The Nonlinear Library

The Nonlinear Library

By The Nonlinear FundEducation
Download on the App Store

All The Nonlinear Library episodes

  • LW - Take 7: You should talk about "the human's utility function" less. by Charlie Steiner
    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: Take 7: You should talk about "the human's utility function" less., published by Charlie Steiner on December 8, 2022 on LessWrong.
    As a writing exercise, I'm writing an AI Alignment Hot Take Advent Calendar - one new hot take, written every day (more or less - I'm getting back on track!) for 25 days. Or until I run out of hot takes.
    When considering AI alignment, you might be tempted to talk about "the human's utility function," or "the correct utility function." Resist the temptation when at all practical. That abstraction is junk food for alignment research.
    As you may already know, humans are made of atoms. Collections of atoms don't have utility functions glued to them a priori - instead, we assign preferences to humans (including ourselves!) when we model the world, because it's a convenient abstraction. But because there are multiple ways to model the world, there are multiple ways to assign these preferences; there's no "the correct utility function."
    Maybe you understand all that, and still talk about an AI "learning the human's utility function" sometimes. I get it. It makes things way easier to assume there's some correct utility function when analyzing the human-AI system. Maybe you're writing about inner alignment and want to show that some learning procedure is flawed because it wouldn't learn the correct utility function even if humans had one. Or that some learning procedure would learn that correct utility function. It might seem like this utility function thing is a handy simplifying assumption, and once you have the core argument you can generalize it to the real world with a little more work.
    That seeming is false. You have likely just shot yourself in the foot.
    Because the human-atoms don't have a utility function glued to them, building aligned AI has to do something that's actually materially different than learning "the human's utility function." Something that's more like learning a trustworthy process. If you're not tracking the difference and you're using "the human's utility function" as a target of convenience, you can all too easily end up with AI designs that aren't trying to solve the problems we're actually faced with in reality - instead they're navigating their own strange, quasi-moral-realist problems.
    Another way of framing that last thought might be that wrapper-minds are atypical. They're not something that you actually get in reality when trying to learn human values from observations in a sensible way, and they have alignment difficulties that are idiosyncratic to them (though I don't endorse the extent to which nostalgebraist takes this).
    What to do instead? When you want to talk about getting human values into an AI, try to contextualize discussion of the human values with the process the AI is using to infer them. Take the AI's perspective, maybe - it has a hard and interesting job trying to model the world in all its complexity, if you don't short-circuit that job by insisting that actually it should just be trying to learn one thing (that doesn't exist). Take the humans' perspective, maybe - what options do they have to communicate what they want to the AI, and how can they gain trust in the AI's process?
    Of course, maybe you'll try to consider the AI's value-inference process, and find that its details make no difference whatsoever to the point you were trying to make. But in that case, the abstraction of "the human's utility function" probably wasn't doing any work anyhow. Either way.
    Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.
    4 min
  • EA - You should consider launching an AI startup by Joshc
    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: You should consider launching an AI startup, published by Joshc on December 8, 2022 on The Effective Altruism Forum.
    Edit: I've noticed I've been getting a lot of downvotes. If you want to reduce the probability that people go this route (myself included), it would be helpful if you voiced your disagreements. Also, opinions are my own.
    The prevailing view in this community is that we are talent constrained, not resource constrained. Earn to give has gone out of fashion. Now, community builders mostly push people towards AI research or strategy.
    Maybe this is a mistake. Or if it made sense previously, perhaps we should update in light of recent events.
    We could use a lot more money
    AI Safety researchers will need a lot of money for compute later...
    Given how expensive AI research is becoming, research orgs like Redwood research and The Center for AI Safety will need a lot of funding in the future. Where is that money going to come from? How are they going to fine-tune the latest models when OpenAI is spending billions? If we want there to be a favorable safety-capabilities balance, there needs to be a favorable safety-capabilities funding balance!
    Money can be converted to talent
    Currently, the primary recruitment strategy appears to be to persuade people to take X-risk arguments seriously. But not many people change their careers because of abstract arguments. Influencing incentives seems more promising to me; first, for directly moving talent, but also for building safety culture. The more appealing it is to work on AI Safety, the more willing people will be to engage with arguments. Research incentives can be shifted in a variety of ways using money:
    Fund AI Safety Industry labs. If there are more well-paying AI safety jobs, people will seek to fill them.
    Build a massive safety compute cluster. Academics feel locked out. They can barely fine-tune GPT-2 sized models while industry models crush every NLP benchmark. Offering compute for safety research is a golden opportunity to shift incentives. Academia is a wellspring of ML talent. If safety dominates there, it will trickle into industry.
    Competitions. The jury is still out on whether you can use ungodly amounts of money to motivate talent; but the fact that you can always raise the bounties make competitions scalable.
    Grants. A common way to grow research fields is to simply fund the research you would like to see more of (i.e. run RfPs). Apparently, the standard interventions tend to work. Accountability mechanisms have to be solid, but government orgs like DARPA appear to be good at this.
    Fellowships. Many professors could take on more graduate students if they had funding for them. For the small cost of $100,000 per year (one-third that in Canada), you can essentially 'hire' a graduate student to produce AI safety research. Graduate students are often more capable than industry research engineers and even come with free mentorship (a professor!). In addition to building new safety labs, we can start converting existing ones and giving them industry-level resources.
    'Talent' might not even be an important input in the future (relative to resources)
    The techniques in Holden's "How might we align transformative AI if it’s developed very soon?" are generally more resource-dependent than talent dependent. This is consistent with a general trend in ML research towards engineering and datasets and away from small numbers of people having creative insights.
    The most effective ways to select against deception could involve extensive red-teaming (e.g. 'honey pots'), or carefully constructing datasets / environments that don't produce instrumental self-preservation tendencies early on in the training process. The most effective ways to guard against distribution shift could be to test the AIs in a wide variety of scenario...
    7 min
  • EA - Thoughts on AGI organizations and capabilities work by RobBensinger
    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: Thoughts on AGI organizations and capabilities work, published by RobBensinger on December 7, 2022 on The Effective Altruism Forum.
    (Note: This essay was largely written by Rob, based on notes from Nate. It’s formatted as Rob-paraphrasing-Nate because (a) Nate didn’t have time to rephrase everything into his own words, and (b) most of the impetus for this post came from Eliezer wanting MIRI to praise a recent OpenAI post and Rob wanting to share more MIRI-thoughts about the space of AGI organizations, so it felt a bit less like a Nate-post than usual.)
    Nate and I have been happy about the AGI conversation seeming more honest and “real” recently. To contribute to that, I’ve collected some general Nate-thoughts in this post, even though they’re relatively informal and disorganized.
    AGI development is a critically important topic, and the world should obviously be able to hash out such topics in conversation. (Even though it can feel weird or intimidating, and even though there’s inevitably some social weirdness in sometimes saying negative things about people you like and sometimes collaborate with.) My hope is that we'll be able to make faster and better progress if we move the conversational norms further toward candor and substantive discussion of disagreements, as opposed to saying everything behind a veil of collegial obscurity.
    Capabilities work is currently a bad idea
    Nate’s top-level view is that ideally, Earth should take a break on doing work that might move us closer to AGI, until we understand alignment better.
    That move isn’t available to us, but individual researchers and organizations who choose not to burn the timeline are helping the world, even if other researchers and orgs don't reciprocate. You can unilaterally lengthen timelines, and give humanity more chances of success, by choosing not to personally shorten them.
    Nate thinks capabilities work is currently a bad idea for a few reasons:
    He doesn’t buy that current capabilities work is a likely path to ultimately solving alignment.
    Insofar as current capabilities work does seem helpful for alignment, it strikes him as helping with parallelizable research goals, whereas our bottleneck is serial research goals. (See A note about differential technological development.)
    Nate doesn’t buy that we need more capabilities progress before we can start finding a better path.
    This is not to say that capabilities work is never useful for alignment, or that alignment progress is never bottlenecked on capabilities progress. As an extreme example, having a working AGI on hand tomorrow would indeed make it easier to run experiments that teach us things about alignment! But in a world where we build AGI tomorrow, we're dead, because we won't have time to get a firm understanding of alignment before AGI technology proliferates and someone accidentally destroys the world. Capabilities progress can be useful in various ways, while still being harmful on net.
    (Also, to be clear: AGI capabilities are obviously an essential part of humanity's long-term path to good outcomes, and it's important to develop them at some point — the sooner the better, once we're confident this will have good outcomes — and it would be catastrophically bad to delay realizing them forever.)
    On Nate’s view, the field should do experiments with ML systems, not just abstract theory. But if he were magically in charge of the world's collective ML efforts, he would put a pause on further capabilities work until we've had more time to orient to the problem, consider the option space, and think our way to some sort of plan-that-will-actually-probably-work. It’s not as though we’re hurting for ML systems to study today, and our understanding already lags far behind today’s systems' capabilities.
    Publishing capabilities advances is even more obviously b...
    12 min
  • EA - Learning from non-EAs who seek to do good by Siobhan M
    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: Learning from non-EAs who seek to do good, published by Siobhan M on December 8, 2022 on The Effective Altruism Forum.
    Is EA a question, or a community based around ideology?
    After a year of close interaction with Effective Altruism – and recognizing that the movement is made up of many people with different views – I’m still confused as to whether EA aims to be a question about doing good effectively, or a community based around ideology.
    In my experience, it’s largely been the latter, but many EAs have expressed – either explicitly or implicitly – that they’d like it to be the former. I see this in the frequent citations of “EA is a question (not an ideology)” and the idea of the scout mindset; and most recently, in a lot of the top comments in the post on suggestions for community changes.
    As an EA-adjacent individual, I think the single most important thing the EA community could do to become more of a question, rather than an ideology, is to take concrete steps to interact more with, learn from, and collaborate with people outside of EA who seek to do good, without necessarily aiming to bring them into the community.
    I was a Fellow with Vox’s Future Perfect section last year, and moved to the Bay Area in part to learn more about EA. I want to thank the EA community for letting me spend time in your spaces and learn from your ideas; my view of the world has definitely broadened over the past year, and I hope to continue to be engaged with the community.
    But EA has never been, and I don’t see it ever becoming, my primary community. The EA community and I have differences in interests, culture, and communication styles, and that’s okay on both ends. As this comment says, the core EA community is not a good fit for everyone!
    A bit more about me. After college, I worked with IDinsight, a global development data analysis and advisory firm that has collaborated with GiveWell. Then I wrote for Future Perfect, focusing on global development, agriculture, and climate. I care a lot about lowercase effective altruism – trying to make the world better in an effective way – and evidence-based decision making. Some specific ways in which I differ from the average highly-engaged EA (although my, or their, priors could change) is that I’m more sympathetic to non-utilitarian ethical theories (and religion), more sympathetic to person-affecting views, and more skeptical that we can predict how our actions now will impact the long-term future.
    My personal experience with the EA community has largely been that it’s a community based around an ideology, rather than a question. I probably disagree with both some critics and some EAs in that I don’t think being a community based around ideology is necessarily bad. I come from a religious background, and while I have a complicated relationship with religion, I have a lot of close friends and family members who are religious, and I have a lot of respect for ideology-based communities in many ways. It’s helpful for me to know what their ideology is, as I know going into discussions where we’ll likely differ and where we’ll potentially find common ground.
    If EA aims to be a community based around ideology, I don’t think much has to change; and the only request I’d have is that EA leadership and general discourse more explicitly position the community this way. It’s frustrating and confusing to interact with a community that publicly claims the importance of moral uncertainty, and then have your ideas dismissed when you’re not a total utilitarian or strong longtermist.
    That said, a lot of EAs have expressed that they do not want to be a community based around ideology. I appreciated the post about EA disillusionment and agree with some of the recent critical posts around the community for women, but this post is not about the community itse...
    12 min
  • AF - Thoughts on AGI organizations and capabilities work by Rob Bensinger
    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: Thoughts on AGI organizations and capabilities work, published by Rob Bensinger on December 7, 2022 on The AI Alignment Forum.
    (Note: This essay was largely written by Rob, based on notes from Nate. It’s formatted as Rob-paraphrasing-Nate because (a) Nate didn’t have time to rephrase everything into his own words, and (b) most of the impetus for this post came from Eliezer wanting MIRI to praise a recent OpenAI post and Rob wanting to share more MIRI-thoughts about the space of AGI organizations, so it felt a bit less like a Nate-post than usual.)
    Nate and I have been happy about the AGI conversation seeming more honest and “real” recently. To contribute to that, I’ve collected some general Nate-thoughts in this post, even though they’re relatively informal and disorganized.
    AGI development is a critically important topic, and the world should obviously be able to hash out such topics in conversation. (Even though it can feel weird or intimidating, and even though there’s inevitably some social weirdness in sometimes saying negative things about people you like and sometimes collaborate with.) My hope is that we'll be able to make faster and better progress if we move the conversational norms further toward candor and substantive discussion of disagreements, as opposed to saying everything behind a veil of collegial obscurity.
    Capabilities work is currently a bad idea
    Nate’s top-level view is that ideally, Earth should take a break on doing work that might move us closer to AGI, until we understand alignment better.
    That move isn’t available to us, but individual researchers and organizations who choose not to burn the timeline are helping the world, even if other researchers and orgs don't reciprocate. You can unilaterally lengthen timelines, and give humanity more chances of success, by choosing not to personally shorten them.
    Nate thinks capabilities work is currently a bad idea for a few reasons:
    He doesn’t buy that current capabilities work is a likely path to ultimately solving alignment.
    Insofar as current capabilities work does seem helpful for alignment, it strikes him as helping with parallelizable research goals, whereas our bottleneck is serial research goals. (See A note about differential technological development.)
    Nate doesn’t buy that we need more capabilities progress before we can start finding a better path.
    This is not to say that capabilities work is never useful for alignment, or that alignment progress is never bottlenecked on capabilities progress. As an extreme example, having a working AGI on hand tomorrow would indeed make it easier to run experiments that teach us things about alignment! But in a world where we build AGI tomorrow, we're dead, because we won't have time to get a firm understanding of alignment before AGI technology proliferates and someone accidentally destroys the world. Capabilities progress can be useful in various ways, while still being harmful on net.
    (Also, to be clear: AGI capabilities are obviously an essential part of humanity's long-term path to good outcomes, and it's important to develop them at some point — the sooner the better, once we're confident this will have good outcomes — and it would be catastrophically bad to delay realizing them forever.)
    On Nate’s view, the field should do experiments with ML systems, not just abstract theory. But if he were magically in charge of the world's collective ML efforts, he would put a pause on further capabilities work until we've had more time to orient to the problem, consider the option space, and think our way to some sort of plan-that-will-actually-probably-work. It’s not as though we’re hurting for ML systems to study today, and our understanding already lags far behind today’s systems' capabilities.
    Publishing capabilities advances is even more obviously bad
    Fo...
    12 min
  • EA - The Spanish-Speaking Effective Altruism community is awesome by Jaime Sevilla
    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 Spanish-Speaking Effective Altruism community is awesome, published by Jaime Sevilla on December 7, 2022 on The Effective Altruism Forum.
    Epistemic status: fanboying and biased. It also focuses a lot on the part of the Spanish-Speaking community that I am familiar with, but there are many other aspects I didn’t cover.
    TL;DR: Exactly what it says in the title. Come meet us at EAGx LatAm! And if you speak Spanish, join the virtual community now!
    I can’t overstate how impressed I am with the successes of the Spanish-speaking EA community in the last year.
    The community was in a somewhat sorry state. Several fragmented efforts in Spain, Argentina and elsewhere had been tried. Some great people joined us that way (e.g., the equally feared and loved Nuño Sempere from QURI, the incredible entrepreneur Pablo Melchor of Ayuda Efectiva or the wondrous researcher Juan García from ALLFED). And yet, without anyone dedicated to nurturing it, the community languished.
    Several things were tried. An online group of Spanish-Speaking community builders met periodically to discuss the situation. EA Argentina and EA Spain joined their slack workspaces. Nothing worked, not for long.
    In the summer of 2021, at the pandemic's peak, we took a turn for the better. Sandra Malagón and Laura González sought, with the support of the community, a grant to dedicate themselves professionally to growing the Spanish-speaking community.
    Since then:
    They have run ~40 cohorts of the online introductory fellowship to Effective Altruism in Spanish, reaching ~150 people. This led, among other things, to the establishment of two university groups in Bogotá and Santiago, thanks in no small part to the initiative of Alejandro Acelas, Laura Arana, Agustín Covarrubias and David Solar.
    They have run two camps in Mexico and Colombia, widely promoted in Universities and reaching ~80 people. This led to about ~10 people becoming highly engaged community members.
    They have supported the ongoing community efforts in Colombia, Mexico and Chile, which each now have dedicated community leaders: Jaime Andres Fernández in Colombia; Claudette Salinas and Michelle Bruno in México; the whole Universidad Católica de Chile group in, well, Chile.
    Laura González led a translation project to get the EA Handbook and other key materials translated into Spanish, and supported the pitching and translation of EA books. More broadly, she contributes to a larger effort to streamline the translation of EA materials to non-English languages.
    Ángela Aristizábal, who had been doing community building in Colombia, organised a fellows program to nurture young professionals. These fellows have gone on to participate in the United Nations, lead local EA communities and do other amazing projects.
    The activity in Slack has grown sevenfold since 2020, and it’s on an upward trend.
    Sandra Malagón ideated and is running a community fellowship program in Mexico City, establishing the city as a budding hub.
    This is just the beginning. EAGx in Mexico City is nearing. The coordination team has more plans to make the community more professional and foster more impactful projects. And the amount of Spanish EA memes is going through the roof.
    Here is a highlight of some of my favourite people in the community:
    Our fearless captain Sandra Malagón, who within one year has kickstarted an EA hub in Mexico and helped raise a community in Chile and Colombia.
    Laura González, who co-coordinates the Spanish speaking community and leads the Spanish translation project. An awesome master of ceremonies for solemn occasions of every sort, too.
    Pablo Melchor, founder of Ayuda Efectiva, promoting and enabling effective giving in Spanish-Speaking countries.
    Javier Prieto, AI Governance and Policy Program Assistant at Open Philanthropy. Mug not included.
    The elusive Pablo...
    7 min
  • EA - Promoting compassionate longtermism by jonleighton
    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: Promoting compassionate longtermism, published by jonleighton on December 7, 2022 on The Effective Altruism Forum.
    This post is in 6 parts, starting with some basic reflections on suffering and ethics, and ending with a brief project description. While this post might seem overly broad-ranging, it’s meant to set out some basic arguments and explain the rationale for the project initiative in the last section, for which we are looking for support and collaboration. I go into much greater detail about some of the core ethical ideas in a new book about to be published, which I will present soon in a separate post. I also make several references here to Will MacAskill’s What We Owe the Future, because many of the ideas he expresses are shared by many EAs, and while I agree with many of the things he says, there are some important stances I disagree with that I will explain in this post.
    My overall motivation is a deep concern about the persistence of extreme suffering far into the future, and the possibility to take productive steps now to reduce the likelihood of that happening, thereby increasing the likelihood that the future will be a flourishing one.
    Summary:
    Suffering has an inherent call to action, and some suffering literally makes non-existence preferable.
    For various reasons, there are mixed attitudes within EA towards addressing suffering as a priority.
    We may not have the time to delay value lock-in for too long, and we already know some of the key principles.
    Increasing our efforts to prevent intense suffering in the short term may be important for preventing the lock-in of uncompassionate values.
    There’s an urgent need to research and promote mechanisms that can stabilise compassionate governance at the global level.
    OPIS is initiating research and film projects to widely communicate these ideas and concrete steps that can already be taken, and we are looking for support and collaboration.
    1. Some reflections on suffering
    Involuntary suffering is inherently bad – one could argue that this is ultimately what “bad” means – but extreme, unbearable suffering is especially bad, to the point that non-existence is literally a preferable option. At this level, people choose to end their lives if they can in order to escape the pain.
    We probably cannot fully grasp what it’s like to experience extreme suffering unless we have experienced it ourselves. To get even an approximate sense of what it’s like requires engaging with accounts and depictions of it. If not, we may underestimate its significance and attribute much lower priority to it than it deserves. As an example, a patient with a terrible condition called SUNCT whom I provided support to, who at one point attempted suicide, described in a presentation we recently gave together in Geneva the utter hell he experienced, and how no one should ever have to experience what he did.
    Intense suffering has an inherent call to action – we respond to it whenever we try to help people in severe pain, or animals being tortured on factory farms.
    There is no equivalent inherent urgency to fill the void and bring new sentient beings into existence, even though this is an understandable desire of intelligent beings who already exist.
    Intentionally bringing into existence a sentient being who will definitely experience extreme/unbearable suffering could be considered uncompassionate and even cruel.
    I don’t think the above reflections should be particularly controversial. Even someone who would like to fill the universe with blissful beings might still concede that the project doesn’t have an inherent urgency – that is, that it could be delayed for some time, or even indefinitely, without harm to anyone (unless you believe, as do some EAs, that every instance of inanimate matter in space and time that isn’t being optimally used ...
    20 min
  • EA - What can we learn from the empirical social science literature on the expected contingency of value change? by jackva
    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 can we learn from the empirical social science literature on the expected contingency of value change?, published by jackva on December 7, 2022 on The Effective Altruism Forum.
    Note: Quickly written on vacation as I don't and won't have time to flesh this out much more but wanted to get the idea out so that others can work on this insofar as this seems important. I usually write on climate and the last time I seriously thought about value change is a while back.
    A missing perspective?
    Reading What We Owe The Future's chapter on the contingency of value change one thing that struck me was that there was -- as far as I can tell -- no reference to the rich empirical literature in political science and sociology on the drivers of value change, e.g. the work on the emergence of post-materialist values by Inglehart and the work of other scholars in this tradition, including the World Value Survey, which maps and seeks to explain value changes in societies around the world.
    Note that I am not claiming this literature being entirely right, more that this is a large body of literature (including criticisms) that seems very relevant to EA interest in value change, but not discussed.
    Introductions to this literature
    Ezra Klein podcast 2022: Conversation with Pippa Norris
    Welzel 2021: Why the Future is Democratic
    Inglehart and Welzel 2010: Changing Mass Priorities: The Link between Modernization and Democracy
    Inglehart and Welzel 2009: How Development Leads to Democracy
    Welzel et al 2003: The theory of human development, a cross-cultural analysis.
    These sources are quickly chosen and are biased towards Welzel (whom I was a student of in college), so don't take this as a definite treatment - just a window into this literature.The basic point of this literature is that there are good reasons to expect that value change is actually quite predictable and that a certain set of values tend to emerge out of the conditions of modernization. There is a lot more nuance to it, this is not just "old-school" modernization theory, but it is an empirically grounded update against massive contingency in the development of values, as there are predictable relationships between economic development and value change that seem to hold across a broad set of geographies and stages of economic development.To give a flavor, the abstract of Welzel 2021 puts it as follows:"Recent accounts of democratic backsliding neglect the cultural foundations of autocracy-versus-democracy.
    To bring culture back in, this article demonstrates that 1) countries' membership in culture zones explains some 70 percent of the total cross-national variation in autocracy-versus-democracy; and 2) this culture-bound variation has remained astoundingly constant over time—in spite of all the trending patterns in the global distribution of regime types over the last 120 years. Furthermore, the explanatory power of culture zones over autocracy-versus-democracy is rooted in the cultures' differentiation on "authoritarian-versus-emancipative values." Therefore, both the direction and the extent of regime change are a function of glacially accruing regime-culture misfits—driven by generational value shifts in a predominantly emancipatory direction. Consequently, the backsliding of democracies into authoritarianism is limited to societies in which emancipative values remain underdeveloped.
    Contrary to the widely cited deconsolidation thesis, the ascendant generational profile of emancipative values means that the momentary challenges to democracy are unlikely to stifle democracy's long-term rise."
    How it matters: different priors on the contingency of value change
    I am wondering
    (1) whether this literature could be quite useful in forming priors about the contingency of value change and,
    (2) in particular, serve as useful correctiv...
    8 min
  • EA - Why development aid is a really exciting field by MathiasKB
    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 development aid is a really exciting field, published by MathiasKB on December 7, 2022 on The Effective Altruism Forum.
    Each year, wealthy countries collectively spend around 178 billion dollars (!!) on Development aid.
    Development aid has funded some of the most cost-effective lifesaving programs that have ever been run. Such examples include PEPFAR, the US emergency aids relief programme rolled out at the height of the African aids pandemic, which estimates suggest saved 25 million lives at a cost of some 85 billion ($3400 per life saved, competitive with Givewell’s very best). EAs working with global poverty will know just how difficult it is to achieve high cost effectiveness at these scales.
    Development aid has also funded some of the very worst development projects conceived, in some instances causing outright harm to the recipients.
    Development aid is spent with a large variety of goals in mind. Climate mitigation projects, gender equality campaigns, and free-trade agreements are all funded by wealthy governments under a single illusory budgetary item: ‘development assistance’.
    In short, the scope of aid is enormous and so is the impact that can be had by positively influencing how it is spent. I'm not the only one who thinks so! In January of 2022 Open Philanthropy announced Global Aid Policy as a priority area within its global health and wellbeing portfolio.
    In this post I will:
    Demystify the processes that decide how aid is allocated.
    Argue that aid policy is neglected (by EAs especially), high in scale, and maybe tractable.
    Sneakily attempt to make you excited about aid, in preparation for the announcement of a non-profit I’m co-founding.
    Who decides how aid is allocated?
    When I first dug into development aid, I found the field very opaque and difficult to get an overview of. This made everything seem much more static and difficult to influence, than I now think it is.
    Come with me, and I’ll show you how the aid-sausage is made. The explanation tries to capture the grand picture, but each country is different and the explanation is overfit to western democracies.
    The aid pipeline:
    It all begins with a government decision to spend money on aid. For many countries this decision was formalized in 1970 after a UN resolution was signed between members to spend 0.7% of GNI on official development assistance.
    Politicians decide on a national aid strategy
    Each country that gives aid will have an official strategy for its aid spending. The strategy lists a number of priorities the government wants to focus on. It is typically re-negotiated and updated once every few years or when a new government takes seat. The agreed upon aid strategy sets a broad direction for the civil service and relevant parliamentary committees in projects they choose to carry out.
    The recently released UK aid strategy is a good example of what typical priorities look like:
    deliver honest, reliable investment through British Investment Partnerships, building on the UK’s financial expertise and the strengths of the City of London, in line with the Prime Minister’s vision for the Clean Green Initiative
    provide women and girls with the freedom they need to succeed, unlocking their future potential, educating girls, supporting their empowerment and protecting them against violence
    step up the UK’s life-saving humanitarian work to prevent the worst forms of human suffering around the world. The UK will lead globally for a more effective international response to humanitarian crises
    take forward UK leadership on climate change, nature and global health. The strategy will put the UK commitments made during the UK’s Presidency of G7 and COP26, UK global leadership in science and technology, and the UK’s COVID-19 response, at the core of its international development work
    The Government passes a...
    11 min
  • EA - EA London Rebranding to EA UK by DavidNash
    Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: EA London Rebranding to EA UK, published by DavidNash on December 7, 2022 on The Effective Altruism Forum.
    This post is mainly from the original proposal document to rebrand EA London to EA UK. There is now an EA UK website and newsletter since August 2022.
    Summary
    There are people in the UK outside of London, Oxford and Cambridge who have an interest in EA and don’t have a local group. They may miss out on chances to connect with others, stay motivated and find out about job and donation opportunities. It could be useful to set up EA UK to support this wider group of people.
    Why is there no EA UK already?
    Historically local EA groups have been founded where there happened to be someone with enough time to run one rather than with a larger strategy in mind. Newer organisers around the world tend to create national groups that support local groups within that country and provide a support network for everyone in their country rather than a few cities. Because EA groups in the UK were made pretty early they grew faster before the idea of national groups was as widespread.
    Another contributing factor to why there is no EA UK may be that some of the support that EA country groups often provide is given by CEA and other organisations, for example most country groups have to set up a donation platform for people who want to give effectively, but this is already done in the UK by EA Funds. There is also no need to translate materials into a local language for the UK.
    Why set up EA UK?
    The UK has a population of 67 million, with roughly 14% living in London, Oxford and Cambridge, the three places in the UK with paid group organisers. This means that people in the rest of the country who have an interest in EA will not have a group they can get support from
    A lot of the support that can be provided doesn’t have to be in person (1-1s, newsletter, directory, project advice)
    People move around the UK and so if there is a national group they can stay up to date with EA no matter whether they are moving to, it also could reduce the chance that people drop out of EA when they move away from or between hubs
    There will be some individuals who wont be interested in attending local group events but will want to keep up to date with what happens with EA in the UK and occasionally want to talk about donations/careers
    Even though EA London can provide support to those outside of London, people may not reach out to a group that isn’t near them unless they feel like they have permission to
    By having a national group it may lead to more local groups being set up as it is easier for people in the same place to find each other
    What EA UK wont do
    Provide support to UK student groups, this is already done by CEA, Open Philanthropy and the Global Challenges Project. EA UK could act as an extra place for them to go for support, but generally would be more focused on professionals in the UK
    Act as a layer of management between other local groups in the UK and CEA, they will still interact with CEA for support
    What does this mean for EA London?
    Most things will stay the same, the website and newsletter have been rebranded to EA UK rather than EA London
    There will still be events and support for people in London
    Potential Issues
    EA London, Oxford and Cambridge community members may get less support if there is more focus on a national group
    People outside of UK EA hubs may feel neglected if most of the jobs/events etc are based in hubs. Although hopefully providing more support will allow more relevant opportunities in the wider UK to be discovered and then this can be shared back to the UK community
    There may be much more demand for the support that is offered. As this is a problem of being potentially too successful we could scale back different projects, or maybe consider hiring extra people if it...
    4 min