotten wide adoption. I’m perhaps not the very most qualified person to write this post, but I’ve been into interior decorating since before the Lightcone team got into it, and I basically know what they do, plus they’re all very busy whereas I wasn’t doing anything else with my time anyway. So here’s this post, which I have written all by myself like a loose cannon; blame me for everything.
    I should point out that this post is anthropological, not normative. That is to say, this isn't a description of what I believe to be ‘optimal’ interior decorating; instead it's a guide to how to make your space more like what's come to be standard for rationalist coworking spaces and group houses in Berkeley (and to a lesser extent elsewhere — I've seen a Lightcone-style coworking space in New York and lumenators in Oxford!). That said, I do think the reason a lot of the products have gotten adoption is that they’re pretty good!
    Lighting
    Lighting is so important. The wrong light can give you a migraine or make you depressed. The right light can make you marvel in wonder at its beauty (like sunshine through a forest canopy, or incandescent fairy lights). I love lighting. All hail lighting.
    Lighting basics
    I’m going to cover mostly what you should be looking for when buying (mainly LED) lights, but if you’re interested in the topic of interior lighting more broadly, I recommend this article, which goes into more depth and technical details.
    You should always be able to find the CRI, color temperature, and lumens of a product listed on the box or the online product page. I don’t think these should be different in different countries.
    CRI (color rendering index)
    CRI, or color rendering index, “is the measurement of how light affects how you see color”. You ideally want lights with CRI >95 but this is hard to find; 90+ is a more realistic goal and will be nearly as good, while 80+ is cheaper but noticeably worse.
    CRI only applies to buying LEDs — incandescent lights all have a CRI of 100 because they, like the sun, produce light via blackbody radiation.
    Color temperature
    Color temperature matters a lot when choosing lights! 2700K (‘warm white’) will be noticeably yellow, 5000K (‘daylight’) will be noticeably blue, and 4000K (‘soft white’) is approximately white. The K stands for Kelvin, by analogy to blackbody radiators.
    Color temperature is largely a matter of personal preference, but that doesn’t mean it isn’t important. The original lumenators used a mix of 2700K and 5000K bulbs (to imitate sunlight?), but personally I find the bluer bulbs depressing, so I use all 2700K.
    Lumens
    Lumens are a measure of brightness. As a reference point, the most common type of incandescent bulb (60 Watts) was about 800 lumens. Brighter bulbs draw more power (though this is mostly negligible for LEDs) and are generally more expensive.
    Lumens specifically measure how much light something emits; note that the placement of lights and the characteristics of the room will affect how much light you actually receive. I like this article on the difference between lux and lumens (mainly because of the cool graphic).
    Diffusion
    This one isn’t a technical term, but everyone knows that looking directly at a bright light is painful / unpleasant / maybe bad for your vision. If you’re using really bright lights, you’ll probably want something that diffuses the light without absorbing too much of it. This is the purpose of lampshades, light fixtures, and the paper lanterns used on lumenators.
    Bright lighting
    Rationalists first got into really bright...
    18 min
  • EA - We can all help solve funding constraints. What stops us? by Luke Freeman
    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: We can all help solve funding constraints. What stops us?, published by Luke Freeman on June 19, 2023 on The Effective Altruism Forum.
    This post is a personal reflection that follows my journey to effective altruism, my experiences within it, the concerns I've developed along the way and my hopes for addressing them. It culminates in my views on funding constraints — the role we can all play in solving them and a key question I have for you all: What stops us? (Please let me know in the comments).
    My journey
    While this starts with a reflection on my personal journey, I suspect it might feel familiar, it might strike a chord, at times it might rhyme with yours.
    I was about eight years old when I was first confronted with the tragic reality that an overwhelming number of children my age were suffering and dying from preventable diseases and unjust economic conditions.
    It broke my heart.
    I knew that I had done nothing to deserve my incredibly privileged position of being born healthy to a loving, stable, middle-income family in Australia (a country with one of the highest standards of living).
    Throughout my early years, I took many opportunities to do what I could to right this wrong. In school, that meant participating in fundraisers and advocacy. As a young professional, that meant living frugally but still giving a relatively meagre amount to help others. When I got my first stable job, I decided it was time to give 10% to help others... But when I calculated that that would be $5,000, this commitment began to feel like a pretty big deal. I wasn't going to back down, but I wanted to be more confident that it'd actually result in something good. I felt a responsibility to donate wisely.
    Some Googling quickly led me to discover Giving What We Can, GiveWell, and Julia Wise's blog Giving Gladly. From this first introduction to what would soon be known as the effective altruism (EA) community, I found the information I needed to help guide me, and the inspiration I needed to help me follow through.
    I also took several opportunities to pursue a more impact-oriented career, and even tried getting involved in politics. These attempts had varying success, but that was okay: I had one constant opportunity to help others by giving.
    Around this time, the EA community started expanding their lines of reasoning beyond effective giving advice to other areas like careers and advocacy. I was thrilled to see this. We all have an opportunity to use various resources to make a dent in the world's problems, and the same community that had made good progress on philanthropy seemed to me well-positioned to make progress on other fronts too.
    By 2016, effective altruism was well and truly “a thing” and I discovered that there was an EA group and conference near me. So, I ventured out to actually meet some of these "effective altruism" people in person.
    It hit me: I'd finally found "my people."
    These were people who actually cared enough to put their money where their mouths were, to use the best tools they could find to make the biggest possible difference, and to advocate for others to join them. None of these things were easy, but these people really owned the stakes and did the work.
    I admired the integrity and true altruism that I found. It motivated me to do better.
    How I saw effective altruism change
    As time went on, however, I noticed some changes that concerned me. The EA community’s expanded focus started to feel less like a "yes, and" message — supporting both effective giving and pursuing other effective paths to impact — and more like a "no, instead" message: giving began to feel a bit passé within the community.
    Above: My response to the shift away from effective giving
    It started slow, but the change became overwhelming:
    2015: 80,000 Hours started to advocate to focus on talent...
    19 min
  • EA - My lab's small AI safety agenda by Jobst Heitzig (vodle.it)
    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: My lab's small AI safety agenda, published by Jobst Heitzig (vodle.it) on June 18, 2023 on The Effective Altruism Forum.
    My lab has started devoting some resources to AI safety work. As a transparency measure and to reach out, I here describe our approach.
    Overall Approach
    I select small theoretical and practical work packages that...
    seem manageable in view of our very limited resources,
    match our mixed background in applied machine learning, game theory, agent-based modeling, complex networks science, dynamical systems theory, social choice theory, mechanism design, environmental economics, behvioural social science, pure mathematics, and applied statistics, and
    appear under-explored or neglected but promising or even necessary, according to our subjective assessment based on our reading of the literature and exchanges with individuals from applied machine learning, computer linguistics, AI ethics researchers, and most importantly, AI alignment researchers (you?).
    Initial Reasoning
    I believe that the following are likely to hold:
    We don't want the world to develop into a very low-welfare state.
    Powerful AI agents that optimizes for an objective not almost perfectly aligned with welfare can produce very low-welfare states.
    Powerful AI agents will emerge soon enough.
    It is impossible to specify sufficiently well what "welfare" means (welfare theorists have tried for centuries and still disagree, common people disagree even more).
    My puzzling conclusion from this is:
    We can't make sure that powerful AI agents optimize for an objective that is almost perfectly aligned with welfare.
    Hence we must try to prevent that any powerful AI agent optimizes for any objective whatsoever.
    Those of you who are Asimov fans like me might like the following...
    Six Laws of Non-Optimizing
    Never attempt to optimize your behavior with regards to any metric.
    Constrained by 1, don't cause suffering or do other harm.
    Constrained by 1-2, prevent other agents from violating 1. or 2.
    Constrained by 1-3, do what the stakeholders in your behavior would collectively decide you should do.
    Constrained by 1-4, cooperate with other agents.
    Constrained by 1-5, protect and improve yourself.
    Rather than trying to formalize this or even define the terms precisely, I just use them to roughly guide my work.
    When saying "optimize" I mean it in the strict mathematical sense: aiming to find an exact or approximate, local or global maximum or minimum of some function. When I mean mere improvements w.r.t. some metric, I just say "improve" rather than "optimize".
    Agenda
    We currently slowly pursue two parallel approaches, the first related to laws 1,3,5 from above, the other related to law 4.
    Non-Optimizing Agents
    Explore several novel variants of "satisficing" policies and related learning algorithms for POMDPs, produce corresponding non-optimizing versions of classical to state-of-the art tabular and ANN-based RL algorithms, and test and evaluate them in benchmark and safety-relevant environments from the literature, plus in tailormade environments for testing particular hypotheses. (Currently underway)
    Test them in near-term relevant application areas such as autonomous vehicles, via state-of-the-art complex simulation environments. (Planned with partner from autonomous vehicles research)
    Using our game-theoretical and agent-based modeling expertise, study them in multi-agent environments both theoretically and numerically.
    Design evolutionarily stable non-optimizing strategies for non-optimizing agents that cooperate with others to punish violations of law 1 in paradigmatic evolutionary games.
    Use our expertise in adaptive complex networks and dynamical systems theory to study dynamical properties of mixed populations of optimizing and non-optimizing agents: attractors, basins of attraction, their stability and...
    6 min
  • LW - Solomonoff induction still works if the universe is uncomputable, and its usefulness doesn't require knowing Occam's razor by Christopher King
    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: Solomonoff induction still works if the universe is uncomputable, and its usefulness doesn't require knowing Occam's razor, published by Christopher King on June 18, 2023 on LessWrong.
    Note: I don't think this idea is original, but I couldn't find a good post going over the implications.
    I used to think that Solomonoff induction was a bit arbitrary for the following reason: it assigned a 100% probability to the universe being computable. I'm pretty sure the universe is computable (ignoring randomness), but nowhere near 100% sure. Who's to say we won't find a halting oracle floating in space for no reason? That seems like a pretty simple hypothesis. Why the focus on recursive languages. You have to make some choice of how descriptions work (you can't assign positive probability to every infinite bit string), but that didn't change the feelings of arbitrariness.
    But then I realized this understanding of why to use Solomonoff induction is incorrect. We do not use it because of the physical church-turing thesis, we use it because of the original church-turing thesis:
    L.C.M.s [logical computing machines: Turing’s expression for Turing machines] can do anything that could be described as ‘rule of thumb’ or ‘purely mechanical’. - Alan Turing
    Because what matters is not whether the universe is computable, but whether our methods of reasoning are computable. Or in other words, whether the map is computable. Solomonoff's induction is at least as "good" as any computable inference method (up to a constant), regardless of the complexity of the universe. So if you, as a human, are trying to come up with a systematic way to predict things (even uncomputable things), Solomonoff's induction is better. Here is the precise statement:
    Theorem: Let D be some probability distributions on infinite sequences of bits such that inferring the next bit from a prefix is computable. The likelihood ratio from D to Solomonoff induction's prior is bounded above by some finite constant (despite the sequence containing infinitely many bits), and this constant is independent of the sequence of bits.
    Proof sketch: (Note: this is already a well-known result.) There is a program m that is a decoder for an entropy code over finite bit strings based on D. Given any finite bit string x, we can find a y such that m(y)=x. The length of y is approximately −log2(px) (where px is the probability of a bit string starting with x according to D). The probability of a string starting with x under Solomonoff induction's prior is greater than 2−|m(y)|=2−|m|2log2(px)=px2|m|. So regardless of the content or length of |x|, the ratio is bounded by 2|m| (which only depends on D). □
    (Notice how we never invoked Occam's razor to argue that Solomonoff induction is superior. We can instead go the other way; Occam's razor is good because it's an informal version of an ideal inference procedure.)
    Solomonoff induction v.s. a human in an uncomputable world
    How does this shake out in an uncomputable universe? What if you're convinced that there is some orb in the universe emitting the digits of Chaitin's constant or something? We'll let Alice be a reasoner using Solomonoff induction, and Bob be a human.
    Bob: Ah yes, I have found it! An orb emitting the digits of Chaitin's constant is floating around in our solar system! Alice: How do you figure? Bob: I calculated the first two digits, and they matched! Alice: Surprising! But not that surprising (about six and half bits). Bob: I was surprised too, but now I can do something better than you. I can predict the digits that orb will emit. Alice: How do you plan to predict an uncomputable sequence, given that you're a human? Bob: Oh yeah... Alice: In fact, if you're correct it will eventually look like a uniformly random sequence to you since it is algorithmically random. So I'll be able...
    7 min
  • LW - A summary of current work in AI governance by constructive
    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 summary of current work in AI governance, published by constructive on June 17, 2023 on LessWrong.
    A summary of current work in AI governance
    If you’d like to learn more about AI governance, apply to the AI Safety Fundamentals: Governance Track, a 12-week, part-time fellowship before June 25.
    Context
    For the past nine months, I spent ~50% of my time upskilling in AI alignment and governance alongside my role as a research assistant in compute governance.
    While I discovered great writing characterizing AI governance on a high level, few texts covered which work is currently ongoing. To improve my understanding of the current landscape, I began compiling different lines of work and made a presentation. People liked my presentation and suggested I could publish this as a blog post.
    Disclaimers:
    I’ve only started working in the field ~9 months ago
    I haven’t run this by any of the organizations I am mentioning. My impression of their work is likely different from their intent behind it.
    I’m biased toward the work by GovAI as I engage with that most.
    My list is far from comprehensive.
    What is AI governance?
    Note that I am primarily discussing AI governance in the context of preventing existential risks.
    Matthijs Maas defines AI long-term governance as
    “The study and shaping of local and global governance systems—including norms, policies, laws, processes, politics, and institutions—that affect the research, development, deployment, and use of existing and future AI systems in ways that positively shape societal outcomes into the long-term future.”
    Considering this, I want to point out:
    AI governance is not just government policy, but involves a large range of actors. (In fact, the most important decisions in AI governance are currently being made at major AI labs rather than at governments.)
    The field is broad. Rather than only preventing misalignment, AI governance is concerned with a variety of ways in which future AI systems could impact the long-term prospects of humanity.
    Since "long-term" somewhat implies that those decisions are far away, another term used to describe the field is “governance of advanced AI systems.”
    Threat Models
    Researchers and policymakers in AI governance are concerned with a range of threat models from the development of advanced AI systems. For an overview, I highly recommend Allan Dafoe’s research agenda and Sam Clarke’s "Classifying sources of AI x-risk".
    To illustrate this point, I will briefly describe some of the main threat models discussed in AI governance.
    Feel free to skip right to the main part.
    Takeover by an uncontrollable, agentic AI system
    This is the most prominent threat model and the focus of most AI safety research. It focuses on the possibility that future AI systems may exceed humans in critical capabilities such as deception and strategic planning. If such models develop adversarial goals, they could attempt and succeed at permanently disempowering humanity.
    Prominent examples of where this threat model has been articulated:
    Is power-seeking AI an existential risk?, Joe Carlsmith, 2022
    AGI Ruin: A list of lethalities, Eliezer Yudkowsky, 2022 (In a very strong form, see also this in-depth response from Paul Christiano)
    The alignment problem from a deep learning perspective, Ngo et al., 2022
    Loss of control through automation
    Even if AI systems remain predominantly non-agentic, the increasing automation of societal and economic decision-making, driven by market incentives and corporate control, could pose the risk of humanity gradually losing control - e.g., if the optimized measures are only coarse proxies of what humans value and the complexity of emerging systems is incomprehensible to human decision-makers.
    This threat model is somewhat harder to convey but has been articulated well in the following texts:
    Will Humanit...
    21 min
  • EA - SoGive rates Open-Phil-funded charity NTI “too rich” by Sanjay
    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: SoGive rates Open-Phil-funded charity NTI “too rich”, published by Sanjay on June 18, 2023 on The Effective Altruism Forum.
    Exec summary
    Under SoGive’s methodology, charities holding more than 1.5 years’ expenditure are typically rated “too rich”, in the absence of a strong reason to judge otherwise. (more)
    Our level of confidence in the appropriateness of this policy depends on fundamental ethical considerations, and could be “clearly (c.95%) very well justified” or “c.50% to c.90% confident in this policy, depending on the charity” (more)
    We understand that the Nuclear Threat Initiative (NTI) holds > 4 years of spend (c$85m), as at the most recently published Form 990, well in excess of our warning threshold. (more)
    We are now around 90% confident that NTI’s reserves are well in excess of our warning threshold, indeed >3x annual spend, although there are some caveats. (more)
    Our conversation with NTI about this provides little reason to believe that we should deviate from our default rating of “too rich”. (more)
    It is possible that NTI could show us forecasts of their future income and spend that might make us less likely to be concerned about the value of donations to NTI, although this seems unlikely since they have already indicated that they do not wish to share this. (more)
    We do not typically recommend that donors donate to NTI. However we do think it’s valuable for donors to communicate that they are interested in supporting their work, but are avoiding donating to NTI because of their high reserves. (more)
    Although this post is primarily to help donors decide whether to donate to NTI, readers may find it interesting for understanding SoGive's approach to charities which are too rich, and how this interacts with different ethical systems.
    We thank NTI for agreeing to discuss this with us knowing that there was a good chance that we might publish something on the back of the discussion. We showed them a draft of this post before publishing; they indicated that they disagree with the premise of the piece, but declined to indicate what specifically they disagreed with.
    0. Intent of this post
    Although this post highlights the fact that NTI has received funding from Open Philanthropy (Open Phil), the aim is not to put Open Philanthropy on the spot or demand any response from them.
    Rather, we have argued that it is often a good idea for donors to “coattail” (i.e. copy) donations made by Open Phil. For donors doing this, including donors supported by SoGive, we think it’s useful to know which Open Phil grantees we might give lower or higher priority to.
    1. Background on SoGive’s methodology for assessing reserves
    The SoGive ratings scale has a category called “too rich”. It is used for charities which we deem to have a large enough amount of money that it no longer makes sense for donors to provide them with funds. We set this threshold at 18 months of spend (i.e. if the amount of unrestricted reserves is one and a half times as big as its annual spend then we typically deem the charity “too rich”). To be clear, this allows the charity carte blanche to hold as much money as it likes as long as it indicates that it has a non-binding plan for that money.
    So, having generously ignored the designated reserves, we then notionally apply the (normally severe) stress of all the income disappearing overnight. Our threshold considers the scenario where the charity has so much reserves that it could go for one and a half years without even having to take management actions such as downsizing its activities.
    In this scenario, we think it is likely better for donors to send their donations elsewhere, and allow the charity to use up its reserves.
    Originally we considered a different, possibly more lenient policy. We considered that charities should be considered too rich if they...
    25 min
  • LW - UK Foundation Model Task Force - Expression of Interest by ojorgensen
    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: UK Foundation Model Task Force - Expression of Interest, published by ojorgensen on June 18, 2023 on LessWrong.
    Ian Hogarth has just been announced as the Chair of the UK's AI Foundation Model Taskforce. He's the author of the FT article "We must slow down the race to God-like AGI", and seems to take X-risks from AI seriously.To quote his twitter thread:
    And to that end I put out a call to people across the world. If you are an AI specialist or safety researcher who wants to build out state capacity in AI safety and help shape the future of AI policy then get in touch:We have £100m to spend on AI safety and the first global conference to prepare for. I want to hear from you and how you think you can help. The time is now and we need more people to step up and help.
    The google form to leave an expression of interest is here.
    (I am in no way affiliated with Ian or the UK Foundation Model Task Force)
    Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.
    2 min
  • EA - Question and Answer-based EA Communities by Joey
    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: Question and Answer-based EA Communities, published by Joey on June 18, 2023 on The Effective Altruism Forum.
    The EA community has expanded to encompass a broad spectrum of interests, making its identity and definition a hotly debated topic. In my view, the community's current diversity could easily support multiple distinct communities, and if we were building a movement from scratch, it would likely look different from the current EA movement.
    Defining sub-communities within the EA movement can be approached in numerous ways. One proposed division that I believe captures much of what people appreciate about the EA community, is as follows:
    Question-based communities
    An Effective Giving Community
    An Impactful Career Community
    Answer-based communities
    An AI X-Risk Community
    An Effective Animal Advocacy Community
    Question-based communities
    An Effective Giving Community
    The concept of effective giving is where EA originated and remains a significant component of the community. Notable organizations such as GWWC, Effektiv Spenden, One for the World, Founders Pledge, and others, share a common mission and practical outcomes. The primary metric for this community is directing funds towards highly impactful areas. GiveWell, for instance, is perhaps the first and most recognized organization within this effective giving community outside the EA movement. This community benefits from its diversity and plurality, as many people could, for example, take the 10% pledge, and an even larger number could enhance their giving effectiveness using EA principles. Key concepts for this community could include determining the best charities to donate to, identifying the most effective charity evaluators, and deciding how much one should donate. This, in many ways, echoes the fundamentals of the EA 1.0 community.
    An Impactful Career Community
    In addition to funding, individuals can contribute to the world through their careers. Much like the effective giving community, there's the question of how to maximize the impact of one's career across multiple cause areas. Organizations such as Probably Good, High Impact Professionals, or Charity Entrepreneurship focus on this area (I intentionally exclude career-focused organizations with a narrow cause area focus, like 80,000 Hours or Animal Advocacy Careers). The objective of this community would be related to career changes and enhancing understanding of the most impactful career paths. Although this is a broadly inclusive community benefiting from cause plurality, it's likely less extensive than the effective giving community, as a smaller percentage of the population will prioritize impact when considering a career switch.
    Relevant topics for this community could include identifying high absorbency, impactful careers, assessing the most impactful paths for individuals with specific value or skill sets, and determining underrated careers.
    Answer-based communities, e.g., AI X-Risk Community
    The second community category that is a bit different from these others is anwer-based communities. I think there are two somewhat distinctive answer-based communities in EA: AI and animals. I think AI X-risk is a better example as it's more often mixed with the other above two communities and has significantly grown as a unique area within EA. This community consists of meta-organizations like Longview, Effective Giving and 80,000 Hours as well as the organizations working directly on the problem. It has begun to hold separate forums, conferences, and events. Its shared goal is to mitigate existential risks from AI, a specific objective that doesn't necessarily require members to embrace effective giving or prioritize impact in their careers. However, it does require specific values and epistemic assumptions, leading to this cause being prioritized over ot...
    6 min
  • LW - Cryonics Career Survey (more jobs than you think) by Mati Roy
    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: Cryonics Career Survey (more jobs than you think), published by Mati Roy on June 18, 2023 on LessWrong.
    (x-post LessDead)
    Survey presentation
    Link to survey: Cryonics Career Survey
    Background. During the annual Teens and Twenties Cryonics Gathering (a fully funded yearly gathering for young cryonics members), there is usually a survey that is distributed to all participants to gauge their interests around getting involved in various roles in cryonics, professionally or as a volunteer. This has been useful, and I think it would be even more valuable if it were accessible for anyone to fill out–so I made my own version.
    Goals. The goals of this survey are to gather people that are either:
    (potentially) interested in utilizing their skills/knowledge/experience to support cryonics endeavors – there are many work opportunities you probably don’t know about;
    (potentially) interested in learning and seeking guidance/mentorships/information about having a career in cryonics – we’re already in touch with cryonics researchers willing to mentor.
    Note: Ramping up to being able to do meaningful cryonics research might be easier than you think, so I invite you to err on the side of signaling your interest to us.
    Email use. We will email the list with work opportunities, as well as (if you opt-in) share your information with interested and approved cryonics(-relevant) organizations for the purpose of helping them find people to work for them. We will also email the list with any information relevant to working in cryonics, such as relevant interviews of cryonics researchers or mentorship opportunities. We expect to send emails infrequently—less than once per month. You can unsubscribe at any time of course.
    Related opportunities:
    If you’re an MD and/or PhD and support the Scientists’ Open Letter on Cryonics, please also let us know.
    If you’re involved in cryonics and willing to mentor others, please reach out!
    Fill out the survey here: Cryonics Career Survey
    Currently open roles
    After an extensive web search of cryonics organizations we know of, here is a list of some current job postings we were able to find (when the country isn’t specified it means it’s in the USA or remote) – some of those roles might have recently close, but I’m leaving them up to show the diversity of options (most or all should still be open at the time of posting though). Fill out the survey to get information about future such opportunities.
    Advanced Neural Biosciences (Portland, OR)
    Research Perfusionist (LinkedIn)
    Alcor Life Extension Foundation (Scottsdale)
    CPA/Corporate Controller/Bookkeeper (LinkedIn)
    Senior Accountant (LinkedIn)
    Paramedics/EMTs (Ongoing Contracts) (Alcor’s website)
    Volunteer (Alcor’s website)
    Biostasis Technologies (New York City, NY)
    Maker / Fabricator (private communication)
    Brain Preservation Foundation (remote)
    Volunteer (Brain Preservation Foundation’s website)
    European Biostasis Foundation (Switzerland)
    Facility Manager for Research Institute (Full-Time) (Join.com)
    Senior Researcher/Medical Engineer (Full-Time) (Join.com)
    Tomorrow Biostasis (Germany)
    Full-Time Junior Product Manager - Berlin (Join.com)
    Internship - Business Development (Join.com)
    Part-time account manager (Join.com)
    Full-time Engineering Internship (m/w/x) - Berlin (Join.com)
    Trans Time Inc. (Bay Area)
    Marketing Specialist (Trans Time)
    Southern Cryonics (Australia)
    Facility Caretaker (Part-time) (Southern Cryonics)
    Volunteer (Southern Cryonics)
    21st Century Medicine (Fontana, California)
    Professional Research Surgeon (21st Century Medicine’s website)
    Lorentz Bio (Bay Area)
    Lab Manager (Lever.co)
    Research Assistant (Tissue/Cell Culture) (Lever.co)
    Research Assistant II/III (Lever.co)
    Research Associate (Neuroscience) (LinkedIn)
    Nectome (San Francisco)
    Senior Scientist (Nectome’s website)
    Prototype Engineer (Nect...
    5 min