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: Things usually end slowly, published by OllieBase on June 7, 2022 on The Effective Altruism Forum.

Epistemic status: Low confidence. These findings updated me slightly towards worrying more about slow declines vs. sudden collapse/extinction. Most sources are things I quickly googled so are probably off the mark.

TL;DR It’s sometimes claimed or suggested that if an extinction event were to happen, it would happen very quickly (i.e. days, weeks, months). However, when empires fall or species go extinct, they usually do so very slowly (years, decades, millennia). These base rates suggest we should have a reasonably strong prior on slow decline when thinking about civilisational collapse and, at more of a stretch, existential risks. This sketch is very rough, and I’d love to see someone work out these base rates, and how they might be changing, more carefully.

Rapid extinction

Thinkers in and around the EA community sometimes claim that x-risk scenarios will cause extinction rapidly. For example:

Eliezer Yudkowsky argues that the likely scenario with AGI takeoff is “growth stays at 5%/year and then everybody falls over dead in 3 seconds and the world gets transformed into paperclips”

Toby Ord, in The Precipice (p.28), writes that “As technology continues to advance, new threats appear on the horizon. These threats promise to be more like nuclear weapons than like climate change: resulting from sudden breakthroughs, precipitous catastrophes”

Relatedly, the metaphor of a precipice implies a very steep descent from stability to extinction.

Perhaps it’s right to expect that an existential catastrophe will be rapid. Your inside view might well take into account things happening awfully fast (such as rapidly increasing compute efficiency or pathogens spreading). But holding this view should require a very strong update from the prior that extinction events and civilisational collapses usually take a very long time.

Data from civilisations

Two caveats:

I expect the way I’m most likely to be wrong about all this is that historical extinctions and civilisational collapses are not appropriate reference classes for x-risks. I think they’re very useful because they’re a rough approximation of x-risks from actual data and I’m sceptical of claims along the lines of “the thing I’m worried about is totally unique ”, but I could be wrong.

While we often refer to the “fall” of Rome and the “collapse” of the Mayans, it’s rarely appropriate to consider human empires and civilisations extinct. In this post, I usually use years when the empire or civilisation is considered by scholars to have “ended.” The specific year, or whether they “ended” at all, is usually up for debate.

To form a rough prior on how long things (civilisations, empires, species) take to decline, I scanned some encyclopedic pages to find:

The start of the end: the last period in time when the civilisation was still ‘flourishing’, usually just before the start of a collapse or invasion.

The earliest plausible date and the latest plausible date which could constitute the “end” of the civilization, to form a kind of confidence interval.

This sheet lists some empires, their last points of flourishing, their earliest reasonable dissolution points and their latest reasonable dissolution points. From this, I calculated a lower guess on decline length and an upper guess on decline length (in years). You can see this summary graph below. I expect anyone could review these dates and find points of disagreement, even by reviewing the same sources I did. Furthermore, I’d love to see someone expand this list (perhaps using this useful list compiled by Luke Kemp).

You don’t need to read this data too carefully to notice that rapid dissolution is clearly the exception to the rule.

One interesting thing to note, however, is that declines...

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: [Link] Luke Muehlhauser: Effective Altruism As I See It, published by Aryeh Englander on June 7, 2022 on The Effective Altruism Forum.

This is a link post for/

Very good article IMHO. I expect to link to this in the future when introducing people to EA.

Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.

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: Deference Culture in EA, published by Joey on June 7, 2022 on The Effective Altruism Forum.

As has been noted by other writers, the EA (Effective Altruism) movement has a pretty strong deference culture. In many ways this makes sense as lots of EAs come from a background of reading about and being compelled by organizations like GiveWell. These organizations are built on the assertion that charity evaluation is hard and benefits from a full-time team doing it. A typical donor has the most impact by deferring to this research. This culture of deference has become pretty strong in EA, I pretty frequently have conversations with highly involved EAs who are still deferring on major topics (e.g. cause area, career choice) without investigating them personally.

On the one hand, it is impossible to be an expert in everything, and making hard decisions around doing good is no different. On the other hand, a culture that is too high in deference or defers using the wrong metrics can become homogenized and lead to great opportunities being missed. Often, pieces written about deferral end up falling on the “you should defer more” side or the “you should defer less” side.

I think the optimal perspective on the right amount of deferral depends on background expertise and expected time. (I don’t think any previous writers would disagree with this, but their posts are typically not interpreted in that way.) If you want to become a tinker and put thousands of hours into learning about cars; this puts you in a different position than someone who drives but has never looked under the hood, in terms of deference to a mechanic. The same piece of advice; ‘defer less’ or ‘defer more’, would not be equally applicable to those two demographics. One might defer too little (e.g. the person who has never opened up their hood being pretty confident they can fix something) another might defer too much (the tinker being disappointed to learn that the local mechanic knows far less than expected about a rare engine part).

Another component of this is discerning which questions are more objectively answerable vs ones which are based on values or unclear epistemic trade-offs. To use GiveWell as an example; if you want to save the most lives possible with a high degree of confidence, one of their top choices in fighting malaria is a really strong bet, and deferring to their research is advisable. However, they are far less confident about their trade-offs between income and lives saved and thus it makes less sense to defer on that topic.

So when does it concretely make sense to defer in EA? Let's examine some clear examples on either side and then work our way to more ambiguous cases.

High deference - new EA

John is brand new to EA and has read a single book on the topic. Although he loves the concepts, he feels overwhelmed with all the new information and does not plan on engaging with it super deeply. He is already well into a solid career and does not imagine EA becoming a big part of his life. Nonetheless, he wants his donations to make the maximum impact from a fairly standard view of saving more lives and reducing pain. He defers to the EA community and ends up donating 10% to GiveWell recommended charities, seeing it as a safe, impactful option that does not take a ton of time.

I think in this case John has made the ideal call; he has made an optimal decision given the amount of time and energy he wants to put into the topic. But let's look at the same amount of deference with a much more involved EA.

High deference - Experienced EA

Sally has been involved in the EA movement for a number of years, she led her local university chapter for a couple of years before joining an EA organization full time. She has spent several hundred hours engaging with EA content, and has a pretty deep understandin...

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 the Legal Priorities Project Writing Competition: Improving Cost-Benefit Analysis to Account for Existential and Catastrophic Risks, published by Mackenzie on June 7, 2022 on The Effective Altruism Forum.

Summary

We’re announcing a writing competition to provide practical guidance to the US federal government on how to incorporate existential and catastrophic risks into agency cost-benefit analysis.

We created this competition because we believe

This guidance is particularly important to mitigating existential and catastrophic risks and improving agency decision-making,

We are entering a unique window of opportunity to inform and shape how the US conducts cost-benefit analysis, and

In order to do that well, we need to attract and direct the attention of relevant experts.

In total, we plan to distribute $72,500 in prize money for up to 10 prizes.

Submissions can be up to 10,000 words (though we encourage shorter submissions where the author can present their argument in fewer words). Submissions will include (a) a summary of the author’s proposal, (b) amended or proposed language to the regulations governing cost-benefit analysis in the US (with particular attention to Circular A-4), and (c) an essay explaining, justifying, and qualifying the proposed language and approach.

The deadline is July 31, 2022. You can find submission instructions and further information about the competition below and on our website.

The remainder of this post outlines why we are excited about this topic and approach, what the contest will entail, how to apply, and ways you can help even if you aren’t in a position to submit a proposal.

Importantly, we are offering $1,000 to anyone who successfully refers an applicant who becomes a finalist in the competition. You can find more information on the referral bonus here.

Prompt

This competition aims to produce practical guidance on how US federal agencies can more fully account for the interests of future generations and catastrophic and existential risks in their decision-making processes. We seek submissions that present and develop regulatory methods to:

Account for changes in the probability of national or human extinction (posed by threats like pandemics, climate change, nuclear weapons, or artificial intelligence)

Account for changes in the probability of catastrophic public harms

Submissions must contain the following elements:

A single cover page summarizing the author’s argument

A second page that contains the language of your proposed amendment(s) to Circular A-4 or other regulations governing regulatory decision-making in the United States [Note: because the topic has already been highly theorized and debated, we discourage applicants from proposing or arguing for changes to the general discount rate]

An essay explaining, justifying, and qualifying the proposed amendments and methodologies used. This essay must be at most 10,000 words, including citations and footnotes. This is an upper bound, and we encourage shorter submissions where the author can present their argument in fewer words. This word count does not include the cover page or amended regulatory text. Citations need not follow any particular format so long as they are consistent and allow for verification of the claims made.

How to apply

All submissions shall be made through this form, in .doc or .docx format. Instructions on how to submit your entry can be found at that link. Because the review process will be anonymous, please do not include identifying information in your document.

Prizes

First Prize: $30,000

Second Prizes (up to 4): $5,000 each

Third Prizes (up to 5): $2,500 each

An additional Incentive Prize of $10,000 is available to any Finalists that (a) are cited by, or (b) whose suggested language is substantially adopted by the revised Circular A-4. ...

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: AGI Safety FAQ / all-dumb-questions-allowed thread, published by Aryeh Englander on June 7, 2022 on LessWrong.

While reading Eliezer's recent AGI Ruin post, I noticed that while I had several points I wanted to ask about, I was reluctant to actually ask them for a number of reasons:

I have a very conflict-avoidant personality and I don't want to risk Eliezer or someone else yelling at me;

I get easily intimidated by people with strong personalities, and Eliezer... well, he can be intimidating;

I don't want to appear dumb or uninformed (even if I am in fact relatively uninformed, hence me wanting to ask the question!);

I feel like there's an expectation that I would need to do a lot of due diligence before writing any sort of question, and I don't have the time or energy at the moment to do that due diligence.

So, since I'm probably not the only one who feels intimidated about asking these kinds of questions, I am putting up this thread as a safe space for people to ask all the possibly-dumb questions that may have been bothering them about the whole AGI safety discussion, but which until now they've been too intimidated, embarrassed, or time-limited to ask.

I'm also hoping that this thread can serve as a FAQ on the topic of AGI safety. As such, it would be great to add in questions that you've seen other people ask, even if you think those questions have been adequately answered elsewhere. [Notice that you now have an added way to avoid feeling embarrassed by asking a dumb question: For all anybody knows, it's entirely possible that you are literally asking for someone else! And yes, this was part of my motivation for suggesting the FAQ style in the first place.]

Guidelines for questioners:

No extensive previous knowledge of AGI safety is required. If you've been hanging around LessWrong for even a short amount of time then you probably already know enough about the topic to meet any absolute-bare-minimum previous knowledge requirements I might have suggested. I will include a subthread or two asking for basic reading recommendations, but these are not required reading before asking a question. Even extremely basic questions are allowed!

Similarly, you do not need to do any due diligence to try to find the answer yourself before asking the question.

Also feel free to ask questions that you're pretty sure you know the answer to yourself, but where you'd like to hear how others would answer the question.

Please separate different questions into individual comments, although if you have a set of closely related questions that you want to ask all together that's fine.

As this is also intended to double as a FAQ, you are encouraged to ask questions that you've heard other people ask, even if you yourself think there's an easy answer or that the question is misguided in some way. You do not need to mention as part of the question that you think it's misguided, and in fact I would encourage you not to write this so as to keep more closely to the FAQ style.

If you have your own (full or partial) response to your own question, it would probably be best to put that response as a reply to your original question rather than including it in the question itself. Again, I think this will help keep more closely to an FAQ style.

Keep the tone of questions respectful. For example, instead of, "I think AGI safety concerns are crazy fearmongering because XYZ", try reframing that as, "but what about XYZ?" Actually, I think questions of the form "but what about XYZ?" or "but why can't we just do ABC?" are particularly great for this post, because in my experience those are exactly the types of questions people often ask when they learn about AGI Safety concerns.

Follow-up questions have the same guidelines as above, so if someone answers your question but you're not sure you fully underst...

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: AGI Ruin: A List of Lethalities, published by EliezerYudkowsky on June 6, 2022 on The Effective Altruism Forum.

Preamble:

(If you're already familiar with all basics and don't want any preamble, skip ahead to Section B for technical difficulties of alignment proper.)

I have several times failed to write up a well-organized list of reasons why AGI will kill you. People come in with different ideas about why AGI would be survivable, and want to hear different obviously key points addressed first. Some fraction of those people are loudly upset with me if the obviously most important points aren't addressed immediately, and I address different points first instead.

Having failed to solve this problem in any good way, I now give up and solve it poorly with a poorly organized list of individual rants. I'm not particularly happy with this list; the alternative was publishing nothing, and publishing this seems marginally more dignified.

Three points about the general subject matter of discussion here, numbered so as not to conflict with the list of lethalities:

-3. I'm assuming you are already familiar with some basics, and already know what 'orthogonality' and 'instrumental convergence' are and why they're true. People occasionally claim to me that I need to stop fighting old wars here, because, those people claim to me, those wars have already been won within the important-according-to-them parts of the current audience. I suppose it's at least true that none of the current major EA funders seem to be visibly in denial about orthogonality or instrumental convergence as such; so, fine. If you don't know what 'orthogonality' or 'instrumental convergence' are, or don't see for yourself why they're true, you need a different introduction than this one.

-2. When I say that alignment is lethally difficult, I am not talking about ideal or perfect goals of 'provable' alignment, nor total alignment of superintelligences on exact human values, nor getting AIs to produce satisfactory arguments about moral dilemmas which sorta-reasonable humans disagree about, nor attaining an absolute certainty of an AI not killing everyone. When I say that alignment is difficult, I mean that in practice, using the techniques we actually have, "please don't disassemble literally everyone with probability roughly 1" is an overly large ask that we are not on course to get. So far as I'm concerned, if you can get a powerful AGI that carries out some pivotal superhuman engineering task, with a less than fifty percent change of killing more than one billion people, I'll take it.

Even smaller chances of killing even fewer people would be a nice luxury, but if you can get as incredibly far as "less than roughly certain to kill everybody", then you can probably get down to under a 5% chance with only slightly more effort. Practically all of the difficulty is in getting to "less than certainty of killing literally everyone". Trolley problems are not an interesting subproblem in all of this; if there are any survivors, you solved alignment. At this point, I no longer care how it works, I don't care how you got there, I am cause-agnostic about whatever methodology you used, all I am looking at is prospective results, all I want is that we have justifiable cause to believe of a pivotally useful AGI 'this will not kill literally everyone'. Anybody telling you I'm asking for stricter 'alignment' than this has failed at reading comprehension.

The big ask from AGI alignment, the basic challenge I am saying is too difficult, is to obtain by any strategy whatsoever a significant chance of there being any survivors.

-1. None of this is about anything being impossible in principle. The metaphor I usually use is that if a textbook from one hundred years in the future fell into our hands, containing all of the simple idea...

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: Grokking “Forecasting TAI with biological anchors”, published by Anson Ho on June 6, 2022 on The AI Alignment Forum.

Notes:

I give a visual explanation of Ajeya Cotra’s draft report, Forecasting TAI with biological anchors (Cotra, 2020), summarising the key assumptions, intuitions, and conclusions

The diagrams can be found here – you can click on some of the boxes to get linked to the part of the report that you’re interested in

Thanks to Michael Aird, Ashwin Acharya, and the Epoch team for suggestions and feedback! Special thanks to Jaime Sevilla and Ajeya Cotra for detailed feedback.

Executive Summary

Click here to skip the summary

Ajeya Cotra’s biological anchors framework attempts to forecast the development of Transformative AI (TAI) by treating compute as a key bottleneck to AI progress. This lets us focus on a concrete measure (compute, measured in FLOP) as a proxy for the question “when will TAI be developed?” Given this, we can decompose the question into two main questions:

2020 training compute requirements: How much compute will we need to train TAI, using 2020 Machine Learning architectures and algorithms?

Affordability of compute: How likely is it that we’ll be able to afford the compute required to train TAI in a particular year?

The second question can be tackled by turning to existing trends in three main factors: (1) algorithmic progress e.g. improved algorithmic efficiency, (2) decreasing computation prices e.g. due to hardware improvements, and (3) increased willingness to spend on compute.

The first question is significantly trickier. Cotra attempts to answer it by treating the brain as a “proof of concept” that the “amount of compute” used to “train” the brain can train a general intelligence. This lets her relate the question “how much compute will we need to train TAI?” with the question “how much ‘compute’ was used to ‘train’ the human brain?”. However, there’s no obvious single interpretation for the latter question, so Cotra comes up with six hypotheses for what this corresponds to, referring to these hypotheses as “biological anchors” or “bioanchors”:

Evolution anchor: Compute to train TAI = Compute performed over evolution since the first neurons

Lifetime anchor: Compute to train TAI = Compute performed by the human brain when maturing to an adult (0 to 32 years old)

Three neural network anchors: Anchor to the processing power of the human brain, and to empirical parameter scaling laws.

Technically there are three of these, corresponding to short, medium, and long “effective horizon lengths” – the amount of data required to determine whether or not a perturbation to the AI system improves or worsens performance

Genome anchor: Anchor to the processing power of the human brain, set the number of parameters = number of bytes in the human genome, and extrapolate the amount of training data required using the same empirical scaling laws mentioned above and assuming a long horizon length (one “data point” = multiple years)

In calculating the training compute requirements distribution, Cotra places 90% weight collectively across these bioanchor hypotheses, leaving 10% to account for the possibility that all of the anchors significantly underestimate the required compute.

Here’s a visual representation of how Cotra breaks down the question “How likely is the development is TAI by a given year?”:

The above was essentially a summary of Cotra’s factorization of the question of AI timelines; for a summary of her key findings, see here.

Motivation

One of the biggest unresolved debates in AI Safety is the question of AI Timelines – when will Transformative AI (TAI) be developed? In 2020, Ajeya Cotra released a draft report, Forecasting TAI with biological anchors (Cotra, 2020), that aims to answer this question. It’s over 200 pages long including the 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: Here are the finalists from FLI’s $100K Worldbuilding Contest, published by Jackson Wagner on June 6, 2022 on The Effective Altruism Forum.

Earlier this year, the Future of Life Institute announced a competition to design visions of a plausible, aspirational future where humanity coexists with artificial general intelligence. Out of ~150 submissions, twenty finalists have been chosen — you can read them all here!

The Future of Life institute would love if you read some finalists and submitted feedback about the scenarios. Please note that I, Jackson Wagner, am not affiliated with FLI — rather, I’m one of the contestants! (Please also note that the US-based FLI is apparently totally different from the Oxford-based “Future of Humanity Institute”.) Since I’m not affiliated with FLI, I’m free to play favorites among the different finalists, so here are my suggestions of the most interesting reads.

My Favorite Entries

World-313

This entry comes from an impressive seven-person team who have backgrounds in the study of AI governance and existential risk. In contrast to the wild sci-fi of other entries, this scenario has a kind of professional, sensible, “consensus optimistic baseline scenario” feeling. If you would like to read a polished future-history of AI as it might appear in a UN report or Kim Stanley Robinson novel, these people have got what you’re looking for.

World-281

Maybe you don’t want UN reports. Maybe you want a story where a superintelligent AI designs a symbiotic fungal organism to live inside human bodies and improve peoples’ health and intelligence, and where humanity hashes out the Long Reflection by building hive-mind superstructures orbiting the sun! World-281 is here for you.

World-88

Of course I think my own entry is the best and most deserving of attention!! Here is the summary I wrote for my scenario:

Look around at humanity’s fractious politics, at the essentially lawless competition of international relations and war, and at our fumbled response to familiar threats like viral pandemics. Our civilization isn’t exactly well-prepared to safely manage the deployment of unpredictable and incredibly powerful AI technology. But how do we get from where we are now, to where we want to go? In our utopian scenario, we imagine:1. How governments could use mechanisms like prediction markets and liquid democracy, along with AI-enhanced research/analysis tools, to collectively make wiser decisions and capably execute big projects.2. How positive changes could spread to many countries via competition to achieve faster economic growth than rivals, and via snowball effects whereby passing one reform makes it easier to pass others.3.

How the resulting, more-capable civilization could recognize the threat posed by misalignment of powerful AI systems, and internationally coordinate to create a solid plan for safely developing aligned superintelligent AI.

World-429

This entry brings a strong dose of Robin-Hanson-esque pessimism (and also realism) about human social motivations and the inadequacy of society’s leading institutions. Its summary:

AGI is kept at a distance from practically impacting the world while narrow AI remakes the world completely. Most people don’t know or care about the difference and have no idea how they could distinguish between a human or artificial stranger in the Metaverse.

It also features a beautifully-written short story and accompanying musical media piece, which explore what it might be like to experience superhumanly high-quality AI-generated art.

Other Finalists

There are lots of great art pieces and interesting ideas mixed in with the other finalists. World-262 features an entire four-page illustrated comic narrating their activist/leftist vision of a future utopia featuring government by sortition and AI that lets us communicate wi...

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 descriptive, not prescriptive, overview of current AI Alignment Research, published by Jan Hendrik Kirchner on June 6, 2022 on The AI Alignment Forum.

TL;DR: In this project, we collected and cataloged AI alignment research literature and analyzed the resulting dataset in an unbiased way to identify major research directions. We found that the field is growing quickly, with several subfields emerging in parallel. We looked at the subfields and identified the prominent researchers, recurring topics, and different modes of communication in each. Furthermore, we found that a classifier trained on AI alignment research articles can detect relevant articles that we did not originally include in the dataset.

Dataset Announcement

In the context of the 6th AISC, we collected a dataset of alignment research articles from a variety of different sources. This dataset is now available for download here and the code for reproducing the scrape is on GitHub here. When using the dataset, please cite our manuscript as described in the footnote.

Here follows an abbreviated version of the full manuscript, which contains additional analysis and discussion.

Rapid growth of AI Alignment research from 2012 to 2022 across two platforms

After collecting the dataset, we analyzed the two largest non-redundant sources of articles, Alignment Forum (AF) and arXiv. We found rapid growth in publications on the AF (Fig. 1a) and a long-tailed distribution of articles per researcher (Fig. 1b) and researchers per article (Fig. 1c). We were surprised to find a decrease in publications on the arXiv in recent years, but identified the cause for the decrease as spurious and fixed the issue in the published dataset (details in Fig. 4).

Unsupervised decomposition of AI Alignment research into distinct clusters

Given access to this unique dataset, we were curious to see if we could identify distinct clusters of research. We mapped the title + abstract of each article into vector form using the Allen Institute for AI's SPECTER model and reduced the dimensionality of the embedding with UMAP (Fig. 2a). The resulting manifold shows a continuum of AF posts and arXiv articles (Fig. 2b) and a temporal gradient from the top right to the bottom left (Fig. 2c). Using k-means and the elbow method, we obtain five clusters of research articles that map onto distinct regions of the UMAP projection (Fig. 2d).

We were curious to see if the five clusters identified by k-means map onto existing distinctions in the field. When identifying the most prolific authors in each cluster, we noticed strong differences (consistent with previous work that suggests that author identity is an important indicator of research direction).

By skimming articles in each cluster and given the typical research published by the authors, we suggest the following putative descriptions of each cluster:

cluster one: Agent alignment is concerned with the problem of aligning agentic systems, i.e. those where an AI performs actions in an environment and is typically trained via reinforcement learning.

cluster two: Alignment foundations research is concerned with deconfusion research, i.e. the task of establishing formal and robust conceptual foundations for current and future AI Alignment research.

cluster three: Tool alignment is concerned with the problem of aligning non-agentic (tool) systems, i.e. those where an AI transforms a given input into an output. The current, prototypical example of tool AIs is the "large language model".

cluster four: AI governance is concerned with how humanity can best navigate the transition to advanced AI systems. This includes focusing on the political, economic, military, governance, and ethical dimensions.

cluster five: Value alignment is concerned with understanding and extracting human preferences and designing me...

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: Transcript of Twitter Discussion on EA from June 2022, published by Zvi on June 6, 2022 on The Effective Altruism Forum.

(Poster's note: Given subject matter I am posting an additional copy here in the EA Forum. The theoretically canonical copy of this post is on my Substack and I also post to Wordpress and LessWrong.)

Recently on Twitter, in response to seeing a contest announcement asking for criticism of EA, I offered some criticism of that contest’s announcement.

That sparked a bunch of discussion about central concepts in Effective Altruism. Those discussions ended up including Dustin Moskovitz, who showed an excellent willingness to engage and make clear how his models worked. The whole thing seems valuable enough to preserve in a form that one can navigate, hence this post.

This compiles what I consider the most important and interesting parts of that discussion into post form, so it can be more easily seen and referenced, including in the medium-to-long term.

There are a lot of offshoots and threads involved, so I’m using some editorial discretion to organize and filter.

To create as even-handed and useful a resource as possible, I am intentionally not going to interject commentary into the conversation here beyond the bare minimum.

As usual, I use screenshots for most tweets to guard against potential future deletions or suspensions, with links to key points in the threads.

(As Kevin says, I did indeed mean should there.)

At this point there are two important threads that follow, and one additional reply of note.

Thread one, which got a bit tangled at the beginning but makes sense as one thread:

Thread two, which took place the next day and went in a different direction.

Link here to Ben’s post, GiveWell and the problem of partial funding.

Link to GiveWell blog post on giving now versus later.

Dustin’s “NO WE ARE FAILING” point seemed important so I highlighted it.

There was also a reply from Eliezer.

And this on pandemics in particular.

Sarah asked about the general failure to convince Dustin’s friends.

These two notes branch off of Ben’s comment that covers-all-of-EA didn’t make sense.

Ben also disagreed with the math that there was lots of opportunity, linking to his post A Drowning Child is Hard to Find.

This thread responds to Dustin’s claim that you need to know details about the upgrade to the laptop further up the main thread, I found it worthwhile but did not include it directly for reasons of length.

This came in response to Dustin’s challenge on whether info was 10x better.

After the main part of thread two, there was a different discussion about pressures perhaps being placed on students to be performative, which I found interesting but am not including for length.

This response to the original Tweet is worth noting as well.

Again, thanks to everyone involved and sorry if I missed your contribution.

Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.