The Nonlinear Library

The Nonlinear Library

By The Nonlinear FundEducation
Download on the App Store

All The Nonlinear Library episodes

  • EA - Getting Cats Vegan is Possible and Imperative by Karthik Sekar
    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: Getting Cats Vegan is Possible and Imperative, published by Karthik Sekar on May 4, 2023 on The Effective Altruism Forum.
    Summary
    Carnivore is a classification, not a diet requirement.
    The amount of meat that cats eat is significant. Transitioning domestic cats to eating vegan would do much good for the environment and animal welfare.
    Having vegan cats now is not convenient, but we (humanity) should make that so.
    We do not need to wait around for cultivated meat. There are tractable opportunities now.
    We also need randomized control trials with measured health outcomes; funding is the main limitation here.
    Making domestic cats vegan meets all of the Effective Altruism criteria: significant, tractable, and neglected.
    Main
    Imagine you are a surveyor traveling to remote parts of the world. Within a thick rainforest, you come across an indigenous group long separated from the modern world. They fashion spears to hunt fish and thicket baskets to collect foraged berries. Notably, they wear distinctive yellow loincloths dyed with local fruit. You are not one with words, so you call them the Yellowclothea.
    This is not a farfetched story. Most species worldwide are classified similarly–someone observes them and then contrives a classification named on what they see. Carnivora was coined in 1821 to describe an Order of animals by the observation that they consumed the meat of other animals–carnem vorāre is Latin for “to eat flesh”.
    Let us go back to the Yellowclothea. You can already intuit that these natives do not have to wear the yellow loincloths–it is simply what you initially observed. If the natives swapped the dye with purple or green, that would work out fine. However, the rainforest lacks those colors, so Yellowclothea is resigned to their monotone. In other words, wearing yellow cloth is not a requirement for them to live, just what works for them and is available.
    1821, the year of Carnivora’s naming, is ages ago in the scientific world. It was before the Theory of Evolution, first described in The Origin of Species in 1859. It was before the molecular biology revolution. It was before we understood the basis of metabolism and nutrition. So it is easy to confuse classification/observation with the requirement. It is the same fallacy as assuming that the Yellowclothea people can only wear yellow clothing.
    Since 1821, we learned more about nutrition, molecular biology, and metabolism to demystify meat. Meat is mostly muscle fibers with some marbled fat and critical nutrients. Carnivora animals generally have more acidic stomachs and shorter gastrointestinal (GI) tracts than nominal herbivores. The extra acid helps chop proteins into the alphabet amino acid molecules, which are readily taken up, so a long GI tract is unnecessary.
    So Carnivora animals cannot have salads or raw vegetables, which are rich in fiber and would not break down in their GI tracts in time. Nevertheless, we can make protein-rich and highly digestible foods for Carnivora starting from plant and microbial ingredients. Just as a cow will chemically process the plants into their muscle–flesh, we can similarly turn the plants into food that a carnivore would thrive off without an animal intermediary. In other words, we can source all the required nutrients from elsewhere, without meat.
    There is—at least in theory—no reason why diets comprised entirely of plants, minerals, and synthetically-based ingredients (i.e., vegan diets) cannot meet the necessary palatability, bioavailability, and nutritional requirements of cats
    Andrew Knight, Director, Centre for Animal Welfare, University of Winchester
    I have written about how succeeding meat, dairy, and eggs with plant and microbial-based alternatives will be one of the best things we ever do–I argue that it is better than curing cancer or transitioning ful...
    14 min
  • EA - [Link Post: New York Times] White House Unveils Initiatives to Reduce Risks of A.I. by Rockwell
    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: [Link Post: New York Times] White House Unveils Initiatives to Reduce Risks of A.I., published by Rockwell on May 4, 2023 on The Effective Altruism Forum.
    This is a linkpost for
    The White House on Thursday announced its first new initiatives aimed at taming the risks of artificial intelligence since a boom in A.I.-powered chatbots has prompted growing calls to regulate the technology.
    The National Science Foundation plans to spend $140 million on new research centers devoted to A.I., White House officials said. The administration also pledged to release draft guidelines for government agencies to ensure that their use of A.I. safeguards “the American people’s rights and safety,” adding that several A.I. companies had agreed to make their products available for scrutiny in August at a cybersecurity conference.
    The announcements came hours before Vice President Kamala Harris and other administration officials were scheduled to meet with the chief executives of Google, Microsoft, OpenAI, the maker of the popular ChatGPT chatbot, and Anthropic, an A.I. start-up, to discuss the technology. A senior administration official said on Wednesday that the White House planned to impress upon the companies that they had a responsibility to address the risks of new A.I. developments. The White House has been under growing pressure to police A.I. that is capable of crafting sophisticated prose and lifelike images. The explosion of interest in the technology began last year when OpenAI released ChatGPT to the public and people immediately began using it to search for information, do schoolwork and assist them with their job.
    Since then, some of the biggest tech companies have rushed to incorporate chatbots into their products and accelerated A.I. research, while venture capitalists have poured money into A.I. start-ups.
    But the A.I. boom has also raised questions about how the technology will transform economies, shake up geopolitics and bolster criminal activity. Critics have worried that many A.I. systems are opaque but extremely powerful, with the potential to make discriminatory decisions, replace people in their jobs, spread disinformation and perhaps even break the law on their own.
    President Biden recently said that it “remains to be seen” whether A.I. is dangerous, and some of his top appointees have pledged to intervene if the technology is used in a harmful way.
    Spokeswomen for Google and Microsoft declined to comment ahead of the White House meeting. A spokesman for Anthropic confirmed the company would be attending. A spokeswoman for OpenAI did not respond to a request for comment.
    The announcements build on earlier efforts by the administration to place guardrails on A.I. Last year, the White House released what it called a “Blueprint for an A.I. Bill of Rights,” which said that automated systems should protect users’ data privacy, shield them from discriminatory outcomes and make clear why certain actions were taken. In January, the Commerce Department also released a framework for reducing risk in A.I. development, which had been in the works for years.
    The introduction of chatbots like ChatGPT and Google’s Bard has put huge pressure on governments to act. The European Union, which had already been negotiating regulations to A.I., has faced new demands to regulate a broader swath of A.I., instead of just systems seen as inherently high risk.
    In the United States, members of Congress, including Senator Chuck Schumer of New York, the majority leader, have moved to draft or propose legislation to regulate A.I. But concrete steps to rein in the technology in the country may be more likely to come first from law enforcement agencies in Washington.
    A group of government agencies pledged in April to “monitor the development and use of automated systems and promote responsible...
    5 min
  • LW - Clarifying and predicting AGI by Richard Ngo
    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: Clarifying and predicting AGI, published by Richard Ngo on May 4, 2023 on LessWrong.
    This post is a slightly-adapted summary of two twitter threads, here and here.
    The t-AGI framework
    As we get closer to AGI, it becomes less appropriate to treat it as a binary threshold. Instead, I prefer to treat it as a continuous spectrum defined by comparison to time-limited humans. I call a system a t-AGI if, on most cognitive tasks, it beats most human experts who are given time t to perform the task.
    What does that mean in practice?
    A 1-second AGI would need to beat humans at tasks like quickly answering trivia questions, basic intuitions about physics (e.g. "what happens if I push a string?"), recognizing objects in images, recognizing whether sentences are grammatical, etc.
    A 1-minute AGI would need to beat humans at tasks like answering questions about short text passages or videos, common-sense reasoning (e.g. Yann LeCun's gears problems), simple computer tasks (e.g. use photoshop to blur an image), justifying an opinion, looking up facts, etc.
    A 1-hour AGI would need to beat humans at tasks like doing problem sets/exams, writing short articles or blog posts, most tasks in white-collar jobs (e.g. diagnosing patients, giving legal opinions), doing therapy, doing online errands, learning rules of new games, etc.
    A 1-day AGI would need to beat humans at tasks like writing insightful essays, negotiating business deals, becoming proficient at playing new games or using new software, developing new apps, running scientific experiments, reviewing scientific papers, summarizing books, etc.
    A 1-month AGI would need to beat humans at coherently carrying out medium-term plans (e.g. founding a startup), supervising large projects, becoming proficient in new fields, writing large software applications (e.g. a new OS), making novel scientific discoveries, etc.
    A 1-year AGI would need to beat humans at... basically everything. Some projects take humans much longer (e.g. proving Fermat's last theorem) but they can almost always be decomposed into subtasks that don't require full global context (even tho that's often helpful for humans).
    Some clarifications:
    I'm abstracting away from the question of how much test-time compute AIs get (i.e. how many copies are run, for how long). A principled way to think about this is probably something like: "what fraction of the world's compute is needed?". But in most cases I expect that the bottleneck is being able to perform a task at all; if they can then they'll almost always be able to do it with a negligible proportion of the world's compute.
    Similarly, I doubt the specific "expert" theshold will make much difference. But it does seem important that we use experts not laypeople, because the amount of experience that laypeople have with most tasks is so small. It's not really well-defined to talk about beating "most humans" at coding or chess; and it's not particularly relevant either.
    I expect that, for any t, the first 100t-AGIs will be way better than any human on tasks which only take time t. To reason about superhuman performance we can extend this framework to talk about (t,n)-AGIs which beat any group of n humans working together on tasks for time t. When I think about superintelligence I'm typically thinking about (1 year, 8 billion)-AGIs.
    The value of this framework is ultimately an empirical matter. But it seems useful so far: I think existing systems are 1-second AGIs, are close to 1-minute AGIs, and are a couple of years off from 1-hour AGIs. (FWIW I formulated this framework 2 years ago, but never shared it widely. From your perspective there's selection bias—I wouldn't have shared it if I'd changed my mind. But at least from my perspective, it gets points for being useful for describing events since then.)
    And very briefly, some of th...
    8 min
  • EA - Introducing Animal Policy International by Rainer Kravets
    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: Introducing Animal Policy International, published by Rainer Kravets on May 4, 2023 on The Effective Altruism Forum.
    Animal Policy International is a new organisation launched through the Charity Entrepreneurship Incubation Program focused on ensuring that animal welfare standards are upheld in international trade policy.
    Problem
    There are significant differences between farmed animal welfare standards across the globe, with billions of animals still confined in factory farms. Even those regions with higher standards like the EU, the UK, Switzerland and New Zealand tend to import a significant portion of their animal products from countries where animals experience significant suffering due to lack of protective measures.
    Solution
    The higher welfare countries can apply their standards to imported animal products by restricting the access of low-welfare animal products that would have been illegal to produce domestically. This can incentivise farmers elsewhere to increase their standards to keep existing supply chains.
    A law restricting the importation of low-welfare products provides a unique win-win opportunity for both animal advocates and farmers in higher welfare countries, especially in our likely first country of operation: New Zealand. Some farmers are facing tough competition from low-priced low-welfare imports and demand more equal standards between imports and local produce after New Zealand’s decision to phase out farrowing crates on local pig farms by December 2025.
    Potential Impact
    A law passed in New Zealand restricting the importation of animal products that do not adhere to local standards could save approximately 8 million fish per year from suffering poor living conditions, transportation, and slaughter practices; spare 330,000 pigs from cruel farrowing crates and 380,000 chickens from inhumane living conditions.
    Differences in animal welfare standards: New Zealand
    Below is an outline of differences between animal welfare standards in New Zealand and its main importers of particular animals.
    Fish: In China, Vietnam and Thailand (total 79% of imports in 2020) there is no legislation for fish meaning they may endure slow, painful deaths by asphyxiation, crushing, or even being gutted alive. New Zealand outlines some protections for fish at the time of killing and during transport.
    Hens: 80% of eggs imported into New Zealand come from China where hens are allowed to be kept in battery cages. Battery cages are illegal in New Zealand from 2023. (Colony (enriched) cages are still used).
    Pigs: The US, an importer of pork to New Zealand, has no federal ban on the use of sow stalls or farrowing crates, leading to sows being cruelly confined to narrow cages where they cannot perform basic behaviours, turn around, or properly mother their piglets. New Zealand has banned sow stalls, and farrowing crates are being phased out by 2025.
    Sheep: Australia, which imports wool products to New Zealand, allows several practices that are prohibited in New Zealand, including the extremely cruel practice of mulesing, which involves removing parts of the skin from live sheep without anaesthetic.
    Next steps
    Establishing connections with potential partner NGOs and industry
    Producing a policy brief
    Conducting public polling
    Addressing the question of legality of import restrictions
    Meeting policymakers
    Open questions
    Will farmers in low-welfare countries be motivated and capable of increasing their animal welfare standards?
    What enforcement mechanisms should be used?
    How would a restriction on importation affect the country’s relationships with its trade partners?
    What externalities (e.g. changes in animal product prices) would such a trade law have?
    How you can help
    Expertise: if you have experience/knowledge in international trade, policy work, WTO laws and can help answer th...
    6 min
  • EA - 500 Million, But Not A Single One More by jai
    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: 500 Million, But Not A Single One More, published by jai on May 4, 2023 on The Effective Altruism Forum.
    We will never know their names.
    The first victim could not have been recorded, for there was no written language to record it. They were someone’s daughter, or son, and someone’s friend, and they were loved by those around them. And they were in pain, covered in rashes, confused, scared, not knowing why this was happening to them or what they could do about it — victims of a mad, inhuman god. There was nothing to be done — humanity was not strong enough, not aware enough, not knowledgeable enough, to fight back against a monster that could not be seen.
    It was in Ancient Egypt, where it attacked slave and pharaoh alike. In Rome, it effortlessly decimated armies. It killed in Syria. It killed in Moscow. In India, five million dead. It killed a thousand Europeans every day in the 18th century. It killed more than fifty million Native Americans. From the Peloponnesian War to the Civil War, it slew more soldiers and civilians than any weapon, any soldier, any army. (Not that this stopped the most foolish and empty souls from attempting to harness the demon as a weapon against their enemies.)
    Cultures grew and faltered, and it remained. Empires rose and fell, and it thrived. Ideologies waxed and waned, but it did not care. Kill. Maim. Spread. An ancient, mad god, hidden from view, that could not be fought, could not be confronted, could not even be comprehended. Not the only one of its kind, but the most devastating.
    For a long time, there was no hope — only the bitter, hollow endurance of survivors.
    In China, in the 10th century, humanity began to fight back.
    It was observed that survivors of the mad god’s curse would never be touched again: They had taken a portion of that power into themselves, and were so protected from it. Not only that, but this power could be shared by consuming a remnant of the wounds. There was a price, for you could not take the god’s power without first defeating it — but a smaller battle, on humanity’s terms.
    By the 16th century, the technique spread to India, then across Asia, the Ottoman Empire and, in the 18th century, Europe. In 1796, a more powerful technique was discovered by Edward Jenner.
    An idea began to take hold: Perhaps the ancient god could be killed.
    A whisper became a voice; a voice became a call; a call became a battle cry, sweeping across villages, cities, nations. Humanity began to cooperate, spreading the protective power across the globe, dispatching masters of the craft to protect whole populations. People who had once been sworn enemies joined in a common cause for this one battle. Governments mandated that all citizens protect themselves, for giving the ancient enemy a single life would put millions in danger.
    And, inch by inch, humanity drove its enemy back. Fewer friends wept; fewer neighbors were crippled; fewer parents had to bury their children.
    At the dawn of the 20th century, for the first time, humanity banished the enemy from entire regions of the world. Humanity faltered many times in its efforts, but there were individuals who never gave up, who fought for the dream of a world where no child or loved one would ever fear the demon ever again. Viktor Zhdanov, who called for humanity to unite in a final push against the demon; the great tactician Karel Raška, who conceived of a strategy to annihilate the enemy; Donald Henderson, who led the efforts in those final days.
    The enemy grew weaker. Millions became thousands, thousands became dozens. And then, when the enemy did strike, scores of humans came forth to defy it, protecting all those whom it might endanger.
    The enemy’s last attack in the wild was on Ali Maow Maalin, in 1977. For months afterwards, dedicated humans swept the surrounding area, seeking out an...
    5 min
  • EA - Upcoming EA conferences in 2023 by OllieBase
    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: Upcoming EA conferences in 2023, published by OllieBase on May 4, 2023 on The Effective Altruism Forum.
    The Centre for Effective Altruism will be organizing and supporting conferences for the EA community all over the world for the remainder of 2023, including the first-ever EA conferences in Poland, NYC and the Philippines.
    We currently have the following events scheduled:
    EA Global
    EA Global: London | (May 19–21) | Tobacco Dock - applications close 11:59 pm UTC Friday 5 May
    EA Global: Boston | (October 27–29) | Hynes Convention Center
    EAGx
    EAGxWarsaw | (June 9–11) | POLIN
    EAGxNYC | (August 18–20) | Convene, 225 Liberty St.
    EAGxBerlin | (September 8–10) | Urania
    EAGxAustralia | (September 22–24, provisional) | Melbourne
    EAGxPhilippines | (October 20–22, provisional)
    EAGxVirtual | (November 17–19, provisional)
    Applications for EAG London, EAG Boston, EAGxWarsaw and EAGxNYC are open, and we expect applications for the other conferences to open approximately 3 months before the event. Please go to the event page links above to apply. Please note again that applications to EAG London close 11:59 pm UTC Friday 5 May.
    If you'd like to add EA events like these directly to your Google Calendar, use this link.
    Some notes on these conferences:
    EA Globals are run in-house by the CEA events team, whereas EAGx conferences are organized independently by local community builders with financial support and mentoring from CEA.
    EA Global conferences have a high bar for admission and are for people who are very familiar with EA and are taking significant actions (e.g. full-time work or study) based on EA ideas.
    Admissions for EAGx conferences are processed independently by the EAGx conference organizers. These events are primarily for those who are newer to EA and interested in getting more involved and who are based in the region the conference is taking place in (e.g. EAGxWarsaw is primarily for people who are interested in EA and are based in Eastern Europe).
    Please apply to all conferences you wish to attend once applications open — we would rather get too many applications for some conferences and recommend that applicants attend a different one than miss out on potential applicants to a conference.
    Travel support funds for events this year are limited (though will vary by event), and we can only accommodate a small number of requests. If you do not end up receiving travel support, this is likely the result of limited funds, rather than an evaluation of your potential for impact. When planning around an event, we’d recommend you act under the assumption that we will not be able to grant your travel funding request (unless it has already been approved).
    Find more info on our website.
    Feel free to email [email protected] with any questions, or comment below. You can also contact EAGx organisers using the format [location]@eaglobalx.org (e.g. [email protected], [email protected]).
    Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.
    4 min
  • AF - Clarifying and predicting AGI by Richard Ngo
    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: Clarifying and predicting AGI, published by Richard Ngo on May 4, 2023 on The AI Alignment Forum.
    This post is a slightly-adapted summary of two twitter threads, here and here.
    The t-AGI framework
    As we get closer to AGI, it becomes less appropriate to treat it as a binary threshold. Instead, I prefer to treat it as a continuous spectrum defined by comparison to time-limited humans. I call a system a t-AGI if, on most cognitive tasks, it beats most human experts who are given time t to perform the task.
    What does that mean in practice?
    A 1-second AGI would need to beat humans at tasks like quickly answering trivia questions, basic intuitions about physics (e.g. "what happens if I push a string?"), recognizing objects in images, recognizing whether sentences are grammatical, etc.
    A 1-minute AGI would need to beat humans at tasks like answering questions about short text passages or videos, common-sense reasoning (e.g. Yann LeCun's gears problems), simple computer tasks (e.g. use photoshop to blur an image), justifying an opinion, looking up facts, etc.
    A 1-hour AGI would need to beat humans at tasks like doing problem sets/exams, writing short articles or blog posts, most tasks in white-collar jobs (e.g. diagnosing patients, giving legal opinions), doing therapy, doing online errands, learning rules of new games, etc.
    A 1-day AGI would need to beat humans at tasks like writing insightful essays, negotiating business deals, becoming proficient at playing new games or using new software, developing new apps, running scientific experiments, reviewing scientific papers, summarizing books, etc.
    A 1-month AGI would need to beat humans at coherently carrying out medium-term plans (e.g. founding a startup), supervising large projects, becoming proficient in new fields, writing large software applications (e.g. a new OS), making novel scientific discoveries, etc.
    A 1-year AGI would need to beat humans at... basically everything. Some projects take humans much longer (e.g. proving Fermat's last theorem) but they can almost always be decomposed into subtasks that don't require full global context (even tho that's often helpful for humans).
    Some clarifications:
    I'm abstracting away from the question of how much test-time compute AIs get (i.e. how many copies are run, for how long). A principled way to think about this is probably something like: "what fraction of the world's compute is needed?". But in most cases I expect that the bottleneck is being able to perform a task at all; if they can then they'll almost always be able to do it with a negligible proportion of the world's compute.
    Similarly, I doubt the specific "expert" theshold will make much difference. But it does seem important that we use experts not laypeople, because the amount of experience that laypeople have with most tasks is so small. It's not really well-defined to talk about beating "most humans" at coding or chess; and it's not particularly relevant either.
    I expect that, for any t, the first 100t-AGIs will be way better than any human on tasks which only take time t. To reason about superhuman performance we can extend this framework to talk about (t,n)-AGIs which beat any group of n humans working together on tasks for time t. When I think about superintelligence I'm typically thinking about (1 year, 8 billion)-AGIs.
    The value of this framework is ultimately an empirical matter. But it seems useful so far: I think existing systems are 1-second AGIs, are close to 1-minute AGIs, and are a couple of years off from 1-hour AGIs. (FWIW I formulated this framework 2 years ago, but never shared it widely. From your perspective there's selection bias—I wouldn't have shared it if I'd changed my mind. But at least from my perspective, it gets points for being useful for describing events since then.)
    And very briefl...
    8 min
  • EA - Air Safety to Combat Global Catastrophic Biorisks [REVISED] by Gavriel Kleinwaks
    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: Air Safety to Combat Global Catastrophic Biorisks [REVISED], published by Gavriel Kleinwaks on May 3, 2023 on The Effective Altruism Forum.
    This report is a collaboration between researchers from 1Day Sooner and Rethink Priorities.
    Overview
    This post is a revision of a report previously published on how improvements in indoor air quality can address global catastrophic risk from pandemics. After feedback from expert reviewers, we revised the report in accordance with comments. The comments greatly improved the report and we consider the earlier version to be misphrased, misleading, or mathematically underspecified in several places, but we are leaving the post available to illustrate the revision process.
    Unlike in the previous post, we are not including the full report, given its length. Instead, this post contains a summary of the reviews and of the report, with a link to the full report.
    Many thanks to the expert reviewers (listed below) for their detailed feedback. Additional thanks to Rachel Shu for research and writing assistance. We also received help and feedback from many other people over the course of this process—a full list is in the “Acknowledgements” section of the report.
    Summary of Expert Review
    We asked biosecurity and indoor air quality experts to review this report: Dr. Richard Bruns of the John Hopkins Center for Health Security, Dr. Jacob Bueno de Mesquita and Dr. Alexandra Johnson of Lawrence Berkeley National Lab, Dr. David Manheim of ALTER, and Professor Shelly Miller of the University of Colorado.
    These experts suggested a variety of both minor and substantive changes to the document, though these changes do not alter the overall conclusion of the report that indoor air safety is an important lever for reducing GCBRs and that there are several high-leverage funding opportunities around promoting indoor air quality and specific air cleaning interventions.
    The main changes suggested were:
    Providing confidence intervals on key estimates, such as our estimate of the overall impact of IAQ interventions, and reframing certain estimates to improve clarity.
    Modifying the phrasing around the section concerning ‘modelling’, to better clarify our position around the specific limitations of existing models (specifically that there aren’t models that move from the room and building-level transmission to population-level transmission).
    Clarifying the distinction between mechanical interventions, specific in-duct vs upper-room systems (254nm) and HVAC-filtration vs portable air cleaners and adding additional information about some interactions between different intervention types
    Adding general public advocacy for indoor air quality as a funding opportunity and related research that could be done support advocacy efforts.
    Adding additional relevant literature and more minor details regarding indoor air quality across different sections.
    Improving the overall readability of the report, by removing repetitive elements.
    Report Executive Summary
    (Full report available here.)
    Top-line summary
    Most efforts to address indoor air quality (IAQ) do not address airborne pathogen levels, and creating indoor air quality standards that include airborne pathogen levels could meaningfully reduce global catastrophic biorisk from pandemics.
    We estimate that an ideal adoption of indoor air quality interventions, like ventilation, filtration, and ultraviolet germicidal irradiation (GUV) in all public buildings in the US, would reduce overall population transmission of respiratory illnesses by 30-75%, with a median estimate of 52.5%.
    Bottlenecks inhibiting the mass deployment of these technologies include a lack of clear standards, cost of implementation, and difficulty changing regulation/public attitudes.
    The following actions can accelerate deployment and improve IAQ to red...
    12 min
  • EA - Here's a comprehensive fact sheet of almost all the ways animals are mistreated in factory farms by Omnizoid
    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: Here's a comprehensive fact sheet of almost all the ways animals are mistreated in factory farms, published by Omnizoid on May 4, 2023 on The Effective Altruism Forum.
    Crosspost of this on my blog.
    1 Introduction
    See, there's the difference between us. I don't care about animals at all. I'm well aware of the cramped, squalid, and altogether unpleasant conditions suffered by livestock animals on many factory farms, and I simply could not care less. I see animals as a natural resource to be exploited. I don't care about them any more than I care about the trees that were cut down to make my house.
    Random person on the bizarre anti-vegan subreddit
    I’ve previously argued against factory farming at some length, arguing that it is the worst thing ever. Here I will just lay out the facts about factory farming. I will describe what happens to the 80 or so billion beings we factory farm every year, who scream in agony and terror in the great juggernauts of despair, whose cries we ignore. They scream because of us—because of our apathy, because of our demand for their flesh—and it’s about time that people learned exactly what is going on. Here I describe the horrors of factory farms, though if one is convinced that factory farms are evil, they should stop paying for their products—an act which demonstrably causes more animals to be tormented in concentration camp-esque conditions.
    If factory farms are as cruel as I suggest, then the obligation not to pay for them is a point of elementary morality. Anyone who is not a moral imbecile recognizes that it’s wrong to contribute to senseless cruelty for the sake of comparatively minor benefits. We all recognize it’s wrong to torture animals for pleasure—paying others to torture animals for our pleasure is similarly wrong. If factory farms are half as cruel as I make them out to be, then factory farming is inarguably the worst thing in human history. Around 99% of meat comes from factory farms—if you purchase meat without careful vetting, it almost definitely comes from a factory farm.
    Here, I’ll just describe the facts about what goes on in factory farms. Of course, this understates the case, because much of what goes on is secret—the meat industry has fought hard to make it impossible to film them. As Scully notes
    It would be reasonable for the justices to ask themselves this question, too: If the use of gestation crates is proper and defensible animal husbandry, why has the NPPC lobbied to make it a crime to photograph that very practice?
    Here, I will show that factory farming is literally torture. This is not hyperbolic, but instead the obvious conclusion of a sober look at the facts. If we treated child molesters the way we treat billions of animals, we’d be condemned by the international community. The treatment of animals is unimaginably horrifying—evocative of the worst crimes in human history.
    Some may say that animals just cannot be tortured. But this is clearly a crazy view. If a person used pliers to cut off the toes of their pets, we’d regard that as torture. Unfortunately, what we do to billions of animals is far worse.
    2 Pigs
    Just like those who defended slavery, the eaters of meat often have farcical notions about how the beings whose mistreatment they defend are treated. But unfortunately, the facts are quite different from those suggested by meat industry propaganda, and are worth reviewing.
    Excess pigs were roasted to death. Specifically, these pigs were killed by having hot steam enter the barn, at around 150 degrees, leading to them choking, suffocating, and roasting to death. It’s hard to see how an industry that chokes and burns beings to death can be said to be anything other than nightmarish, especially given that pigs are smarter than dogs.
    Factory-farmed pigs, while pregnant, are stuffed in tiny gestation cr...
    45 min
  • LW - Formalizing the "AI x-risk is unlikely because it is ridiculous" argument 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: Formalizing the "AI x-risk is unlikely because it is ridiculous" argument, published by Christopher King on May 3, 2023 on LessWrong.
    There is a lot of good writing on technical arguments against AI x-risk (such as Where I agree and disagree with Eliezer (which mostly argues for more uncertainty) and others). However, in the wider world the most popular argument is more of the form "it is ridiculous" or "it is sci-fi" or some sort of other gut feeling. In particular, I think this is the only way people achieve extremely low credence on AI doom (low enough that they worry about other disasters instead).
    Although this seems like a fallacy, in this post I will attempt to formalize this argument. Not only is it good, but I think it turns out to be extremely strong!
    In my judgement, I still find the arguments for x-risk stronger or at least balanced with the "it is ridiculous" argument, but it still deserves serious study. In particular, I think analyzing and critiquing it should become a part of the AI public discourse. For example, I think there are flaws in the argument that, when revealed, would cause people to become more worried about AI x-risk. I do not quite know what these flaws are yet. In any case, I hope this post will allow us to start studying the argument.
    The argument is actually a bunch of separate arguments that tend to be lumped together into one "it is ridiculous" argument.
    For the purposes of this article, Bob is skeptical of AI x-risk and Alice argues in favor of it.
    Existential risk would stop a 10,000 year trend
    Forgetting all of human history, Bob first assumes our priors for the long term future are very much human agnostic. The vast majority of outcomes have no humans, but are just arbitrary arrangements of matter (paperclips, diamonds, completely random, etc...). So our argument will need to build a case for the long term future actually being good for humans, despite this prior.
    Next, Bob takes into account human history. The total value of the stuff humans consume tends to go up. In particular, it seems to follow a power law, which is a straight line on this log-log graph.
    Which means Bob has the gods of straight lines on his side!
    This should result in a massive update of the priors towards "the future will have lots of things that humans like".
    Most people of course don't track economics or think about power-laws, but they have an intuitive sense of human progress. This progress is pretty robust to a wide variety of disasters, but not to x-risk, and thus the model is evidence that x-risk simply won't occur.
    Clever arguments fail in unexpected ways
    However, trends do break sometimes, and AI seems pretty dangerous. In fact, Alice has very good technical theories of why it is dangerous.
    But if you go through history, you find that even very good theories are hit and miss. It is good enough to locate the hypothesis, but still has a decent chance of being wrong.
    Alice might say "but if the theory fails, that might just mean AI is bad to humans in a way I didn't expect, not that AI is safe". But our prior does say AI is safe, thanks to the gods of straight lines. And Alice does not have a god of straight lines for AI doom; if anything, AI has tended to get more useful to humans over time, not less useful.
    Out of the ideas Bob has heard, sci-fi-ness and bad-ness is correlated
    Science fiction is a pretty good predictor of the future (in that future progress has often been predicted by some previous sci-fi story). However, if Bob discovers that a new idea he heard previously occurred in sci-fi, on average this provides evidence against the idea.
    That's because if an idea is both bad and not from sci-fi, Bob is unlikely to hear it. And thus being sci-fi and being bad becomes correlated conditioned on Bob having heard about it.
    Bob should partially di...
    7 min