The Nonlinear Library

The Nonlinear Library

By The Nonlinear FundEducation
Download on the App Store

All The Nonlinear Library episodes

  • LW - Nose / throat treatments for respiratory infections by juliawise
    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: Nose / throat treatments for respiratory infections, published by juliawise on March 13, 2023 on LessWrong.
    After hearing various people interested in nasal sprays for preventing covid or the common cold, I did a shallow dive on what might be worth trying in this area.
    Overall, there are several treatments that seem safe and cheap and some of them are probably effective. A few (carrageenan and HPMC) seem to prevent infection, and the others reduce the severity of an infection once you have it. Some may also reduce your risk of infecting others if you’re already sick.
    Povidone iodine
    Nitric oxide
    Nasal washing (neti pot or various other devices)
    Hydroxypropyl methyl cellulose (HPMC)
    Carrageenan
    Xylitol
    Saline spray / wash
    Various mouthwashes
    Alcohol
    Other stuff being researched but not on the market as far as I know
    Side effects
    Best acronym
    Personal takeaways
    Obvious disclaimer: None of this replaces a vaccine or medical care. But it would be awesome if there were fewer colds in my house, if we could all get over them faster, and if we had gotten over covid sooner.
    I looked at ingredients you can currently buy (in some country), not stuff that’s not on the market. In some cases you can get them e.g. on Ebay from other countries where they’re available without prescription. I only looked at things with human trials, not things with only in vitro or animal trials.
    My understanding is that treatments aimed at the nose and throat can make sense because that’s where upper respiratory infections often first take hold. E.g. covid reproduces in nasal cilia cells for the first 24 hours after exposure before spreading to the rest of the body. A lot of antiviral treatments are known to kill covid in a petri dish, so they might also kill it in your nose and throat.
    A lot of the research here is small preliminary studies on covid, often without randomization or control groups. I’m frustrated that in 3 years there haven’t been larger, better studies published – though there are some in the works. I suspect part of this is that the ingredients are cheap and unpatentable, so no one is going to make a lot of money off them. Some of the studies are funded by the companies making the product.
    If you’re excited to dig into this more, I would love for you to make this review better!
    I would also be excited to see more research on some of these cheap treatments, if anyone reading has the power to make that happen.
    Povidone iodine
    Povidone iodine has long been used as an antiseptic, and is considered safe to use in the nose and throat.
    Study in Bangladesh: 606 people with covid were randomized to use a 1% iodine mouthwash, nose drops, and eye drops vs. water. The ones using iodine were much less likely to test positive at day 7 (3% vs 70%), and were less likely to need oxygen (3% vs 21%) or to die (1% vs 6%). If this is representative of what would happen at a larger scale, it seems like a huge deal and I’m confused why there hasn’t been more study of this. There are a bunch of studies that never finished or never published results – maybe that’s because they didn’t find iodine effective and the researchers didn’t put the time into publishing the null result?
    Another study in Bangladesh (189 people): indicates that iodine spray or nasal wash can make you test negative at least temporarily. They took symptomatic people with covid, gave them a nasal spray or nasal wash to use one time, and then collected a second PCR test right away. Some people do test negative after applying the iodine, but this doesn’t seem that relevant to whether you get better sooner. .5% nasal irrigation worked best.
    This does make me think it might be good to use something like this if you know you have covid and need to be in a public place e.g. to get medical care. Or before a situation where you’re exposing people...
    13 min
  • LW - What Discovering Latent Knowledge Did and Did Not Find by Fabien Roger
    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 Discovering Latent Knowledge Did and Did Not Find, published by Fabien Roger on March 13, 2023 on LessWrong.
    Thanks to Marius Hobbhahn and Oam Patel for helpful feedback on drafts. Thanks to Collin and Haotian for answering many questions about their work.
    Discovering Latent Knowledge in Language Models Without Supervision describes Contrast-Consistent Search (CCS), a method to find a classifier which accurately answers yes-no questions given only unlabeled model activations. It might be a stepping stone towards recovering superhuman beliefs of AI systems, as unsupervised methods are more scalable and might be less likely to simply recover “what a human would say”.
    I think this research direction is interesting and promising. But I feel like people often got a bit carried away with the approximate takeaway they got from the experimental results of the initial paper.
    In this post, I present experimental results which highlight the strengths and weaknesses of CCS.
    CCS is able to find a single linear probe which correctly classifies statements across datasets, and it doesn’t hurt performance;
    CCS does so better than random, but not by a huge margin: on average, random linear probes have a 75% accuracy on some “easy” datasets;
    CCS does not find the single linear probe with high accuracy: there are more than 20 orthogonal linear probes (i.e. using completely different information) that have similar accuracies as the linear probe found by CCS (for most datasets);
    CCS does not always find a probe with low test CCS loss (Figure 1 of the paper is misleading). CSS finds probes which are sometimes overconfident in inconsistent predictions on the test set, resulting in a test loss that is sometimes higher than always predicting a constant probability;
    CCS’ performance on GPT-J heavily depends on the last tokens of the input, especially when looking at the last layers’ activations (the setting used in the paper).
    Main takeaways:
    CCS does not simply find weird heuristics on each dataset independently, i.e. it finds some feature or property that is shared between datasets. However, we still don’t know if this feature corresponds to the model’s “beliefs”.
    Future work should compare their work against the random probe baseline. Comparing to a 50% random guessing baseline is misleading, as random probes have higher accuracy than that.
    CCS will likely miss important information about the model’s beliefs because there is more than one linear probe which achieves low loss and high CCS accuracy, i.e. there is more than one truth-like feature.
    There are many orthogonal linear probes which achieve low loss and high CCS accuracy, i.e. there are many truth-like features. Narrowing down which linear probe corresponds to the model’s beliefs might be hard.
    There exists a direction which contains all linearly available information about truth, i.e. you can’t train a linear classifier to classify true from untrue texts after projecting the activations along this direction. CCS doesn’t find it. This means CCS is ill-suited for ablation-related experiments.
    Future work should use more data or more regularization than the original paper did if it wants to find features which are actually truth-like.
    To get clean results, use CCS on UQA, and don’t get too close to GPT models. Investigating when and why CCS sometimes fails with GPT models could be a promising research direction.
    When using CCS on GPT models, don’t use CCS only on the last layer, as probes trained on activations earlier in the network are less sensitive to the format of the input.
    Experimental setup
    I’m using a modified version of the code Collin and Haotian used to run the experiments (the zip file linked in this readme).
    I report results for two models:
    UnifiedQA (T5, 11B parameters), which has the highest accuracies, and which C...
    19 min
  • AF - Discussion with Nate Soares on a key alignment difficulty by HoldenKarnofsky
    Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Discussion with Nate Soares on a key alignment difficulty, published by HoldenKarnofsky on March 13, 2023 on The AI Alignment Forum.
    In late 2022, Nate Soares gave some feedback on my Cold Takes series on AI risk (shared as drafts at that point), stating that I hadn't discussed what he sees as one of the key difficulties of AI alignment.
    I wanted to understand the difficulty he was pointing to, so the two of us had an extended Slack exchange, and I then wrote up a summary of the exchange that we iterated on until we were both reasonably happy with its characterization of the difficulty and our disagreement.1 My short summary is:
    Nate thinks there are deep reasons that training an AI to do needle-moving scientific research (including alignment) would be dangerous. The overwhelmingly likely result of such a training attempt (by default, i.e., in the absence of specific countermeasures that there are currently few ideas for) would be the AI taking on a dangerous degree of convergent instrumental subgoals while not internalizing important safety/corrigibility properties enough.
    I think this is possible, but much less likely than Nate thinks under at least some imaginable training processes.
    I didn't end up agreeing that this difficulty is as important as Nate thinks it is, although I did update my views some (more on that below). My guess is that this is one of the two biggest disagreements I have with Nate's and Eliezer's views (the other one being the likelihood of a sharp left turn that leads to a massive capabilities gap between AI systems and their supervisors.2)
    Below is my summary of:
    Some key premises we agree on.
    What we disagree about, at a high level.
    A hypothetical training process we discussed in order to get more clear and mechanistic about Nate's views.
    Some brief discussion of possible cruxes; what kind of reasoning Nate is using to arrive at his relatively high (~85%) level of confidence on this point; and future observations that might update one of us toward the other's views.
    MIRI might later put out more detailed notes on this exchange, drawing on all of our discussions over Slack and comment threads in Google docs.
    Nate has reviewed this post in full. I'm grateful for his help with it.
    Some starting points of agreement
    Nate on this section: “Seems broadly right to me!”
    An AI is dangerous if:
    It's powerful (like, it has the ability to disempower humans if it's "aiming" at that)
    It aims (perhaps as a side effect of aiming at something else) at CIS (convergent instrumental subgoals) such as "Preserve option value," "Gain control of resources that can be used for lots of things," "Avoid being turned off," and such. (Note that this is a weaker condition than "maximizes utility according to some relatively simple utility function of states of the world")
    It does not reliably avoid POUDA (pretty obviously unintended/dangerous actions) such as "Design and deploy a bioweapon."
    "Reliably" just means like "In situations it will actually be in" (which will likely be different from training, but I'm not trying to talk about "all possible situations").
    Avoiding POUDA is kind of a low bar in some sense. Avoiding POUDA doesn't necessarily require fully/perfectly internalizing some "corrigibility core" (such that the AI would always let us turn it off even in arbitrarily exotic situations that challenge the very meaning of "let us turn it off"), and it even more so doesn't require anything like CEV. It just means that stuff where Holden would be like "Whoa whoa, that is OBVIOUSLY unintended/dangerous/bad" is stuff that an AI would not do.
    That said, POUDA is not something that Holden is able to articulate cleanly and simply. There are lots of actions that might be POUDA in one situation and not in another (e.g., developing a chemical that's both poisonous and useful...
    34 min
  • LW - Plan for mediocre alignment of brain-like [model-based RL] AGI by Steven Byrnes
    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: Plan for mediocre alignment of brain-like [model-based RL] AGI, published by Steven Byrnes on March 13, 2023 on LessWrong.
    (This post is a more simple, self-contained, and pedagogical version of Post #14 of Intro to Brain-Like AGI Safety.)
    (Vaguely related to this Alex Turner post and this John Wentworth post.)
    I would like to have a technical plan for which there is a strong robust reason to believe that we’ll get an aligned AGI and a good future. This post is not such a plan.
    However, I also don’t have a strong reason to believe that this plan wouldn’t work. Really, I want to throw up my hands and say “I don’t know whether this would lead to a good future or not”. By “good future” here I don’t mean optimally-good—whatever that means—but just “much better than the world today, and certainly much better than a universe full of paperclips”. I currently have no plan, not even a vague plan, with any prayer of getting to an optimally-good future. That would be a much narrower target to hit.
    Even so, that makes me more optimistic than at least some people. Or at least, more optimistic about this specific part of the story. In general I think many things can go wrong as we transition to the post-AGI world—see discussion by Dai & Soares—and overall I feel very doom-y, particularly for reasons here.
    This plan is specific to the possible future scenario (a.k.a. “threat model” if you’re a doomer like me) that future AI researchers will develop “brain-like AGI”, i.e. learning algorithms that are similar to the brain’s within-lifetime learning algorithms. (I am not talking about evolution-as-a-learning-algorithm.) These algorithms, I claim, are in the general category of model-based reinforcement learning. Model-based RL is a big and heterogeneous category, but I suspect that for any kind of model-based RL AGI, this plan would be at least somewhat applicable. For very different technological paths to AGI, this post is probably pretty irrelevant.
    But anyway, if someone published an algorithm for x-risk-capable brain-like AGI tomorrow, and we urgently needed to do something, this blog post is more-or-less what I would propose to try. It’s the least-bad plan that I currently know.
    So I figure it’s worth writing up this plan in a more approachable and self-contained format.
    1. Intuition: Making a human into a moon-lover (“selenophile”)
    Try to think of who is the coolest / highest-status-to-you / biggest-halo-effect person in your world. (Real or fictional.) Now imagine that this person says: “You know what’s friggin awesome? The moon. I just love it. The moon is the best.”
    You stand there with your mouth agape, muttering to yourself in hushed tones: “Wow, huh, the moon, yeah, I never thought about it that way.” (But 100× moreso. Maybe you’re on some psychedelic at the time, or this is happening during your impressionable teenage years, or whatever.) You basically transform into a “moon fanboy” / “moon fangirl” / “moon nerd” / “selenophile”.
    How would that change your motivations and behaviors going forward?
    You’re probably going to be much more enthusiastic about anything associated with the moon.
    You’re probably going to spend a lot more time gazing at the moon when it’s in the sky.
    If there are moon-themed trading cards, maybe you would collect them.
    If NASA is taking volunteers to train as astronauts for a trip to the moon, maybe you’d enthusiastically sign up.
    If a supervillain is planning to blow up the moon, you’ll probably be extremely opposed to that, and motivated to stop them.
    Hopefully this is all intuitive so far.
    What’s happening mechanistically in your brain? As background, I think we should say that one part of your brain (the cortex, more-or-less) has “thoughts”, and another part of your brain (the basal ganglia, more-or-less) assigns a “value” (in RL terminology) a....
    21 min
  • EA - Shallow Investigation: Stillbirths by Joseph Pusey
    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: Shallow Investigation: Stillbirths, published by Joseph Pusey on March 13, 2023 on The Effective Altruism Forum.
    This topic has the potential to be deeply upsetting to those reading it, particularly to those who have personal experience of the topic in question. If you feel that I’ve missed or misunderstood something, or could have phrased things more sensitively, please reach out to me.
    Throughout the review, words like “woman” or “mother” are used in places where some people might prefer “birthing person” or similar. This choice reflects the language used in the available literature and does not constitute a position on what the most appropriate terminology is.
    This report is a shallow dive into stillbirths, a sub-area within maternal and neonatal health, and was produced as part of the Cause Innovation Bootcamp. The report, which reflects approximately 40-50 hours of research, offers a brief dive into whether a particular problem area is a promising area for either funders or founders to be working in. Being a shallow report, it should be used to decide whether or not more research and work into a particular problem area should be prioritised.
    Executive Summary
    Importance: This problem is likely very important (epistemic status-strong)- stillbirths are widespread, concentrated in the world’s poorest countries, and decreasing only very slowly compared to the decline in maternal and infant mortality. There are more deaths resulting from stillbirth than those caused by HIV and malaria combined (depending on your personal definition of death- see below), and even in high-income countries stillbirths outnumber infant deaths.
    Tractability: This problem is likely moderately tractable (moderate)- most stillbirths are likely to be preventable, but the most impactful interventions are complex, facility-based, expensive, and most effective at scale e.g. guaranteeing access to high-quality emergency obstetric care
    Neglectedness: This problem is unlikely to be neglected (less strong)- although still under-researched and under-counted, stillbirths are the target of some of the largest organisations in the global health and development world, including the WHO, UNICEF, the Bill and Melinda Gates Foundation, and the Lancet. Many countries have committed to the Every Newborn Action Plan, which aims- amongst other things- to reduce the frequency of stillbirths.
    Key uncertainties
    Key uncertainty 1: Accurately assessing the impact of stillbirths, and therefore the cost-effectiveness of interventions aimed at reducing stillbirths, depends significantly on to what extent direct costs to the unborn child are counted. Some organisations view stillbirths as having negative effects on the parents and wider communities but do not count the potential years of life lost by the unborn child; others use time-discounting methods to calculate a hypothetical number of expected QALYS lost, and still others see it as completely equivalent to losing an averagely-long life. Differences in the weighting of this loss can alter the calculated impacts of stillbirth by several orders of magnitude and is likely the most important consideration when considering a stillbirth-reducing intervention
    Key uncertainty 2: Interventions which reduce the risk of stillbirth tend to be those which also address maternal and neonatal health more broadly; therefore, it is very difficult to accurately assess the cost-effectiveness of these interventions solely in terms of their impact on stillbirths, and more complex models which take into account the impacts on maternal, neonatal, and infant health are likely more accurate in assessing the overall cost-effectiveness of interventions.
    Key uncertainty 3: A large proportion of the data around interventions to reduce stillbirths comes from high-income countries, but most still...
    26 min
  • LW - your terminal values are complex and not objective by carado
    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: your terminal values are complex and not objective, published by carado on March 13, 2023 on LessWrong.
    a lot of people seem to want terminal (aka intrinsic aka axiomatic) values (aka ethics aka morality aka preferences aka goals) to be simple and elegant, and to be objective and canonical. this carries over from epistemology, where we do favor simplicity and elegance.
    we have uncertainty about our values, and it is true that our model of our values should, as per epistemology, generally tend to follow a simplicity prior. but that doesn't mean that our values themselves are simple; they're definitely evidently complex enough that just thinking about them a little bit should make you realize that they're much more complex than the kind of simple model people often come up with.
    both for modeling the world and for modeling your values, you should favor simplicity as a prior and then update by filtering for hypotheses that match evidence, because the actual territory is big and complex.
    there is no objectively correct universal metaethics. there's just a large, complex, tangled mess of stuff that is hard to categorize and contains not just human notions but also culturally local notions of love, happiness, culture, freedom, friendship, art, comfort, diversity, etc. and yes, these are terminal values; there is no simple process that re-derives those values. i believe that there is no thing for which i instrumentally value love or art, which if you presented me something else that does that thing better, i would happily give up on love/art. i value those things intrinsically.
    if you talk of "a giant cosmopolitan value handshake between everyone", then picking that rather than paperclips, while intuitive to you (because you have your values) and even to other humans doesn't particularly track anything universally canonical.
    even within the set of people who claim to have cosmopolitan values, how conflicts are resolved and what "everyone" means and many other implementation details of cosmopolitanism will differ from person to person, and again there is no canonical unique choice. your notion of cosmopolitanism is a very complex object, laden with not just human concepts but also cultural concepts you've been exposed to, which many other humans don't share both across time and space.
    there is no "metaethics ladder" you can which climb up in order to resolve this in an objective way for everyone, not even all humans — what ladder and how you climb it is still a complex subjective object laden with human concepts and concepts from your culture, and there is no such thing as a "pure" you or a "pure" person without those.
    some people say "simply detect all agents in the cosmos and do a giant value handshake between those"; but on top of the previous problems for implementation details, this has the added issue that the things whose values we want to be satisfied aren't agents but moral patients. those don't necessarily match — superintelligent grabby agents shouldn't get undue amounts of power in the value handshake.
    some people see the simplicity of paperclips as the problem, and declare that complexity or negentropy or something like that is the ultimate good. but a superintelligence maximizing for that would just fill the universe with maximally random noise, as opposed to preserving the things you like. turns out, "i want whatever is complex" is not sufficient to get our values; they're not just anything complex or complexity itself, they're an extremely specific complex set of things, as opposed to other equally complex sets of things.
    entropy just doesn't have much to do with terminal values whatsoever. sure, it has a lot to do with instrumental values: negentropy is the resource we have to allocate to the various things we want. but that's secondary to what it is we want...
    4 min
  • EA - On taking AI risk seriously by Eleni A
    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: On taking AI risk seriously, published by Eleni A on March 13, 2023 on The Effective Altruism Forum.
    Yet another New York Times piece on AI. A non-AI safety friend sent it to me saying "This is the scariest article I've read so far. I'm afraid I haven't been taking it very seriously". I'm noting this because I'm always curious to observe what moves people, what's out there that has the power to change minds. In the past few months, there's been increasing public attention to AI and all sorts of hot and cold takes, e.g., about intelligence, consciousness, sentience, etc. But this might be one of the articles that convey the AI risk message in a language that helps inform and think about AI safety.
    The following is what stood out to me and made me think that it's time for philosophy of science to also take AI risk seriously and revisit the idea of scientific explanation given the success of deep learning:
    I cannot emphasize this enough: We do not understand these systems, and it’s not clear we even can. I don’t mean that we cannot offer a high-level account of the basic functions: These are typically probabilistic algorithms trained on digital information that make predictions about the next word in a sentence, or an image in a sequence, or some other relationship between abstractions that it can statistically model. But zoom into specifics and the picture dissolves into computational static.
    “If you were to print out everything the networks do between input and output, it would amount to billions of arithmetic operations,” writes Meghan O’Gieblyn in her brilliant book, “God, Human, Animal, Machine,” “an ‘explanation’ that would be impossible to understand.”
    Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.
    2 min
  • AF - Plan for mediocre alignment of brain-like [model-based RL] AGI by Steve Byrnes
    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: Plan for mediocre alignment of brain-like [model-based RL] AGI, published by Steve Byrnes on March 13, 2023 on The AI Alignment Forum.
    (This post is a more simple, self-contained, and pedagogical version of Post #14 of Intro to Brain-Like AGI Safety.)
    (Vaguely related to this Alex Turner post and this John Wentworth post.)
    I would like to have a technical plan for which there is a strong robust reason to believe that we’ll get an aligned AGI and a good future. This post is not such a plan.
    However, I also don’t have a strong reason to believe that this plan wouldn’t work. Really, I want to throw up my hands and say “I don’t know whether this would lead to a good future or not”. By “good future” here I don’t mean optimally-good—whatever that means—but just “much better than the world today, and certainly much better than a universe full of paperclips”. I currently have no plan, not even a vague plan, with any prayer of getting to an optimally-good future. That would be a much narrower target to hit.
    Even so, that makes me more optimistic than at least some people. Or at least, more optimistic about this specific part of the story. In general I think many things can go wrong as we transition to the post-AGI world—see discussion by Dai & Soares—and overall I feel very doom-y, particularly for reasons here.
    This plan is specific to the possible future scenario (a.k.a. “threat model” if you’re a doomer like me) that future AI researchers will develop “brain-like AGI”, i.e. learning algorithms that are similar to the brain’s within-lifetime learning algorithms. (I am not talking about evolution-as-a-learning-algorithm.) These algorithms, I claim, are in the general category of model-based reinforcement learning. Model-based RL is a big and heterogeneous category, but I suspect that for any kind of model-based RL AGI, this plan would be at least somewhat applicable. For very different technological paths to AGI, this post is probably pretty irrelevant.
    But anyway, if someone published an algorithm for x-risk-capable brain-like AGI tomorrow, and we urgently needed to do something, this blog post is more-or-less what I would propose to try. It’s the least-bad plan that I currently know.
    So I figure it’s worth writing up this plan in a more approachable and self-contained format.
    1. Intuition: Making a human into a moon-lover (“selenophile”)
    Try to think of who is the coolest / highest-status-to-you / biggest-halo-effect person in your world. (Real or fictional.) Now imagine that this person says: “You know what’s friggin awesome? The moon. I just love it. The moon is the best.”
    You stand there with your mouth agape, muttering to yourself in hushed tones: “Wow, huh, the moon, yeah, I never thought about it that way.” (But 100× moreso. Maybe you’re on some psychedelic at the time, or this is happening during your impressionable teenage years, or whatever.) You basically transform into a “moon fanboy” / “moon fangirl” / “moon nerd” / “selenophile”.
    How would that change your motivations and behaviors going forward?
    You’re probably going to be much more enthusiastic about anything associated with the moon.
    You’re probably going to spend a lot more time gazing at the moon when it’s in the sky.
    If there are moon-themed trading cards, maybe you would collect them.
    If NASA is taking volunteers to train as astronauts for a trip to the moon, maybe you’d enthusiastically sign up.
    If a supervillain is planning to blow up the moon, you’ll probably be extremely opposed to that, and motivated to stop them.
    Hopefully this is all intuitive so far.
    What’s happening mechanistically in your brain? As background, I think we should say that one part of your brain (the cortex, more-or-less) has “thoughts”, and another part of your brain (the basal ganglia, more-or-less) assigns a “value” (in RL ter...
    21 min
  • EA - How bad a future do ML researchers expect? by Katja Grace
    Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: How bad a future do ML researchers expect?, published by Katja Grace on March 13, 2023 on The Effective Altruism Forum.
    Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.
    1 min
  • EA - It's not all that simple by Brnr001
    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: It's not all that simple, published by Brnr001 on March 13, 2023 on The Effective Altruism Forum.
    TLTR: I feel that recently the EA forum became pretty judgmental and unwelcoming. I also feel that the current discourse about sex misses two important points and, in a huge part of it, lacks maturity and is harmful. Let me attempt to address it. Trigger warning, point 2 involves a long description of personal stories connected to sex, some of them were difficult and may be triggering. It also may not be very well structured, but I preferred to write one long post instead of three short ones.
    This is obviously a burner account, but when you see those stories you’ll be able to see why. For the record, they don’t involve people from the community. I'm a woman (it's going to matter later on).
    Acceptable dating and sexual behaviors vary between classes and cultures. The devil is in the detail, and rules you live by and perceive as “obvious” may be so clear for anybody else. Also, the map of the US is not in a shape of geode.
    People vary in gender and sexual orientation. They vary in a level of sexual desire. They have different kinks, ways of expressing sexuality and levels of self-awareness. Different needs. Various physiological reactions to sexually tense situations. Various ways of presenting themselves when it comes to all of the above.
    People come from different cultures – regions, countries, social classes and religions. As a result, dating cultures vary around the world. Sexual behaviors also. Acceptable level of flirt, jokes, touch and the way consent is asked for and expressed sometimes just vary. Problems and how i.e. sexism looks like also has various shapes and forms. There are some common characteristics, but details matter, to a huge extent. Many people in the recent discussions stated that various nuances are obvious and should be intuitively followed by everyone. I think it’s problematic and leads to abuse.
    Believing that your values and behavior associated with your culture and class are the only right ones and everybody should know, understand and follow them, is fundamentally different from assertively vocalizing your boundaries and needs. The second is a great, mature behavior. The first feels a bit elitist, ignorant and has nothing to do with safety, equality and being inclusive.
    Additionally, I want to draw your attention to one thing. I have a strong belief (correct me if I’m wrong) that the vast majority (if not all) of sexual misconduct causes which were described over the last couple of days in the articles or here, on the forum, come from either US or the UK. EA crowd is definitely not limited to those. So my honest question would be – is it EA who has a problem with sexual misconduct? Or is it an Anglo-Saxon culture which has a problem with sexual misconduct? Or maybe – EA with a mix of Anglo-Saxon culture has this issue? Shouldn’t we zoom in on that a bit?
    Human sexuality is complex. Consent is also sometimes complex.
    People often talk a lot of “what consent norms should be”. But often such disputes do not give a full picture of what people’s actual behaviors around consent actually are – and it’s a bit crucial to this whole conversation. If you start having more intimate talks, however, you end up seeing a much more complex and broad picture. And often consent is easier said than done.
    I encourage you all, regardless what’s your gender, to have those talks with friends, who are open and empathetic. I’ve learned a lot and they made my life easier.
    Yet, some people may have no opportunity to hear such stories. So let me share, why do I think that consent is not all that easy. I'm going to talk about myself here, because maybe somebody needs to hear somebody being open and vulnerable about stuff like that. My message is - it's ok to sometimes stru...
    14 min