The Nonlinear Library

The Nonlinear Library

By The Nonlinear FundEducation
Download on the App Store

All The Nonlinear Library episodes

  • AF - Interlude: But Who Optimizes The Optimizer? by Paul Bricman
    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: Interlude: But Who Optimizes The Optimizer?, published by Paul Bricman on September 23, 2022 on The AI Alignment Forum.
    This post is part of my hypothesis subspace sequence, a living collection of proposals I'm exploring at Refine. Preceded by representational tethers.
    Intro
    Consider diffusion models. DALL-E is able to take in an image as input and imperceptibly edit it so that it becomes slightly more meaningful (i.e. closer to the intended content). In this, DALL-E is an optimizer of images. It can steadily nudge images across image space towards what people generally deem meaningful images. But an image diffusion model need not nudge images around in this specific way. Consider an untrained version of DALL-E, whose parameters have just been initialized a moment ago. It still implicitly exerts currents across image space, it's just that those currents don't systematically lead towards the manifold of meaningful images. We deem an image diffusion model good if the field it implements nudges images in a way we deem appropriate. Conversely, a bad one simply specifies a wildly different field.
    Naturally, we'd like good models. However, we lack an exhaustive closed-form description of the desired field for pushing towards meaningful images. The best we managed to do is come up with a way of roughly telling which of two models is better. We then use (1) this imperfect relative signal, and (2) a sprinkle of glorified hill-climbing as means of obtaining a good model by charting our way through model space. We might fill those in with e.g. (1) incremental denoising performance on ImageNet, and (2) SGD. Those two ingredients help us define a trainer for the diffusion model, a way from getting from bad models towards good ones. We deem a trainer good if the field it implements nudges models in a way we deem appropriate (i.e. from bad to good). Conversely, a bad one simply specifies a wildly different field across model space.
    Naturally, we'd like good trainers. However, we lack an exhaustive closed-form description of the desired field for reliably pushing towards truly good models. You've guessed it, we did manage to come up with a way of roughly telling which of two trainers is better. We similarly use this rough relative signal and our favorite hill-climbing algorithm as means of obtaining a good trainer by charting our way across trainer space. We might fill those in with e.g. (1) convergence speed, and (2) grid search across hyperparameters. Those two help us define a hyperparameter search for the trainer, a way from getting from bad trainers towards good ones. A hyperparam search is good if its field nudges trainers around nicely. Bad ones nudge differently.
    Naturally, we'd like... Ok, you got the point. My next point is that there's nothing special about diffusion. GPT-3 drives text from gibberish towards coherent human-like text. It drives text from one topic to another, from one stage of a narrative to the next. The field it implements is the Janusian physics of language, with the associated transition function defining it. In this framing, GPT-3's trainer is tasked with roughly nudging bad models-in-training towards actually good ones, by means of (1) perplexity on giant text dataset, and (2) probably some advanced SGD.
    There's nothing special about current mainstream training regimes as well. You could have GPT-3 at the base nudging text around language space, a trainer helping it learn by gradient descent, but then instead of conventional hyperparam tuning above the trainer, you could go with learning to learn by gradient descent by gradient descent. And on and on to the cube and beyond. Or instead, you could have a "prompter" as a prompt optimizer trying to generate prompts which get the base GPT-3 to behave in certain ways. And then learn to prompt the prompter ...
    18 min
  • LW - Public-facing Censorship Is Safety Theater, Causing Reputational Damage by Yitz
    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: Public-facing Censorship Is Safety Theater, Causing Reputational Damage, published by Yitz on September 23, 2022 on LessWrong.
    It's so common it's a stereotype.
    A large corporation releases a cutting-edge AI model, and puts out a press release talking about how their new, [larger/smaller]-than-ever model provides unprecedented freedom for [underprivileged artists/small business owners/outside researchers] to do whatever it is their AI does. You go to their website, start playing with the model, and before long
    Results containing potentially sensitive content have been omitted. Further requests of this type may result in account suspension, etc., etc., etc....
    or something along those lines. The prompt you gave was pretty innocuous, but in retrospect you can sort of see how maybe the output might have resulted in something horrifically offensive, like a curse word, or even (heaven forbid) an image that has a known person's face in it. You've been protected from such horrors, and this is reassuring. Of course, your next prompt for whatever reason elicits [insert offensive stereotype/surprisingly gory or uncanny imagery/dangerously incorrect claim presented with high confidence/etc. here], which is slightly less reassuring.
    Checking the details of the press release, you see a small section of the F.A.Q. with the disclaimer that some outputs may be biased due to [yadda yadda yadda you know the drill]. You breathe a sigh of relief, secure in the comforting knowledge that [faceless company] cares about AI safety, human rights, and reducing biases. Their model isn't perfect, but they're clearly working on it!
    The above scenario is how [large corporations] seem to expect consumers to react to their selective censorship. In reality I strongly suspect that the main concern is not so much protecting the consumer as it is protecting themselves from liability. After all, by releasing a model which is clearly capable of doing [harmful capability], and by giving sufficient detail to the public that their model can be replicated, [harmful capability] has effectively been released, if perhaps delayed by a few months at most. However, whoever makes the open-source replication will not be [large corporation], absolving the company of perceived moral (and legal) culpability in whatever follows. If the concern were actually that [harmful capability] would lead to real danger, then the moral thing to do would be not to release the model at all.
    There are a few serious problems with this. The most obvious (and generic) objection is that censorship is bad. When looking at historical incidents of censorship we often find ourselves morally disagreeing with the censors, who got to choose what is considered inappropriate from a position of power. Almost everyone agrees that Hollywood's infamous Hays code was a moral mistake. In the present day, inconsistent or weaponized social media censorship is widespread, with seemingly nobody happy with how large corporations enforce their rules (though the details of how they are failing are arguable). At least one Chinese text-to-image model disallows prompts which include the word "democracy". It would be surprising to me if protections against generating certain forms of content with LLMs don't eventually lead to unexpected negative social consequences.
    Secondly, there is a danger of AI safety becoming less robust—or even optimising for deceptive alignment—in models using front-end censorship. If it's possible for a model to generate a harmful result from a prompt, then the AI is not aligned, even if the user can't see the bad outputs once they are generated. This will create the illusion of greater safety than actually exists, and (imo) is practically begging for something to go wrong. As a "tame" example, severe bugs could crop up which are left unad...
    10 min
  • EA - My Personal Story of Rejection from EA Global by Constance Li
    Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: My Personal Story of Rejection from EA Global, published by Constance Li on September 23, 2022 on The Effective Altruism Forum.
    I am the person rejected from EA Global who was anonymously featured in the Scott Alexander post about Opening EAG.
    I have decided to publicly identify myself and offer my story as an in-depth case study on the EAG admissions process. I’ve dedicated much of my life to EA principles since 2010 and yet I kept getting rejected from attending EAG. This experience brought up a lot of sadness and frustration as I felt like I was being pushed out of a community that I had always closely identified with. In the process of creating this post, I had a lot of time to reflect on EA as a personal identity vs a philosphy vs a community vs organizations. They are distinct from one another and yet are still tangled together. I will share many criticisms about the admissions process and will also be proposing some solutions to improving it. By posting my story, I hope to lower the barrier for an honest discussion about EAG Admissions, EA disillusionment, gatekeeping, community building, and solutions for improving EA communities and organizational processes.
    Goals:
    To allow reader to compare my application to the fact that I was rejected 3 times from EAG
    To offer more insight into the EAG admissions process by showing all my correspondences with CEA
    To critically discuss problems I see with the admissions process and propose some solutions
    To promote more easily accessible EA communities such as Gather Town
    To create a safer space for others to share their own stories so more experiences can be heard and accounted for
    Meet Up at Gather Town During EAG DC
    Since I had already blocked out my schedule in anticipation of EAG DC, I will be available most of this weekend (September 23 - 25) on EA Gather Town to video chat with anyone who wants to.
    Disclosures:
    This is my first forum post.
    Since I know this post could potentially come across as bitter, I want to say upfront that it’s always easy to criticize the actions of others and imagine that one could do things better. I make no claim about being able to do a better job at conference admissions. Through sharing my perspective, I hope to contribute to a discussion that helps everyone create a better event.
    Please feel free to skip to the “Criticisms and Proposed Solutions” section if you want to skip my lengthy personal story. To optimize for both transparency and brevity, I have also created an appendix which I will link to throughout this post.
    TLDR Timeline Summary:
    2001 - 2010 - Hadn’t learned about Effective Altruism yet, but tried my best to do good for non-human animals
    2010 - 2021- Discovered effective altruism and was engaged in EA work for 10+ years. Around 2012, I pivoted career paths because earning to give was the main EA recommendation at the time and went on to become a physician. I still pursued independent projects effectively and altruistically throughout medical school and residency, but was not directly involved in the EA community.
    Fall 2021- Decided to rejoin the EA community after EAGxVirtual
    Winter 2021 - Got rejection #1 for EAG London
    April 19, 2022 - Applied to EAG DC and heard no response (4 months) so assumed I was rejected
    July 2022- Made my original post as a reaction to a friend’s facebook post about EA Disillusionment
    August 15, 2022 - Had a long video chat with a sympathetic CEA staff member, “X,” who saw my post and received feedback, tips for re-integrating with the EA community, and encouragement for reapplying to EAG DC
    August 18, 2022 - Attended my first EA NYC event because of my talk with X
    August 22, 2022 - Got official rejection #2 for EAG DC (felt sad, but figured it was because my application from 4 months earlier was not that great)
    Aug - September, 2022 - Continued ...
    35 min
  • EA - $5k challenge to quantify the impact of 80,000 hours' top career paths by NunoSempere
    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: $5k challenge to quantify the impact of 80,000 hours' top career paths, published by NunoSempere on September 23, 2022 on The Effective Altruism Forum.
    Motivation
    80,000 hours has identified a number of promising career paths. They have a fair amount of analysis behind their recommendations, and in particular, they have a list of top ten priority paths.
    However, 80,000 hours doesn’t quite have quantitative estimates of these paths' value. Although their usefulness would not be guaranteed, quantitative estimates could make it clearer:
    how valuable their top career paths are relative to each other
    how valuable their top career paths are relative to options further down their list
    at which level of personal fit one should switch between different career paths
    where the expected impact is coming from, and which variables we are most uncertain about
    eventually, whether certain opportunities are valuable in themselves or for the value of information or career capital that they provide
    etc.
    The Prize
    Following up on the $1,000 Squiggle Experimentation Challenge and the Forecasting Innovation Prize we are offering a prize of $5k for quantitative estimates of the value of 80,000 hours' top 10 career paths.
    Rules
    Step 1: Make a public post online between now and December 1, 2022. Posts on the EA Forum (link posts are fine) are encouraged.Step 2: Complete this submission form.
    Further details
    Participants can use units or strategies of their choice—these might be QALYs, percentage points of reduction in existential risk, basis points of the future, basis points of existential risk reduced, career-dependent units, etc. Contestants could also use some other method, like relative values, estimating proxies, or some original option.
    We are specifically looking for quantitative estimates that attempt to estimate some magnitude reasonably close to the real world, similar to the units above. So for example, assigning valuations from 0 to 5 stars would not fulfil the requirements of the contest, but estimates in terms of the units above would qualify.
    Participants are free to estimate the value of one, several, or all ten career paths.
    Participants are free to use whatever tool or language they want to produce these estimates. Some possible tooling might be: Excel, Squiggle, Guesstimate, probabilistic languages or libraries (e.g., Turing.jl, PyMC3, Stan), Causal, working directly in a popular programming language, etc.
    Participants can provide point estimates of impact, but they are encouraged to provide their estimates as distributions instead.
    Participants are free to estimate the impact of a marginal person, of a marginal person with a good fit, the average value, etc. Participants are welcome to provide both average and marginal value—for example, they could provide a function which provides an estimate of marginal value at different levels of labor and capital.
    We provide some examples of possible rough submissions in an appendix. We are also happy to comment on estimation strategies: feel free to leave a comment on this post or to send a message to Nuño Sempere using the EA forum message functionality.
    Judging
    The judges will be Nuño Sempere, Eli Lifland, Alex Lawsen and Sam Nolan. These judges will judge on their personal capacities, and their stances do not represent their organizations.
    Judges will estimate the quality and value of the entries, and we will distribute the prize amount of $5k in proportion to an equally weighted aggregate of those subjective estimates.
    To reduce our operational burden, we are looking to send out around three to five prizes. If there are more than five submissions, we plan to implement a lottery system. For example, a participant who would have won $100 would instead get a 10% chance of receiving $1k.
    Acknowledgements
    This contest is a project of...
    11 min
  • LW - A game of mattering by KatjaGrace
    Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: A game of mattering, published by KatjaGrace on September 23, 2022 on LessWrong.
    When I have an overwhelming number of things to do, and insufficient native urge to do them, I often arrange them into a kind of game for myself. The nature and appeal of this game has been relatively stable for about a year, after many years of evolution, so this seems like a reasonable time to share it. I also play it when I just want to structure my day and am in the mood for it. I currently play something like two or three times a week.
    The game
    The basic idea is to lay out the tasks in time a bit like obstacles in a platformer or steps in Dance Dance Revolution, then race through the obstacle course grabbing them under consistently high-but-doable time pressure.
    Here’s how to play:
    Draw a grid with as many rows as there are remaining hours in your hoped for productive day, and ~3 columns. Each box stands for a particular ~20 minute period (I sometimes play with 15m or 30m periods.)
    Lay out the gameboard: break the stuff you want to do into appropriate units, henceforth ‘items’. An item should fit comfortably in the length of a box, and it should be easy enough to verify completion. (This can be achieved through house rules such as ‘do x a tiny bit = do it until I have a sense that an appropriate tiny bit has been done’ as long as you are happy applying them). Space items out a decent amount so that the whole course is clearly feasible. Include everything you want to do in the day, including nice or relaxing things, or break activities. Drinks, snacks, tiny bouts of exercise, looking at news sites for 5 minutes, etc. Design the track thoughtfully, with hard bouts followed by relief before the next hard bout.
    To play, start in the first box, then move through the boxes according to the time of day. The goal in playing is to collect as many items as you can, as you are forced along the track by the passage of time. You can collect an item by doing the task in or before you get to the box it is in. If it isn’t done by the end of the box, it gets left behind. However if you clear any box entirely, you get to move one item anywhere on the gameboard. So you can rescue something from the past, or rearrange the future to make it more feasible, or if everything is perfect, you can add an entirely new item somewhere.
    I used to play this with tiny post-it stickers, which I would gather in a large moving pile, acting as a counter:
    Now I just draw the whole thing. Crossed out = collected; [] = rescued from the past, now implicitly in the final box; dot in the lower right = box cleared; dot next to item = task done but item stuck in the past (can be collected immediately if rescued).
    Why is this good?
    I think a basic problem with working on a big pile of things in a big expanse of time is that if you work or not during any particular minute, it feels like it makes nearly no difference to the expectation of success. I’m not quite sure why this is—in fact if I don’t work this minute, I’m going to get one minute less work done. But it feels like if I don’t work this minute, I only need to work a smidgen faster on average to get any particular amount of work done, so what does it matter if I work now or later? And if i had some particular goal (e.g. finishing writing some massive text today), it’s unlikely that my other efforts will get me exactly to the line where this minute pushed me over—probably I will either succeed with hours to spare (haha) or fail hours from my goals.
    I picture what’s going on as vaguely something like this—there is often some amount of work that is going to make your success likely, and if you know that you are on a locally steep part of the curve, it is more motivating than if you are either far away from the steep part or don’t know where you are:
    Yet on the other hand, th...
    8 min
  • EA - The $100,000 Truman Prize: Rewarding Anonymous EA Work by Drew Spartz
    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 $100,000 Truman Prize: Rewarding Anonymous EA Work, published by Drew Spartz on September 22, 2022 on The Effective Altruism Forum.
    Harry Truman once said: "It's amazing what you can accomplish if you don't care who gets the credit"
    The Truman Prize, now live on EA prize platform Superlinear, recognizes Effective Altruists with $5,000-$10,000 prizes for declining credit in order to increase their impact, in ways that can't be publicized directly.
    Theory of change: EA promotes caring about effectiveness over other goals like getting credit, but wanting credit or recognition for your work is natural. Rewarding people for maximizing impact over credit increases the health and future effectiveness of the community.
    Example #1: Sam toils behind the scenes and makes a breakthrough on an important problem. Sam suggests the idea to, say, a political figure or other organization who can then take credit, because that leads to the breakthrough being more widely accepted.
    Anyone that knows what happened, including the person/org that gets credit, can nominate Sam for The Truman Prize on Superlinear. Superlinear passes on the nomination to a committee of well-respected EAs from diverse backgrounds. If one of them verifies that Sam actually did make a breakthrough and allowed someone better placed to take credit to increase impact, Superlinear awards Sam $10,000.
    The Truman Prize is the brainchild of David Manheim, and the judges are:
    Eliezer Yudkowsky
    Peter Wildeford
    Spencer Greenberg
    Cate Hall
    Julia Wise
    Gregory Lewis
    Luke Freeman
    Ozzie Gooen
    The criteria, generally speaking, is that if you can’t make an EA forum post about someone for doing something noteworthy in order to publicize what they have done, they could be eligible for the prize.
    Example #2: Greg works for the government. There are political or career consequences if it is publicly acknowledged that he’s working on something potentially controversial. Greg contributes an important idea to a research field and helps make it happen behind the scenes. Someone nominates him for The Truman Prize, and the committee asks someone in a position to know about what occurred, and confirms Greg’s contribution. Superlinear awards Greg $5,000, and announces that a prize was awarded to the originator of the research idea to a recipient to be named in 5 years.
    Example #3: Claire, a biosecurity researcher, blows the whistle internally on a potentially dangerous research direction which likely violates the Biological Weapons Convention. The organization doesn’t want this to be public, but an individual inside the organization could still confidentially nominate them for the Truman Prize, and depending on details and the potential infohazards, the prize might be awarded to Claire, without specifying what was done.
    Example #4: Max has a criminal record and troubled past. He’s reformed now, but his background makes him a liability for any person or org to publicly associate with him. He silently does good work behind the scenes, so someone that knows him nominates him for The Truman Prize on the basis of a specific critical contribution which was made to a now successful larger project. The committee awards the prize, and names Max, likely without naming the specific work done.
    Example #5: Steve has extreme political beliefs. It is risky for any person or org to work with him due to reputation risks. Steve knows this, but does apolitical high impact work behind the scenes anyways. Someone that knows Steve nominates him for The Truman Prize on the basis of a specific project which was not previously disclosed. The committee awards the prize and discloses the project, but not the individual, or vice-versa, to avoid undermining the project.
    Example #6: Morgan has recurring depression. Therefore, she does not want to work or associate wi...
    8 min
  • LW - Methodological Therapy: An Agenda For Tackling Research Bottlenecks by adamShimi
    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: Methodological Therapy: An Agenda For Tackling Research Bottlenecks, published by adamShimi on September 22, 2022 on LessWrong.
    This work was done while at Conjecture.
    Epistemic Status: Palimpsest
    Better epistemology should make you stronger.
    Which is why at Conjecture's' epistemology team we are so adamant on improving our models of knowledge-production: this feels like the key to improving alignment research across the board, given the epistemological difficulties of the field.
    Yet we have failed until now to make our theory of impact legible, both to ourselves and to well-meaning external reviewers. The most sorely missing piece is the link between better models of knowledge-production and quick improvements of alignment research, in the shorter timelines that we expect at Conjecture .
    Following interviews that we conducted with a handful of alignment researchers (John Wentworth, Vanessa Kosoy, Evan Hubinger, Abram Demski, Steve Byrnes, Conjecture researchers, and Refine participants), we want to present our current best guess for framing how our epistemology research can make alignment research stronger: revealing, analyzing, and expanding or replacing what we call "Cached Methodologies" — patterns that encode how research is supposed to proceed in a given context, for example the idea that we need to prove a statement to learn if it's true or not. Given that this involves bringing to light and questioning cached thoughts about methodology, we dub this approach methodological therapy.
    Note that we definitely also want to leverage better models of knowledge-production to help field-builders and newcomers; our current focus on explicit applications for established researchers comes from two key factors: we're currently only three people in the epistemology team, which means we have to prioritize; and we expect that models and tools useful to established researchers will prove instrumental in building ones for field-builders and newcomers.
    On credit: this idea emerged from discussions within the epistemology team (composed of Adam Shimi, Lucas Teixeira, and Daniel Clothiaux); Adam contributed the overall perspective, the unification and the therapy metaphor (from Bachelard); Lucas contributed the focus on actual cached patterns in researchers' minds and interviews; Daniel contributed the expansion of the idea to encompass more alternatives (for examples more ways of doing experiments) as well as many insights from the literature. This post is written by Adam; "we" means that the opinion is that of the whole team, and "I" means it's only Adam's opinion.
    Thanks to Clem von Stengel and Andrea Motti for feedback on drafts.
    How Do Researchers Think About Research?
    Our starting point was conducting interviews with researchers, in order to improve the product-market fit of our epistemological research by better understanding and targeting the needs of researchers (and in the future of field-builders and newcomers).
    We ended up asking explicitly about the kind of bottlenecks researchers dealt with, but also more general probes about how they were spending their time. Some interviewees articulated bottlenecks by themselves, while others didn't really see explicit bottlenecks but identified particularly costly activities in their research.
    In particular, four research activities were often highlighted as difficult and costly (here in order of decreasing frequency of mention):
    Running experiments
    Formalizing intuitions
    Unifying disparate insights into a coherent frame
    Proving theorems
    I don't know what your first reaction to this list is, but for us, it was something like: "Oh, none of these activities seems strictly speaking necessary in knowledge-production." Indeed, a quick look at history presents us with cases where each of those activities was bypassed:
    Einstein figured ...
    16 min
  • LW - Funding is All You Need: Getting into Grad School by Hacking the NSF GRFP Fellowship by hapanin
    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: Funding is All You Need: Getting into Grad School by Hacking the NSF GRFP Fellowship, published by hapanin on September 22, 2022 on LessWrong.
    Fellowship is All You Need: Getting into Grad School by Hacking the NSF GRFP
    Last year I started having delusions of grad school. By delusions I mean that on paper, in the numeric and text fields that actually make it into the application, I looked utterly unqualified to do machine learning research, let alone at a top school. I had no computer science publications, no masters degree, and a 3.3 GPA, in materials engineering. I was a self-taught washed up startup founder six years out of school with no advisor and three months work experience in my proposed field of NLP. Only two people had ever seen me write code; one was my current boss and the other was my ex-cofounder, who also happened to be my brother, so I couldn’t even scrape together a single letter of recommendation attesting to my programming talents. I did take a programming class once, CSE142: Intro to Computer Programming I, where I got a C. But next month I’ll be starting my PhD studies at Columbia advised by Zhou Yu and Luis Gravano.
    They’re really really good at what they do and there’s no school I’d rather be at. This is the story of how I leveraged the National Science Foundation Graduate Research Fellowship Program (NSF GRFP) to get my unqualified ass into grad school. The technical guide follows.
    Table of Contents
    The Hack
    Technical Guide
    Abstract
    Background
    Normal Route
    Funding is All You Need
    What Does the NSF Want?
    Methods
    How to Win a GRFP
    Take Risks
    The Hero's Journey
    Editing
    Timeline
    What to Do Once You've Won
    Results & Conclusion
    Footnotes
    Appendix 1: Useful Links
    The Hack
    In June of 2021 I forgot to register my +1 at a wedding and got placed next to a cool dude named Logan who told me just to shoot my shot on PhD applications even though I didn’t feel qualified. After all, they’re kind of a crap shoot anyway. Lacking any proper mentorship I took this advice at face value and spent the rest of the summer reading papers and the autumn drinking cappuccinos and banging out applications at Amy’s Merkato and Fort St. George after work. At the end of the year I fired off my freshly minted applications and swapped caffeine for alcohol while the responses rolled in.
    I was rejected from every school. Don’t get me wrong, I knew my application was weak, but that doesn’t make it suck less. All seven schools turned me down without even waitlisting me. There’s not much more to say on this except that plan A was kaput.
    Plan B, however, was still cooking. Promptness and agility are not virtues commonly ascribed to the Federal Government, so despite grad school fellowship deadlines (GRFP and NDSEG) closing before those of colleges, results are released afterward by several months. The GRFP is a three year fellowship for first or second year PhD students in STEM fields. Unlike most grants, the GRFP funds a person rather than a project. This fact carries several implications, the first of which being that you can sort of win a GRFP without a school or an advisor, in in mid-April I did just that.
    The GRFP is considered rather prestigious because it’s a bit difficult to get (~16% acceptance rate) and because it’s dripping with the world’s best lubricant: money. It’s not a master key to the ivory tower but it is a big shiny ⭐ NSF Sticker of Approval ⭐ that can get a busy prof to reply to a cold email four months past the official application deadline. The reason I had a shot at the GRFP but not at the traditional application process was due to the second implication of the GRFP’s person > project philosophy: doing cool stuff matters a lot, and I did a lot of cool stuff instead of going to class. I invented some polio eradication tech deployed to six countries, founded thr...
    21 min
  • LW - Fake qualities of mind by Kaj Sotala
    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: Fake qualities of mind, published by Kaj Sotala on September 22, 2022 on LessWrong.
    There’s a thing where you’d like to have one “quality of mind”, but it’s not available, but you substitute it with a kind of a fake or alternative version of the same. Which is fine as long as you realize you’re doing it, but becomes an issue if you forget that what’s happening.
    For example, you have a job that you’re sometimes naturally motivated to do and sometimes you totally don’t feel like it. On the days when you don’t feel motivated, you substitute the motivation with an act of just making yourself do it.
    Which of course makes sense: it’s hard to be motivated all the time, and if you need to work anyway, then you need to find some substitute.
    But what happens if you forget that you’re doing this, and forget what it actually feels like to be naturally motivated?
    Then you might find yourself doing the mental motion of “pushing yourself” all the time and wonder why it is that you keep struggling with motivation and why work feels so unenjoyable. You might think that the answer is to push yourself more, or to find more effective ways of pushing yourself.
    And then you might wonder why it is that even when you do manage to more successfully push yourself, you keep feeling depressed. After all, the pushing was a substitute for situations when you’re not enjoying yourself, but need to work anyway!
    But it might be that you constantly pushing yourself is a part of the problem. It’s hard to be naturally motivated if you don’t give yourself the time (or if your external circumstances don’t give you the time) to actually let that motivation emerge on its own.
    That’s not to say that just easing off on the pushing would necessarily be sufficient. Often there’s a reason for why the pushing became the default response; the original motivation was somehow blocked, and you need to somehow identify what’s keeping it blocked.
    It’s easiest to talk about this in the context of motivation. Most people probably have some sense of the difference between feeling naturally motivated and pushing yourself to do something. But in my experience, the same dynamic can emerge in a variety of contexts, such as:
    Trying to ‘do’ creative inspiration, vs. actually having inspiration
    Trying to ‘do’ empathy, vs. actually having empathy
    Trying to ‘do’ sexual arousal, vs. actually getting aroused
    Trying to quiet your feelings, vs. actually having self-compassion
    As well as more subtle mental motions that I have difficulty putting into exact words.
    The more general form of the thing seems to be something like. a part of the brain may sometimes be triggered and create an enjoyable and ‘useful’ state of mind. Typically these states of mind are more accessible if you’re feeling safe and not feeling stressed.
    When you are more stressed, or the original states are otherwise blocked off, another part of the mind observes that it would be useful to have that original state again. So it tries to somehow copy or substitute for it, but because it doesn’t have access to the systems that would actually trigger that state, it ends up with an imperfect substitute that only somewhat resembles the original one.
    What needs to happen next depends on the exact situation, but the first step is to notice that this is happening, and that “keep doing the thing but harder” isn’t necessarily the solution.
    My friend Annie comments:
    The easiest way for me to identify when I’m doing this is if there start to be phrases / mantras / affirmations that frequently pop into my head uninvited, and it’s the exact same phrase each time. Used to happen all the time at my stressful marketing job.
    It’s as if one part of my brain is trying to push the rest of my brain to be the kind of person who would naturally think/say that, but because I think in concepts by...
    4 min
  • EA - Why Wasting EA Money is Bad by Jordan Arel
    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 Wasting EA Money is Bad, published by Jordan Arel on September 22, 2022 on The Effective Altruism Forum.
    The thought crossed my mind today, “should I take the BART or Uber to the airport on the way to EAG DC..?” Among other considerations, I thought “well the BART would be much cheaper, but EA will compensate me for the Uber, so maybe cost shouldn’t be much of a consideration.” After thinking this, I thought “wow, what a sketchy line of logic.” Yet I don’t think this way of thinking is entirely uncommon among EAs.
    Shortly after this I came across this article about how EA Berkeley is wasting money in the EA UC Berkeley Slack channel. While I found the article a little bit confused and it seems to have some factual errors, and some of the claims were made somewhat less credible by the fact that the author then proceeded to post some somewhat aggressive comments toward people in the slack, I nonetheless find the criticism that EAs waste money to be alarming and valid and think it is important to address before the issue balloons out of hand.
    Basically, I think this argument has a few levels.
    On the first level, you could say that money is really valuable and since we can say that something like $200 (please correct me if this number is inaccurate) could save a year of someone’s life via GiveWell top charities, we should take this as a real consideration and have a very high bar for wasting money.
    Against that you could argue that, well, we have an insane amount of money for the size of the movement, if we very roughly have something like $50 billion and 2000 highly engaged EAs which have both been relatively stable over the past few years, if all of that money was spent by current EAs in our lifetime of ~50 years that’s about $500,000 per person, PER YEAR. That’s a lot. So even if it makes me only a minuscule amount more efficient, if the work I’m doing is high value enough in contributing to the community, then maybe it’s worth it.
    But then that only makes sense if the work I’m doing is extremely extremely valuable, because I still have to compare it against the bar of $200 equals ~1 year of life saved. So if a $50 Uber ride saves me half an hour, my half an hour must be more valuable than a three months of someone else’s life. That’s a pretty big claim.
    But, then, the claims of longtemism are quite big indeed. Bostrom calculates that a one second delay in colonizing space may be equivalent to something like the loss of 100 trillion human lives, due to galaxies we could potentially colonize moving away from us in every direction at fast speeds. Working on existential risk reduction, rather than speeding up technological progress and space colonization, likely increases this expected value by several orders of magnitude.. So if I am one of the very small number of people who is most obsessed with these ideas and competent/privileged enough to make a difference, and in expectation it seems that people explicitly working to reduce existential risk are most likely to succeed at doing so, then yes maybe saving half an hour of my time may actually have, in expectation, an un-intuitively massive positive impact.
    But then what about the article above and other criticisms? Couldn’t the reputation risk to EA from this way of thinking be very dangerous, both because it attracts people who want to mooch money off of the community, and repels potential collaborators who don’t want to be seen as wasteful?
    Yes, maybe it does repel certain people, but then again, perhaps it attracts the type of people who understand and agree with our logic, and if our logic is in fact correct and good, then perhaps the type of people who really look at our ideas and actions and evaluate them carefully, and then decide they agree, are exactly the type of people we are trying to attract. Perhaps w...
    9 min