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: Preregistration: Air Conditioner Test, published by johnswentworth on April 21, 2022 on LessWrong.

Warning: None of the participants in the Great Air Conditioner Debate Of 2022 have endorsed my summaries of their positions in this post. Including me.

Background

In Everything I Need To Know About Takeoff Speeds I Learned From Air Conditioner Ratings On Amazon, I complained about the top-rated new air conditioner on Amazon. I claimed that it’s a straightforward example of a product with a major problem, but a major problem which most people will not notice, and which therefore never gets fixed. Specifically: although it does cool the room, the air conditioner also pulls hot air from outside into the house. People do notice the cool air blowing from the air conditioner, but don’t think to blame the air conditioner for hot air drawn into the house elsewhere. Simply adding a second hose would fix the problem at relatively low extra cost, and dramatically improve the effectiveness of the air conditioner. But companies don’t actually do that because (apparently) people mostly don’t notice the problem.

To my surprise, multiple commenters disagreed with my interpretation of the air conditioner example. They argue that in fact one-hose air conditioners work fine. Sure, single-hose air conditioners are less-than-ideally efficient compared to two-hose, but it’s not a very big difference in practice. CEER efficiency ratings account for the problems, and the efficiency difference is typically only about 20-30%. Also, The Wirecutter tested lots of portable air conditioners and found that there wasn’t much difference between one-hose and two-hose designs. (Credit to Paul for both those pieces of evidence.) Really, what this example illustrates is that simple models and clever arguments are not actually very reliable at predicting how things work in practice. One should instead put more trust in experiment and reported experiences, including all those 5-star ratings on Amazon.

I, on the other hand, think the “second hose doesn’t help much” claim is a load of baloney. I think it is far more probable that CEER ratings are bullshit and The Wirecutter messed up their test, than that a second hose makes only a small difference.

And so began The Great Air Conditioner Debate Of 2022.

. Why Do Air Conditioners Need Hoses?

Ideally, an air conditioner should work much like a fridge: it pumps heat from air inside to air outside. Inside and outside air do not touch or mix; only heat flows from one to the other.

A portable air conditioner sits inside the house. So, in order to pump heat to the outside air (without letting it mix with inside air) it needs two hoses. One hose runs from a window to the air conditioner, and sucks in outside air. The other runs from the air conditioner back to the window, and blows the outside air back out. Inside air comes in and out through vents in the air conditioner, and the unit pumps heat from the inside air to the outside air, keeping the two separate throughout the process.

A single-hose air conditioner doesn’t do that. A single hose air conditioner sucks in indoor air, splits it into two streams, and pumps heat from one stream to the other. The hotter stream blows out the window (via the one hose); the cooler stream blows back into the room.

The problem with a single-hose design is that it blows air from inside to outside; it removes air from the room. That lowers the pressure in the room slightly, so new air is pulled back in via whatever openings the house has. That air comes from outside, so presumably it’s warm - and it’s replacing formerly-cool indoor air. (Technical term for this problem: “infiltration”.)

Oversimplified Summary Of The Debate

I’m not even going to try to do justice here, just give what I currently think are the key points, in roughly ch...

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: How much current animal suffering does longtermism let us ignore?, published by Jacob Eliosoff on April 21, 2022 on The Effective Altruism Forum.

Some thoughts on whether/why it makes sense to work on animal welfare, given longtermist arguments. TLDR:

We should only deprioritize the current suffering of billions of farmed animals, if we would similarly deprioritize comparable treatment of millions of humans; and,

We should double-check that our arguments aren't distorted by status quo bias, especially power imbalances in our favor.

This post consists of six arguments:

If millions of people were being kept in battery cages, how much energy should we redirect away from longtermism to work on that?

Power is exploited, and absolute power is exploited absolutely

Sacrificing others makes sense

Does longtermism mean ignoring current suffering until the heat death of the universe?

Animals are part of longtermism

None of this refutes longtermism

Plus some context and caveats at the bottom.

A. If millions of people were being kept in battery cages, how much energy should we redirect away from longtermism to work on that?

Despite some limitations, I find this analogy compelling. Come on, picture it. Check out some images of battery cages and picture millions of humans kept in the equivalent for 100% of their adult lives, and suppose with some work we could free them: would you stick to your longtermist guns?

Three possible answers:

Yes, the suffering of both the chickens and the humans is outweighed by longtermist concerns (the importance of improving our long-term future).

No, the suffering of the humans is unacceptable, because it differs from the suffering of the chickens in key ways.

No, neither is acceptable: longtermism notwithstanding, we should allocate significant resources to combating both.

I lean towards c) myself, but I can see a case for a): I just think if you're going to embrace a), you should picture the caged-humans analogy so you fully appreciate the tradeoff involved. I'm less sympathetic to b) because it feels like suspicious convergence - "That theoretical bad thing would definitely make me change my behavior, but this actual bad thing isn't actually so bad" (see section B below). Still, one could sketch some plausibly relevant differences between the caged chickens and the caged humans, eg:

"Millions of people" are subbing here for "billions of hens", implying something like a 1:1,000 suffering ratio (1 caged chicken = 0.001 caged humans): this ratio is of course debatable based on sentience, self-awareness, etc. Still, 0.001 is a pretty tiny factor (raw neuron ratio would put 1 chicken closer to 0.002-0.005 humans) and again uncertainty does some of the work for us (the argument works even if it's only quite plausible chicken suffering matters). There is a school of thought that we can be 99+% confident that a billion chickens trapped on broken legs for years don't outweigh a single human bruising her shin; I find this view ridiculous.

Maybe caging creatures that are "like us" differs in important ways from caging creatures that are "unlike us". Like, maybe allowing the caging of humans makes it more likely future humans will be caged too, making it (somehow?) of more interest to longtermists than the chickens case. (But again, see section B.)

A lot of longtermism involves the idea that humans (or AIs), unlike hens, will play a special role in determining the future (I find this reasonable). Maybe this makes caging humans worse.

B. Power is exploited, and absolute power is exploited absolutely

A general principle I find useful is, when group A is exploiting group B, group A tends to come up with rationalizations, when in fact it's often just a straight-up result of a power imbalance. I sometimes picture a conversation with a time traveler from a future 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: How to place a bet on the end of the world, published by AllAmericanBreakfast on April 20, 2022 on LessWrong.

Edit: This post is about a practical method for structuring a bet. It's not on how to extract information about people's true confidence levels based on the betting odds.

Bryan Caplan and Eliezer Yudkowsky figured out a way to place a meaningful bet on the end of the world. Bryan described their simple approach in his April 2022 80,000 Hours podcast interview.

In brief, Bryan assigns a lower probability than Eliezer to the world ending by January 1, 2030. Of course, if the world ends by that date, Eliezer can't collect his winnings. So Bryan prepays Eliezer now, and Eliezer pays him back if he loses the bet.

This works for friendly bets. Bryan points out that they need to account for interest. To that, I'd add both risk and opportunity cost. Commenters have changed my mind, however, about whether we can really learn much about the participants' confidence levels from such a bet. I elaborate on this in the comments.

Here's the relevant part of the transcript, under "Bryan's Betting Record" starting at 1:48:19 in the podcast episode.

Bryan Caplan: Let’s see. Well, I’ve already done one huge [bet against the effective altruism community], and this is that machines are going to kill us — or at least do something terrible to us — in the medium term. So I literally have an end-of-the-world bet with Eliezer Yudkowsky.

Rob Wiblin: Oh wow. I didn’t know that.

Bryan Caplan: Yeah. Which many people believe cannot be made, but it’s super easy. The person who disbelieves in the end of the world just pays the money now. And then if the world does not end, the loser pays back with whatever the odds are.

Rob Wiblin: I can’t believe I didn’t think of that.

Bryan Caplan: Yeah. So Eliezer, we actually have a bet. It takes a little effort to understand the bet, because his view is so specific. He said, “Look, I want a bet on there will no longer be any human beings on the surface of the Earth on January 1, 2030.” I was willing to give him, like, “How about all of human extinction?” — “No, no, no, no, no. There could still be humans in mine shafts. That’s OK. But not the surface of the Earth.” And I’m like, “All right. If that’s such a big deal to you, fine, we’ll make it the surface of the Earth, whatever.” But yes.

Bryan Caplan: So anyway, we have a bet, where I don’t remember the exact odds. It might be just like two-to-one. And I prepaid, so implicitly there’s interest. So it’s not as good as it seems.

It's not in Bryan's public betting inventory at the time of publication of this post, but he describes it in more detail here.

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: For every choice of AGI difficulty, conditioning on gradual take-off implies shorter timelines., published by Francis Rhys Ward on April 21, 2022 on The AI Alignment Forum.

Alternate titles: Deconfusing Take-off; Taboo “Fast” and “Slow” Take-off.

This post has two main purposes:

To suggest that we taboo the terms "fast" and "slow" take-off in favour of "gradual" and "sudden" take-off, respectively (in order to avoid confusion with “short” and “long” timelines). I also encourage people to discuss take-off "dynamics" instead of take-off "speeds".

To highlight a point made by Buck S. at EAG London and by Paul C. here: For every choice of "AGI difficulty", gradual take-off implies shorter timelines.

Although these points have been made before, I expect some people will be deconfused by this post.

To clarify terms:

Short vs long timelines is the question of when (i.e. by which year) we will develop AGI/TAI/Superintelligence.

Paul Christiano operationalizes a gradual take-off as:

There will be a complete 4 year interval in which world output doubles, before the first 1 year interval in which world output doubles. (Similarly, we’ll see an 8 year doubling before a 2 year doubling, etc.)

Sudden take-off is just the negation of this statement.

Instead of considering economic output, we can also characterize take-off in terms of AI capabilities. For some measure of AI capabilities, we could say a gradual-take-off is a situation in which:

There will be a complete 4 year interval in which AI capabilities double, before the first 1 year interval in which AI capabilities double. (Similarly, we’ll see an 8 year doubling before a 2 year doubling, etc.)

Daniel Kokotajlo also argues against GDP as a metric for AI timelines and take-off dynamics, pointing to other factors that we actually care about (e.g. warning-shots, multi-polarity, etc).

Take-off is therefore not a question about timelines – it’s a question about AI capabilities (or the economic usefulness of AI) before some threshold level which is sufficient for explosive growth (e.g. the level of intelligence needed for recursive self-improvement, or some discontinuous phase-change in capabilities).

Here is a graph of gradual vs sudden take-off (for completeness I’ve included a “no take-off” curve):

As you can see, the gradual take-off scenario happens at an earlier time.

In a gradual take-off world, AI systems become increasingly capable and useful before reaching the threshold for explosive growth. In this world, pre-AGI systems can fully automate or rapidly speed up many tasks. They provide lots of economic value and transform the world in many ways. One effect of these systems is to speed up AI research – implying we get to AI sooner than in a sudden take-off world (conditioned on a given level of “AGI difficulty”). Furthermore, the increased value and usefulness of these systems causes more funding, talent, and compute to be invested in AI, leading to further improvements (leading to more funding/talent/compute.). In addition, AI replaces an increasing number of jobs, freeing up more people to work on AI (as researchers, engineers, overseers, data-generators, etc.). The interplay of all of these factors leads to a virtuous circle which ultimately causes an increasing rate of AI capability growth.

In a sudden take-off world, AIs haven’t been very valuable up until some threshold of capability (at which point something like recursive improvement kicks in). Therefore they haven’t changed the world much and in particular haven’t made AI research go faster, so it's probably taken us longer to get to AGI.

In some very loose sense, you can think of the area under the take-off curve (cut off at the point where the line crosses the threshold) as having to be similar in both settings (if the y axis corresponds to improving AI research...

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: Neil Buddy Shah has been appointed CEO of the Clinton Health Access Initiative, published by GiveWell on April 20, 2022 on The Effective Altruism Forum.

I am excited to share that GiveWell Managing Director Buddy Shah has been appointed CEO of the Clinton Health Access Initiative (CHAI), a major global health organization working across a range of issues including malaria prevention and maternal and neonatal health. This news is bittersweet for me. I’m sad to be losing the talent, advice, and thought partnership Buddy brought to GiveWell, but I’m thrilled that he is taking up this position—the global health sector will be stronger for it.

CHAI is gaining a great leader in Buddy. But perhaps more importantly for GiveWell and our supporters, this appointment is a signal that effective giving is contributing to more corners of the global health landscape than ever before. Buddy is a strong champion of impact maximization, and I am excited that he will apply this lens in his new role.

I am also pleased that this transition does not mark the end of Buddy’s relationship with GiveWell. It is important that GiveWell maintain strong connections with leading organizations in the global health sector. With Buddy at the helm of CHAI, there will be another important voice advocating for programs that increase health outcomes as much as possible per dollar spent.

We wish our best to Buddy in his new role, and we look forward to maximizing impact together for many years to come.

Elie Hassenfeld, GiveWell Co-Founder and CEO

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: Release of Existential Risk Research database, published by rumtin on April 20, 2022 on The Effective Altruism Forum.

Today, I am releasing an existential risk research database:

This database collects all published academic articles, reports and government submissions specifically relating to existential and global catastrophic risk over 2016-2021. Each article is categorised by policy relevance and risk category.

I have attempted to capture all such publications over this period - though some may have been missed by my search. As at 20 April 2022, 376 publications have been included in the database.

The purpose of this database is to provide researchers and other interested people a resource for finding relevant publications that may assist in their work. By classifying by policy relevance, it may reveal publications that could best inform your policy understanding, research or advice.

It is not meant to be perfect. You may disagree with the categories, and some articles I have deemed not specifically relevant to existential risk are judgement calls. Nor is it a permanent solution; automated processes will be more superior, specifically TERRA. This database was developed incidentally when cataloguing policy ideas for the Policy Ideas Database, so I am sharing in case people find value in it as well.

Q&A

Which sources are included?

So far, I’ve included academic articles, government submissions and reports, published between 2016 and 2021. Publications must have been authored by academics or by organisations that identify as entirely or explicitly focused on existential risk. The publications must be specifically related to existential and global catastrophic risks. Those related to the broad risks but not at the existential or catastrophic scale (such as 'normal' climate change or general AI development) are not included.

Where were publications found?

I first used TERRA’s bibliography. I then went through each organisation’s website and identified publications they’ve listed. Cursory (though not systematic) Google Scholar searches were also done.

Gimme the stats!

My search identified 758 possibly relevant publications, with 376 identified as relevant to existential and global catastrophic risk. Just over half have direct policy relevance - that is, they provide implications or ideas for policy communities.

Since 2018, the number of publications, and their policy relevance, has stayed relatively even.

Most publications are related to existential risk at the general level and to technological risks (primarily AI). This is mirrored when looking at policy-relevant publications only.

Of the 190 policy-relevant publications, 68 have specific policy ideas or options. I have captured ideas from 29 of these publications in the policy idea database.

Seems interesting and all. but what do I do with this information?

There may be a few implications of this:

The field may need to focus more policy work where policy-relevance has been underweighted

The policy of nuclear weapons risk seems underweighted, though the database does not include publications from the Nuclear Threat Initiative (NTI) for now. Extreme climate change may be another area.

Of the risk categories, AI had the most policy relevant publications, but the specificity of policy ideas was low.

Researchers may wish to make more clear the policy implications of their research, even if they decide not to provide specific policy ideas. Around 40% of publications made claims that could be relevant for policy, but did not make a direct link.

The field may wish to look at developing more ‘translation’ products and services - taking policy implications and ideas from publications into more digestible formats (e.g. policy briefs and government submissions, infographics, co-creation of policy proposals with policy-makers).

...

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: Judge Overturns Transportation Mask Mandate, published by Zvi on April 20, 2022 on LessWrong.

A federal judge has overturned the transportation mask mandate, saying the CDC overstepped its authority. The Biden administration is adhering to this ruling. It intends to appeal, but is not being loud about that.

Mask mandates are lifted on planes. This happened in-flight. Reactions from passengers and crew in flight were highly supportive.

Public health advocates naturally took a somewhat different view of events, this being a central reasonable version of their reaction.

There’s that ‘don’t change things people won’t like it’ principle again, and also citing whatever invokes the availability heuristic this week. Most importantly he collected data from this experiment, and remember that connecting to Boston via JetBlue should be an unusually liberal set of passengers.

Lifting the mandate during flights led to some great moments.

As a small point yes I do think this could have waited a few hours given that people boarded the planes with an expectation?

In future, I’d urge such folks to get P100 or superior masks, which should do a much better job on you alone than the KN95s did on everyone combined. Which was the right play before but almost no one did it, and probably almost no one will do it in the future either. In general, there are two reasons to wear a mask – you either want to be someone wearing a mask, or you want actual protection for yourself or others. Now that we no longer need to do the first one, those who want protection can do the second, which means P100 or better.

Uber has lifted its own mandate in response, as has Lyft. Some local authorities like the MTA in NYC (which includes the Subway) and NJ Transit are leaving their mandates in place, others are not.

What should we make of all this?

The Details Of This Decision Imply Bad Faith

The transportation mask mandate was already hanging by a thread. The CDC had multiple times come close to letting it expire. The CDC’s latest extension of the mandate was only for two weeks. The Senate voted 57-40 to repeal the mandate. Airlines, who have the most skin in the game, called for lifting the mandate.

The Biden administration so far held firm, as part of its ‘trust the science’ branding. This was a no-win situation for them, with inevitable heat on them no matter their decision.

Having a judge take it out of their hands does them a favor, letting them dodge blame for lifting the policy while also dodging blame for not lifting the policy. All bad outcomes can be blamed on a Trump-appointed judge that many are calling extraordinarily unqualified and can thus make the scapegoat/hero?

That’s perfect. No wonder there is no rush to appeal.

The timing of this ruling strongly implies to me that it was made in light of these facts.

Does the CDC have the authority to impose a mask mandate on the airlines? That is an interesting legal question. I do not know whether this particular rule overstepped their authority. It seems plausible that it did.

If it did, then the mandate needed to be lifted. We are supposed to be a nation of laws.

Poe’s Law for the win in the replies, I can’t tell either.

During the pandemic we got, even more than usual, into the mindset that we do not have a constitution or separation of powers or limits on Presidential or federal authority. In theory that is very wrong, and in practice it once again is being treated that way.

Consider the parallel to the eviction moratorium. The CDC decided that it was a public health issue to forcibly let people live rent free indefinitely in other people’s houses and apartments. Then Trump gave way to Biden, and what was previously allowed was not allowed. I’d strongly agree in that case that this was very much not within the CDC’s authority, but if that is true, it 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: What do we want the world to look like in 10 years?, published by Owen Cotton-Barratt on April 20, 2022 on The Effective Altruism Forum.

We have a lot of uncertainty over the world we'll see in a decade. There are many different dimensions it might vary on. Some of these are likely much more important than others for ensuring a good future. Which ones?

I'm keen to see more discussion of this in longtermist EA circles. I think we have a lot of analysis of very long-term outcomes we're aiming for ("ensure AI is aligned"; "embark on a long reflection"), and a lot of discussion of immediate plans we're considering, and relatively little of what's good on this intermediate timescale. But I think it's really important to understand for informing more immediate plans. It isn't just enough to identify a handful of top priority variables (like "quality-adjusted amount of AI alignment research"), because comparative advantage can vary between people, and sometimes there are high-leverage opportunities available for achieving various ends, and it's helpful to understand how good those are.

I've been getting mileage out of this prompt as an exercise for groups over the last year. I think people don't need to have a full strategic understanding to fruitfully engage with the question, but that answers aren't simple even for people who've thought about it a lot. Discussion of detailed cases often seems productive (at least I find it productive to think about individual cases, and I've liked conversations I've observed others having).

Examples of desirable states

Differential technological development is a major principle used in the allocation of public research spending

We have ways of monitoring and responding to early-stage potential pandemic pathogens that make civilization-ending pandemics almost impossible

There is a variety of deeply inspiring art, and a sense of hope for the future among society broadly

There is a textbook on the AI alignment problem which crisply sets out the problem in unambiguous technical terms, and is an easy on-ramp for strong technical researchers, while capturing the heart of what's important and difficult about it

Society has found healthier models of relating to social media, in which it is less addictive and doesn't amplify crazy but clickbait-y views

There are more robust fact-checking institutions, and better language for discussing the provenance of beliefs in use among cultural and intellectual elites

We have better models for avoiding rent-seeking behaviour

Tensions between great powers are low

These examples would be even better if they were more concrete/precise (such that it would be unambiguous on waking up in 10 years whether they had been achieved or not), but often the the slightly fuzzy forms will be more achievable as a starting point.

This is a short list of examples; in practice when I've run this exercise for long enough people have had hundreds of ideas. (Of course some of the ideas are much better than others.)

Consider spending time on this question

I wanted to share the prompt as an invitation to others to spend time on, either by themselves or as a prompt for a group exercise. I've liked coming back to this multiple times, and I expect I'll continue doing that.

This question is close to cause prioritization. I think of it as complementary rather than a replacement. Reasons to include this in the mix of things to think about:

The cause prioritization frame nudges towards identifying a single best thing and stopping, but I think it's often helpful in practice to have thoughts on a suite of different things

e.g. for noticing particularly high-leverage opportunities

Immediate plans for making progress on large causes must factor through effects in the intermediate future; it can be helpful to look directly at what we're aiming f...

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: Longtermist EA needs more Phase 2 work, published by Owen Cotton-Barratt on April 20, 2022 on The Effective Altruism Forum.

I think there is an issue with the community-wide allocation of effort. A very large proportion of our effort goes into preparation work, setting the community up for future successes; and very little goes into external-focused actions which would be good even if the community disappeared. I'll talk about why I think this is a problem (even though I love preparation work), and what types of things I hope to see more of.

Phase 1 and Phase 2

A general strategy for doing good things:

Phase 1: acquire resources and work out what to do

Phase 2: spend down the resources to do things

Note that Phase 2 is the bit where things actually happen. Phase 1 is a necessary step, but on its own it has no impact: it is just preparation for Phase 2.

To understand if something is Phase 2 for the longtermist EA community, we could ask “if the entire community disappeared, would the effects still be good for the world?”. For things which are about acquiring resources — raising money, recruiting people, or gaining influence — the answer is no. For much of the research that the community does, the path to impact is either by using the research to gain more influence, or having the research inform future longtermist EA work — so the answer is again no. However, writing an AI alignment textbook would be useful to the world even absent our communities, so would be Phase 2. (Some activities live in a grey area —for example, increasing scope sensitivity or concern for existential risk across broad parts of society.)

It makes sense to frontload our Phase 1 activities, but we do want to also do Phase 2 in parallel for several reasons:

Doing enough Phase 2 work helps to ground our Phase 1 work by ensuring that it’s targeted at making the Phase 2 stuff go well

Moreover we can benefit from the better feedback loops Phase 2 usually has

We can’t pivot (orgs, careers) instantly between different activities

A certain amount of Phase 2 helps bring in people who are attracted by demonstrated wins

We don’t know when the deadline for crucial work is (and some of the best opportunities may only be available early) so we want a portfolio across time

So the picture should look something like this:

I’m worried that it does not. We aren’t actually doing much Phase 2 work, and we also aren’t spending that much time thinking about what Phase 2 work to be doing. (Although I think we’re slowly improving on these dimensions.)

Problem: We are not doing Phase 2 Work

When we look at the current longtermist portfolio, there’s very little Phase 2 work. A large majority of our effort is going into acquiring more resources (e.g. campus outreach, or writing books), or into working out what the-type-of-people-who-listen-to-us should do (e.g. global priorities research).

This is what we could call an inaction trap. As a community we’re preparing but not acting. (This is a relative of the meta-trap, but we have a whole bunch of object-level work e.g. on AI.)

How does AI alignment work fit in?

Almost all AI alignment research is Phase 1 — on reasonable timescales it’s aiming to produce insights about what future alignment researchers should investigate (or to gain influence for the researchers), rather than producing things that would leave the world in a better position even if the community walked away.

But if AI alignment is the crucial work we need to be doing, and it's almost all Phase 1, could this undermine the claim that we should increase our focus on Phase 2 work?

I think not, for two reasons:

Lots of people are not well suited to AI alignment work, but there are lots of things that they could productively be working on, even if AI is the major determinant of the future (see below)

Even within AI 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: Help with the Forum; wiki editing, giving feedback, moderation, and more, published by Lizka on April 20, 2022 on The Effective Altruism Forum.

There are a number of ways you could help make the Forum (even) better, and we’re opening an expression of interest form in case you’d like to get involved in any of the ways listed below.

Note that we’re interested in a variety of models; you could be a contractor working a certain number of hours, a volunteer, or even potentially a full-time hire.

EA Forum Wiki

Editing and adding to the EA Forum Wiki

Giving feedback on drafts of Forum posts

Forum upkeep

Deleting spam and approving non-spam

Adding relevant tags to new posts

Moderation

Enforcing and promoting the Forum norms.

General expression of interest

[Bonus] For everyone: actions you can take to improve the Forum

⇨ Express interest ⇦

You do not need to have any particular experience for any of these tasks. If interested and in doubt about whether you’re qualified, please just submit the expression of interest form!

Please note that this is just an expression of interest. We may end up not adding anyone to the team.

By default, we would suggest a contractor rate of around $30 per hour for most of this work, but you could volunteer your time if you prefer. Declaring a wish to volunteer will not make us more likely to work with you on this.

1. The EA Forum Wiki

The EA Forum Wiki does the following:

Explains key concepts in an accessible way (here’s an example)

Helps organize writing; it collects content related to a given EA-relevant topic in one place, so that it’s more discoverable (for instance, you can find the “improving institutional decision-making” posts at the relevant tag)

It might also serve as a basis for communities or subforums that will be formed around given topics (e.g. “biosecurity” or “animal welfare” communities)

Forum users do a lot of great work to make the Wiki happen, but having regular contributors could help unlock the full potential of the Wiki.

You can read more about the point and functioning of the Wiki here.

What you can do:

Write and improve entries

Help organize the Wiki

Tag posts

You’d likely be a good fit if you:

Are happy to learn about a broad range of EA-relevant topics

Read content on the Forum a lot

Are willing to write sometimes

If any of this is interesting to you and you think you might be able to work on it, please express interest!

2. Giving feedback on drafts of Forum posts

Would you enjoy looking at drafts of posts and giving feedback? We may be interested in putting you in touch with people who are looking for feedback; we think this often leads to good posts that wouldn’t have happened otherwise and is generally a useful community service. You’ll work with Lizka (me). I’ll be giving feedback occasionally, but may not have time to give feedback every time it’s requested, and will direct people to you when that seems appropriate. (Note that you could also just be an active participant in the EA Editing and Review Facebook group.)

You’d likely be a good fit if you:

Are encouraging

Would enjoy reading drafts and giving feedback

Have a sense of what makes for a good Forum post

[Not required] Have domain expertise in a particular field

3. Forum upkeep

As sketched out above, the main upkeep tasks required on the Forum are:

Deleting spam and approving new non-spam users

Adding relevant tags to new posts

Sorting spam and adding tags are some of the most crucial activities listed in this post. We’re doing our best to ensure that the Forum is one of the clean and safe places on the internet, and we’d love help with that.

I’m interested in hearing from you even if you’d just be willing to commit to regularly going to the Frontpage to approve non-spam users and delete spam comments.

4. Moderation

This involves occasionally helping to emp...