The Nonlinear Library

The Nonlinear Library

By The Nonlinear FundEducation
Download on the App Store

All The Nonlinear Library episodes

  • EA - Meta Directory of April Fool’s Ideas by Yonatan Cale
    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: Meta Directory of April Fool’s Ideas, published by Yonatan Cale on April 1, 2023 on The Effective Altruism Forum.
    Intro
    As Effective Altruism grows, we face new challenges as a movement. Funding, cause prioritization, but the most neglected area is definitely just having fun.
    Potential impact
    # If one EA reads one post which makes them happy, how much more productive do they become, in percent, for that day? INCREASED_PERSON_PRODUCTIVITY_PER_HAPPY_FUNNY_POST_FOR_ONE_DAY = sq.to(-0.05, 0.10)
    FRACTION_OF_EA_THAT_READS_A_GOOD_POST = sq.to(0.05, 0.4)
    # This calculation is so intuitive that even copilot could autocomplete it
    IMPACT_PER_APRIL_FOOL_POST = \ EA_YEARLY_IMPACT_IN_QALYS \ INCREASED_PERSON_PRODUCTIVITY_PER_HAPPY_FUNNY_POST_FOR_ONE_DAY \ FRACTION_OF_EA_THAT_READS_A_GOOD_POST
    bins = 100 samples = IMPACT_PER_APRIL_FOOL_POST @ bins plt.hist(samples, bins=bins)
    As you can see, the spike around 0.1 indicates (assuming I'm reading this chart incorrectly) that the potential impact is relatively very high, compared to some other smaller spikes.
    A marketplace of ideas
    Some people have ideas for April Fool’s posts, and some people are good at writing but don’t have ideas.
    Let’s make these high impact connections between EA cofounders!
    Implementation
    The ideal platform will be
    A list of ideas
    Each idea has all the essential fields such as “title”, “long description”, “target audience”, “expected impact of that audience”, “squiggle model estimating the impact of the post given the post's quality”, “what is the joke you told that you’re most proud of (we use this as applicant background, but don't worry, we won't judge you)” and probably many more fields, the more the better, and we would like more ideas for fields below as this seems crucial for the project’s success.
    Of course, we will ask people not to spend no more than 1 hour filling out this form. (some people will be randomly assigned to the "45 minutes only" group, and we'll check which group performs better)
    We will send out a recursive feedback form to each applicant.
    The platform will support for comments, such as “I’d fund this post”, or “looking for a co-author”
    Emoji inclusiveness: If Zoomers use , Millennials will see , which is important to avoid miscommunication.
    User chat
    Profiles
    Sub groups by topic
    A feed that users can scroll
    Algorithms to optimize that feed based on the user’s preferences (open sourced, inspired by Twitter)
    A blue background
    But as an MVP, we’ll just use the comments to this post.
    Creating a critical mass
    This project will only succeed if all the following things will happen:
    The forum will change its font to Comic Sans (this step was easier than expected)
    Readers like you (yes you) will share their ideas for April Fool's posts here in the comments
    Thanks!
    This post was written quickly since we used the top EA tools for getting things done (having a due date), so any suggestions are welcome.
    Thanks Edo Arad for helping write this post. I'm sure he'd want to tell you that all mistakes are his. Edo could not immediately comment on this line I just added.
    Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.
    4 min
  • EA - Honestly I’m just here for the drama by electroswing
    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: Honestly I’m just here for the drama, published by electroswing on April 1, 2023 on The Effective Altruism Forum.
    I'm a young twenty-something who got into EA a couple of years ago. Back then I was really into the whole "learning object-level facts about EA” thing. I'd religiously listen to the 80,000 Hours podcast, read in detail about different interventions, both neartermist and longtermist, and generally just try my darndest to improve my understanding of how to do good effectively. Key to this voracious consumption was of course the EA Forum.
    Now? Lmao.
    It started with the SBF saga. Boy was the EA forum entertaining. The whole front page was full of upset people wildly speculating. So many rumors to sift through, so many spicy arguments to follow. The best parts of any post could be found by scrolling to the very bottom and unfolding highly downvoted comments. So much entertainment. Like reality TV except about my ingroup specifically.
    Then, you know the meme. EA has had a scandal of the week TM ever since. Castle? Hilarious watching people who don't understand logistics butt heads with people who don't value optics and frugality. The weird Tegmark neonazi thing? Absolutely incredible watching the comments turn on him and then pull back. Time and Bloomberg articles and the ensuing "I can fix EA in one blog post" follow-ups? Delicious. Bostrom and the fascinating case of the use-mention distinction? Yikes bro. Spicy takes and arguments hidden in the Lightcone closure announcement? Fantastic sequel to "The Vultures Are Coming".
    When it was announced the Community posts would appear in a separate section of the Forum, the little drama-hungry goblin in my brain was at first disappointed. Oh no! Maybe I'll accidentally click on a post about malaria instead! Then I realized I can simply upweigh Community posts by +100 and I'll never miss another scandal ever again.
    Now, I visit almost daily. I briefly skim the frontpage post titles. Maybe occasionally I'll stop to learn more about some new org or read the executive summary of a high-quality research report. But honestly? Most of the time I just scroll through looking for drama, and if I don't find it, I close the tab and get on with my day.
    Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.
    3 min
  • EA - The EA Relationship Escalator by ProbablyGoodCouple
    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 EA Relationship Escalator, published by ProbablyGoodCouple on April 1, 2023 on The Effective Altruism Forum.
    When dating, you may typically encounter the relationship escalator, which is a social script for how relationships are supposed to unfold that is oriented around default societal expectations. Potential partners are supposed to follow a progressive set of steps and achieve specific visible milestones towards a clear goal.
    This escalator frequently looks like:
    Meet on a dating app
    Go on a few dates
    Hold hands, kiss
    Become romantically exclusive
    Fall in love
    Meet the parents
    Have a long weekend together
    Vacation together
    Move in together
    Get engaged
    Get married
    Buy a house
    Have kids
    Have a dog
    Of course the steps are approximate and may not unfold literally in this order, but it should be pretty close. Moreover, where you are on this escalator and how long it has been since the previous step is often used to judge whether a relationship is sufficiently significant, serious, good, healthy, committed, etc. and to tell whether the relationship is worth pursuing.
    Effective altruists, as a community though, are rarely mainstream like this. This suggests that it may be helpful for EAs looking to date other EAs to have a very different and more customized social script to follow to judge how their unique EA relationship is unfolding.
    I recommend the EA relationship escalator look like this:
    Meet at EA Global but definitely do not flirt
    Re-meet at a house party where flirting is allowed
    Comment on their EA Forum posts
    Reach out on their “Date Me” doc
    Go on a few dates
    Become awkwardly personally and professionally intertwined
    Thoroughly assess together the conflicts of interest inherent in your relationship
    Talk to your HR department
    Talk to their HR department
    Talk to the CEA Community Health team
    Make a spreadsheet together to thoroughly quantify the relevant risks and benefits of your romantic relationship
    Decide to go for it
    Finally hold hands, kiss
    Synchronize your pomodoro schedules together
    Create a shared Complice coworking room / Cuckoo coworking room for just you two
    Take the same personality tests and quantify each other
    Introduce them to your polycule, hope they get along
    Fall in love
    Move in to the same EA group house
    Synchronize your GWWC pledge donation schedules
    Have your relationship details exposed by a burner account on the EA Forum
    Have the EA Forum moderator team encode your relationship details in rot13
    Meet the parents
    Vacation together, but exclusively touring various EA hubs
    Decide to fight the same x-risk
    Raise free range chickens together, start an animal sanctuary
    Create a Manifold market around whether or not you will get married
    Get engaged
    Get married
    Maximize utility
    Hopefully this EA Relationship Escalator helps give young EAs a social script to be able to find love with each other and be able to understand where their relationship currently is at.
    After all, you may spend 80,000 hours of your life in a good marriage so it’s important to get this right!
    Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.
    4 min
  • EA - Cause Exploration: Sending Billionaires to Space by Rasool
    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: Cause Exploration: Sending Billionaires to Space, published by Rasool on April 1, 2023 on The Effective Altruism Forum.
    TL:DR: +➡️️=➡️‍⚕️
    Today, I present to you a groundbreaking proposal with the potential to create a paradigm shift in philanthropy and global stewardship. The mission: to send 10 of the world's most affluent individuals into space to experience the "overview effect," a powerful cognitive transformation that could inspire them to become agents of change for our planet and its inhabitants.
    The "overview effect" is a phenomenon experienced by astronauts during spaceflight, often while observing Earth from orbit. This profound sense of interconnectedness and the realisation of our planet's vulnerability have been known to inspire a heightened sense of responsibility towards preserving and protecting Earth.
    Financial Implications and Potential Returns
    A commercial crew launch to low Earth orbit using SpaceX's Crew Dragon costs approximately $55 million per seat. Thus, sending 10 billionaires into space would require an investment of $550 million.
    Assuming the "overview effect" encourages these individuals to increase their annual philanthropic efforts by a modest 1%, the impact could be staggering. With a combined net worth of roughly $1 trillion, a 1% increase in annual giving would result in an additional $10 billion in charitable contributions each year, an 18-fold ROI.
    Importance, Tractability, and Neglectedness
    Importance: This proposal aims to foster a significant shift in the philanthropic mindset of some of the world's wealthiest individuals by leveraging the "overview effect." This cognitive transformation could result in increased charitable contributions towards global challenges such as climate change, global health and development, and extinction risks.
    Tractability: The proposal is feasible given the current advancements in commercial space travel, such as SpaceX's Crew Dragon. Although the initial investment is substantial, the potential for a substantial return on investment through increased philanthropy provides a strong incentive to pursue the project.
    Neglectedness: While the concept of the "overview effect" is known within the space community, its potential impact on the philanthropic behaviour of billionaires has not been widely explored. This proposal brings forward a unique and innovative approach to encouraging large-scale philanthropy, addressing a relatively neglected aspect of global problem-solving.
    Conclusion
    The proposal to send billionaires into space to experience the "overview effect" is firmly grounded in both scientific understanding and financial analysis. The initial investment of $550 million carries the potential to generate a substantial increase in philanthropic giving, which could dramatically impact our world for the better.
    55,000,000 x 10
    Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.
    3 min
  • LW - Proposal: Butt bumps as a default for physical greetings by Adam Zerner
    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: Proposal: Butt bumps as a default for physical greetings, published by Adam Zerner on April 1, 2023 on LessWrong.
    As Tim Urban explains in The Great Perils of Social Interaction, greetings are hard.[1]
    I’ll be 90 and I still won’t have figured this out. There are different rules for everyone and nothing’s clear—Do I shake my grandfather’s hand or go for the hug? How about my friend’s father? Old friend? New friend? Opposite-sex acquaintance? Longtime work colleague? Sibling’s good friend who I’m meeting for the second time? It’s unbelievably complicated.
    And there aren’t just two options you’re choosing from—there’s the high school bro handshake/backslap douche possibility, there’s the vertical, loose-hand high-five that morphs into a weird springy-finger tension thing as you snap away, there’s even the easy but taking-yourself-really-seriously non-ironic fist pound. And even if you both go for the hug, there’s a question of duration and firmness and who’s in charge of those decisions.
    Frankly, I'm embarassed that we still haven't figured this one out. We're rationalists, after all. We understand coordination. Aren't we supposed to WIN?
    Yes. Yes we are. I refuse to consider us The Kind That Can't Cooperate. I refuse to lose to barbarians.
    That's why, today, I am taking action. Instead of sitting on my hands, I shall stand up and do something about this problem.
    I propse that, as a default, universal greeting, we adopt the butt bump.
    Why? We could instead go with something like "the elbow"...
    ...but, well, that's just a little boring. Butt bumps feel like a more fun theoretically sound solution due to a variety of properties they possess.
    In everyday situations, they are quick and easy to perform.[2]
    If, however, you want to add some flair to it, you can include a namaste.
    If you are celebrating an achievement, you can jump, aiming to bump butts in the air.
    If you are in a good mood and have your party pants on, you can get a little jiggy.
    And of course, if there is ever another pandemic, we'll be well prepared.
    I also predict that, being an unusual form of greeting, they would enhance feelings of ingroup connection. Think about people who have secret handshakes.
    Imagine yourself showing up to a meetup for the first time. You see one other person standing there and approach them to say hi.
    Eye contact is made. The physical distance between you and them is closing.
    You slowly start to swivel your hips, hoping desperately that the other person reciprocates.
    Imagine the sense of belonging you'll feel when they do.
    As an actionable next step, assurance contracts seem like the right tool for the job. Please reply with your username to this comment if you agree to the following vow if 100 others agree, and this comment if you agree if 1,000 others agree. If you agree to the former it is implied that you agree to the later.
    I {username} vow to, for the next three months, when greeting a person for whom I am reasonably confident to be a rationalist, swivel my hips clockwise in an attempt to initiate a butt bump. If they reciprocate, I will proceed to perform the butt bump. If they do not reciprocate within three seconds, I will unswivel my hips and proceed as I desire.
    Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.
    4 min
  • EA - Announcing drama curfews on the Forum by Lorenzo Buonanno
    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: Announcing drama curfews on the Forum, published by Lorenzo Buonanno on April 1, 2023 on The Effective Altruism Forum.
    The moderation team is small, limited in capacity, and quite tired. We often end up having to suddenly work at odd hours, and this is bad for our productivity and ability to focus on our other (usually main) projects.
    So we’re introducing a schedule: Forum users are henceforth not allowed to come even close to violating the Forum norms — or significantly increase the probability that others violate Forum norms by posting drama-prone content — on certain days and hours. Moderation action during the curfew will be significantly more tyrannical.
    Here’s the schedule (“drama” here is used as a shorthand for “posting or potentially causing norm-violating or close-to-the-line discussions”):
    Mondays-Fridays: 10 a.m. GMT - 10 p.m. GMT is ok. Besides those times, though, no drama.
    Saturdays: No drama at all.
    Sundays: You can post drama from 1 p.m. to 8 p.m. GMT.
    Please also respect the following holidays:
    All possible versions of New Year’s Eve and New Year’s Day
    Top 5 holidays in any of the top 10 most popular religions
    Smallpox Eradication Day
    Petrov Day
    Valentine’s Day
    (You can find a calendar that lists these here.)
    Thank you all in advance!
    Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.
    2 min
  • EA - Meetup: EA Burner Accounts Anonymous by BurnerMeetupBurner
    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: Meetup: EA Burner Accounts Anonymous, published by BurnerMeetupBurner on April 1, 2023 on The Effective Altruism Forum.
    Do you think the amount of burner accounts is too high? What if you had a chance to directly talk to some of EA’s most prominent burners and bring your concerns about their concerns directly to them?
    Various burner accounts are getting together to host a meetup.
    Have you ever wanted to meet the face behind Whistleblower3, Whistleblower9, Whistleblower10, or Whistleblower67? What about BurnerAcct or anotherEAonaburner? Want to meet all the concerned EAs behind ConcernedEAs?
    Activities will include:
    Creating a taxonomy of burner accounts
    More burner explainers from BurnerExplainer
    Finally critiquing everyone without fear
    Finally reveal the identity behind who killed JFK
    Investigative journalism: Is Bostrom behind BostromAnonAccount?
    All communication will be done exclusively in the Navajo language, encrypted via ADFGVX, then through the Enigma machine, and finally via AES-256. The communication will then be sent via smoke signals that are sent via Signal.
    This meetup will take place on April 1 in a remote undisclosed location for security purposes.
    Plus we will finally reveal the shocking true identity of one of the most prolific burner accounts, Nathan Young!
    Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.
    2 min
  • EA - April's Fools Posts Can Be Short by EdoArad
    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: April's Fools Posts Can Be Short, published by EdoArad on April 1, 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 - New 80,000 Hours feature: Listen to audio versions of our podcast transcripts by Bella
    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: New 80,000 Hours feature: Listen to audio versions of our podcast transcripts, published by Bella on April 1, 2023 on The Effective Altruism Forum.
    I’m delighted to announce the launch of a new feature on the 80,000 Hours website: as of today, you can now listen to state-of-the-art text-to-speech audio versions of all of our podcast transcripts!
    We're including the entire 146-episode back catalogue at launch — that's over 300 hours of listening material for our audience to enjoy whenever they choose.
    We hope that this new feature can bring our podcast transcripts to an all new audience of audiophiles and those who prefer listening to content rather than reading it.
    And since it also works on mobile, you can now listen to our podcast transcripts while doing the dishes, walking the dog, or on your daily commute!
    I'm really excited about this inititative making our podcast transcripts more widely accessible, and helping spread important ideas about the world's most pressing problems and what you can do to solve them.
    Feel free to ask any questions about our new feature in the comments.
    Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.
    2 min
  • LW - Definitive confirmation of shard theory by TurnTrout
    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: Definitive confirmation of shard theory, published by TurnTrout on April 1, 2023 on LessWrong.
    I've written a lot about shard theory over the last year. I've poured dozens of hours into theorycrafting, communication, and LessWrong comment threads. I pored over theoretical alignment concerns with exquisite care and worry. I even read a few things that weren't blog posts on LessWrong. In other words, I went all out.
    Last month, I was downloading gpt-neox-20b when I noticed the following:
    I've concluded the following:
    Shard theory is correct (at least, for gpt-neox-20b).
    I wasted a lot of time arguing on LessWrong when I could have just touched grass and ran the cheap experiment of actually downloading a model and observing what happens.
    I totally underappreciated how much work the broader AI community has done on alignment. They are literally able to print out how many shards an "uninterpretable" model like gpt-neox-20b has.
    They don't whine about how tough these models are to interpret. They just got the damn job done.
    This really makes team shard's recent work look naive and incomplete. I think we found subcircuits of a cheese-shard. The ML community just found all 46 shards in gpt-neox-20b, and nonchalantly printed that number for us to read.
    This is embarrassing for me, even though shard theory just got confirmed. I seriously spent months arguing about this stuff? Cringe.
    I did a bad job of interpreting theoretical arguments, given that I was still uncertain after months of thought. Probably this is hard and we are bad at it.
    Stop arguing in huge comment threads about what highly speculative approaches are going to work, and start making your beliefs pay rent by finding ways to experimentally test them.
    See how well you can predict AI alignment results today before trying to predict what happens for 500IQ GPT-5.
    I don't ever want to hear another "but I won't be able to learn until the intelligence explosion happens or not." Correct theories leave many clues and hints, and some of those are going to be findable today. If even a navel-gazing edifice like shard theory can get instantly experimentally confirmed, so can your theory.
    I then had GPT-4 identify some of the 46 shards.
    The Procrastination Shard: This shard makes the AI model continually suggest that users read just one more LessWrong post or engage in another lengthy comment thread before taking any real-world action.
    Hindsight Bias Shard: This shard leads the model to believe it knew the answer all along after learning new information, making it appear much smarter in retrospect.
    The "Armchair Philosopher" Shard: With this shard, the AI can generate lengthy and convoluted philosophical arguments on any topic, often without any direct experience or understanding of the subject matter.
    Argumentative Shard: A shard that compels the model to challenge every statement and belief, regardless of how trivial or uncontroversial it may be.
    Trolley Problem Shard: This shard makes the model obsess over hypothetical moral dilemmas and ethical conundrums, while ignoring more practical and immediate concerns.
    The Existential Dread Shard: This shard causes the AI to frequently bring up the potential risks of AI alignment and existential catastrophes, even when discussing unrelated topics.
    I'm pleasantly surprised that gpt-neox-20b shares values which I (and other LessWrong users) have historically demonstrated. The fact that gpt-neox-20b and LessWrong share many shards makes me more optimistic about alignment overall, since it implies convergence in the values of real-world trained systems.
    (As a corollary, this demonstrates that LLMs can make serious alignment progress.)
    Although I'm embarrassed and humbled by this experience, I'm glad that shard theory is true. Here's to—next time—only taking three months before running ...
    5 min