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 2020 Review: Preliminary Voting, published by Raemon on December 2, 2021 on LessWrong.

Today is the first day of the LessWrong 2020 Review. At the end of each year, we take a look at the best posts from the previous year, and reflect on which of them have stood the test of time.

As we navigate the 21st century, a key issue is that we’re pretty confused about which intellectual work is valuable. We have very sparse reward signals when it comes to “did we actually figure out things that matter?”

The Review has a few goals. It improves our incentives, feedback, and rewards for contributing to LessWrong. It creates common knowledge about the LW community's collective epistemic state about the most important posts of 2020. And it turns all of that into a highly curated sequence that people can read. You can read more about the philosophy behind the Review in last year’s announcement post.

A few important announcements about this year’s review:

We’ve replaced the nomination process with Preliminary Voting.

Winning posts will get Donation Buttons.

There’s a new View My Past Upvotes page to help you find posts to vote on.

2019 Books will be shipping in a couple weeks. We’re still evaluating whether and how to do books for 2020.

How does the review work?

The review has three phases:

Preliminary Voting Phase (Dec 1- 14)

Discussion Phase (Dec 14 - Jan 11)

Final Voting (Jan 11 to Jan 25)

Users who registered before January 1st 2020 can vote. The LessWrong moderation team will take the results of the vote as input for a curated sequence of posts, and award prizes. We’ll be giving more weight to the votes of users with 1000+ karma.

Preliminary Voting

The first big change this year is changing the Nomination Phase to the Preliminary Vote Phase. Eligible voters voters will see this UI:

If you think a post was an important intellectual contribution, you can cast a vote indicating roughly how important it was. A vote of 1 means “it was good.” A vote of 4 means “it was quite important”, and is weighted 4x a vote of 1. A vote of 9x means it was a crucial piece of intellectual progress.

You can vote at the top of a post, or anywhere the post appears in a list (like the All Posts page, or the new View Your Past Upvotes page).

Posts that get at least one positive vote go to the Voting Dashboard, where other users can vote on it. You’re encouraged to give at least a rough vote based on what you remember from last year.

If you feel a post was important, you’re also encouraged to write up at least a short review of it saying what stands out about the post and why it matters. (This is essentially the same as writing a nomination comment from the 2018 and 2019 Reviews. In practice nominations and reviews were fairly similar and it didn’t seem worth separating them out in the UI). You’re allowed to write multiple reviews of a post, if you want to start by jotting down your quick impressions, and later review it in more detail.

Why did we switch to preliminary voting?

Each year, more posts get written on LessWrong. The first Review of 2018 considered 1,500 posts. In 2020, there were 3,000. Processing that many posts is a lot of work.

Preliminary voting is designed to help handle the increased number of posts. Instead of simply nominating posts, we start directly with a votes. At the end of the Preliminary Voting phase, the results of the vote will be published. This will help the LessWrong community prioritize reviews. Posts that are highly ranked can invite more investigation of how they stand the tests of time. If you think a post was (unfairly) ranked low, you are welcome to write a positive review arguing it should be considered more strongly.

Posts which everyone agrees are “meh” can get deprioritized, making more time for more interesting posts.

How is preliminary voting calculated?

You...

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 2020 Review: Preliminary Voting, published by Raemon on December 2, 2021 on LessWrong.

Today is the first day of the LessWrong 2020 Review. At the end of each year, we take a look at the best posts from the previous year, and reflect on which of them have stood the test of time.

As we navigate the 21st century, a key issue is that we’re pretty confused about which intellectual work is valuable. We have very sparse reward signals when it comes to “did we actually figure out things that matter?”

The Review has a few goals. It improves our incentives, feedback, and rewards for contributing to LessWrong. It creates common knowledge about the LW community's collective epistemic state about the most important posts of 2020. And it turns all of that into a highly curated sequence that people can read. You can read more about the philosophy behind the Review in last year’s announcement post.

A few important announcements about this year’s review:

We’ve replaced the nomination process with Preliminary Voting.

Winning posts will get Donation Buttons.

There’s a new View My Past Upvotes page to help you find posts to vote on.

2019 Books will be shipping in a couple weeks. We’re still evaluating whether and how to do books for 2020.

How does the review work?

The review has three phases:

Preliminary Voting Phase (Dec 1- 14)

Discussion Phase (Dec 14 - Jan 11)

Final Voting (Jan 11 to Jan 25)

Users who registered before January 1st 2020 can vote. The LessWrong moderation team will take the results of the vote as input for a curated sequence of posts, and award prizes. We’ll be giving more weight to the votes of users with 1000+ karma.

Preliminary Voting

The first big change this year is changing the Nomination Phase to the Preliminary Vote Phase. Eligible voters voters will see this UI:

If you think a post was an important intellectual contribution, you can cast a vote indicating roughly how important it was. A vote of 1 means “it was good.” A vote of 4 means “it was quite important”, and is weighted 4x a vote of 1. A vote of 9x means it was a crucial piece of intellectual progress.

You can vote at the top of a post, or anywhere the post appears in a list (like the All Posts page, or the new View Your Past Upvotes page).

Posts that get at least one positive vote go to the Voting Dashboard, where other users can vote on it. You’re encouraged to give at least a rough vote based on what you remember from last year.

If you feel a post was important, you’re also encouraged to write up at least a short review of it saying what stands out about the post and why it matters. (This is essentially the same as writing a nomination comment from the 2018 and 2019 Reviews. In practice nominations and reviews were fairly similar and it didn’t seem worth separating them out in the UI). You’re allowed to write multiple reviews of a post, if you want to start by jotting down your quick impressions, and later review it in more detail.

Why did we switch to preliminary voting?

Each year, more posts get written on LessWrong. The first Review of 2018 considered 1,500 posts. In 2020, there were 3,000. Processing that many posts is a lot of work.

Preliminary voting is designed to help handle the increased number of posts. Instead of simply nominating posts, we start directly with a votes. At the end of the Preliminary Voting phase, the results of the vote will be published. This will help the LessWrong community prioritize reviews. Posts that are highly ranked can invite more investigation of how they stand the tests of time. If you think a post was (unfairly) ranked low, you are welcome to write a positive review arguing it should be considered more strongly.

Posts which everyone agrees are “meh” can get deprioritized, making more time for more interesting posts.

How is preliminary voting calculated?

You...

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: Effective Altruism: The First Decade (Forum Review), published by Aaron Gertler on December 1, 2021 on The Effective Altruism Forum.

Following in the footsteps of LessWrong, we are running our first-ever EA Forum Review.

The purpose of the Review is to figure out which writings from the EA community have produced actual value, so that we can build on their ideas and share them with many more readers.

The Review uses the following process:

Users cast preliminary votes to nominate posts they’ve found valuable for the final voting stage. (December 1 - December 14)

Users review posts — explaining how the posts have been useful (or not), which of their ideas seem best (or worst), and so on. (December 15 - January 11)

Users cast final votes on posts[1], based on their own experience and others’ reviews. The posts with the most votes are, hopefully, those that have provided the most collective value to the community. (January 12 - January 26) To see what this looks like in practice, check out LessWrong’s 2019 Review and its results.[2]

Why hold a review?

Effective altruism is focused on impact. But it’s hard to tell which intellectual work has actually made an impact.

The karma system shows us how many people liked a post right away. But that’s different from knowing what a post has achieved over the full span of its existence. Did it change the movement forever? Did people forget about it completely? Is it still useful?

The Review is our way of trying to answer questions like:

“What are the most valuable things we’ve written? How might they be improved, or built upon?”

“Which ideas are most important for someone to understand if they want to do as much good as possible? And where can they find the best writing on those ideas?”

“What should people do if they want their own writing to be impactful? What can they learn by seeing trends and patterns in our best writing so far?”

Other benefits:

This is a chance for people to share reflective commentary on important posts, with the benefit of hindsight. What still seems true, a year or five years later? Which ideas didn’t stand the test of time?

It’s also a chance for authors to look back at their own posts and fix errors or make changes. The voting system lets them see which posts might be most impactful to update.

To quote LessWrong: “There needs to be an incentive to clean up ideas that turned out to be important but badly presented.”

Asking people to think about content they’ve valued might also lead them to crosspost it, making the Forum a more complete repository of EA’s intellectual contributions.

How does it work?

The Review will be visible to anyone who had a Forum account before December 1st, 2021.[3]

Our first review covers all posts published before 2021.[4] To quote LessWrong:

A year is probably enough time to no longer be swept away in the news or excitement of the day, but recent enough that we can still remember and write down how an idea or explanation has affected us.

(But don’t be shy about taking your best guess at how something much older affected you.)

Stage 1: Nominations / Preliminary Votes

From today through January 11th, eligible users will see this feature on eligible posts:

You can vote for as many posts as you like; a post will need at least one positive vote to be nominated to the second round. Negative votes don’t affect nomination status — only how posts are ranked.

The nomination scoring system

Here’s how we interpret each voting score[5]:

0 // Neutral: You have no strong feelings about the post.[6]

1 // Good: You found the post valuable. You’ve referred back to it before, or would recommend it to others.

4 // Important: Like “Good”, but more so. Perhaps the post includes one or more key insights that have changed the way you think, or led you to take useful actions.

9 // Crucial: The post is tim...

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: Effective Altruism: The First Decade (Forum Review), published by Aaron Gertler on December 1, 2021 on The Effective Altruism Forum.

Following in the footsteps of LessWrong, we are running our first-ever EA Forum Review.

The purpose of the Review is to figure out which writings from the EA community have produced actual value, so that we can build on their ideas and share them with many more readers.

The Review uses the following process:

Users cast preliminary votes to nominate posts they’ve found valuable for the final voting stage. (December 1 - December 14)

Users review posts — explaining how the posts have been useful (or not), which of their ideas seem best (or worst), and so on. (December 15 - January 11)

Users cast final votes on posts[1], based on their own experience and others’ reviews. The posts with the most votes are, hopefully, those that have provided the most collective value to the community. (January 12 - January 26) To see what this looks like in practice, check out LessWrong’s 2019 Review and its results.[2]

Why hold a review?

Effective altruism is focused on impact. But it’s hard to tell which intellectual work has actually made an impact.

The karma system shows us how many people liked a post right away. But that’s different from knowing what a post has achieved over the full span of its existence. Did it change the movement forever? Did people forget about it completely? Is it still useful?

The Review is our way of trying to answer questions like:

“What are the most valuable things we’ve written? How might they be improved, or built upon?”

“Which ideas are most important for someone to understand if they want to do as much good as possible? And where can they find the best writing on those ideas?”

“What should people do if they want their own writing to be impactful? What can they learn by seeing trends and patterns in our best writing so far?”

Other benefits:

This is a chance for people to share reflective commentary on important posts, with the benefit of hindsight. What still seems true, a year or five years later? Which ideas didn’t stand the test of time?

It’s also a chance for authors to look back at their own posts and fix errors or make changes. The voting system lets them see which posts might be most impactful to update.

To quote LessWrong: “There needs to be an incentive to clean up ideas that turned out to be important but badly presented.”

Asking people to think about content they’ve valued might also lead them to crosspost it, making the Forum a more complete repository of EA’s intellectual contributions.

How does it work?

The Review will be visible to anyone who had a Forum account before December 1st, 2021.[3]

Our first review covers all posts published before 2021.[4] To quote LessWrong:

A year is probably enough time to no longer be swept away in the news or excitement of the day, but recent enough that we can still remember and write down how an idea or explanation has affected us.

(But don’t be shy about taking your best guess at how something much older affected you.)

Stage 1: Nominations / Preliminary Votes

From today through January 11th, eligible users will see this feature on eligible posts:

You can vote for as many posts as you like; a post will need at least one positive vote to be nominated to the second round. Negative votes don’t affect nomination status — only how posts are ranked.

The nomination scoring system

Here’s how we interpret each voting score[5]:

0 // Neutral: You have no strong feelings about the post.[6]

1 // Good: You found the post valuable. You’ve referred back to it before, or would recommend it to others.

4 // Important: Like “Good”, but more so. Perhaps the post includes one or more key insights that have changed the way you think, or led you to take useful actions.

9 // Crucial: The post is tim...

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: Biology-Inspired AGI Timelines: The Trick That Never Works, published by Eliezer Yudkowsky on December 1, 2021 on LessWrong.

- 1988 -

Hans Moravec: Behold my book Mind Children. Within, I project that, in 2010 or thereabouts, we shall achieve strong AI. I am not calling it "Artificial General Intelligence" because this term will not be coined for another 15 years or so.

Eliezer (who is not actually on the record as saying this, because the real Eliezer is, in this scenario, 8 years old; this version of Eliezer has all the meta-heuristics of Eliezer from 2021, but none of that Eliezer's anachronistic knowledge): Really? That sounds like a very difficult prediction to make correctly, since it is about the future, which is famously hard to predict.

Imaginary Moravec: Sounds like a fully general counterargument to me.

Eliezer: Well, it is, indeed, a fully general counterargument against futurism. Successfully predicting the unimaginably far future - that is, more than 2 or 3 years out, or sometimes less - is something that human beings seem to be quite bad at, by and large.

Moravec: I predict that, 4 years from this day, in 1992, the Sun will rise in the east.

Eliezer: Okay, let me qualify that. Humans seem to be quite bad at predicting the future whenever we need to predict anything at all new and unfamiliar, rather than the Sun continuing to rise every morning until it finally gets eaten. I'm not saying it's impossible to ever validly predict something novel! Why, even if that was impossible, how could I know it for sure? By extrapolating from my own personal inability to make predictions like that? Maybe I'm just bad at it myself. But any time somebody claims that some particular novel aspect of the far future is predictable, they justly have a significant burden of prior skepticism to overcome.

More broadly, we should not expect a good futurist to give us a generally good picture of the future. We should expect a great futurist to single out a few rare narrow aspects of the future which are, somehow, exceptions to the usual rule about the future not being very predictable.

I do agree with you, for example, that we shall at some point see Artificial General Intelligence. This seems like a rare predictable fact about the future, even though it is about a novel thing which has not happened before: we keep trying to crack this problem, we make progress albeit slowly, the problem must be solvable in principle because human brains solve it, eventually it will be solved; this is not a logical necessity, but it sure seems like the way to bet. "AGI eventually" is predictable in a way that it is not predictable that, e.g., the nation of Japan, presently upon the rise, will achieve economic dominance over the next decades - to name something else that present-day storytellers of 1988 are talking about.

But timing the novel development correctly? That is almost never done, not until things are 2 years out, and often not even then. Nuclear weapons were called, but not nuclear weapons in 1945; heavier-than-air flight was called, but not flight in 1903. In both cases, people said two years earlier that it wouldn't be done for 50 years - or said, decades too early, that it'd be done shortly. There's a difference between worrying that we may eventually get a serious global pandemic, worrying that eventually a lab accident may lead to a global pandemic, and forecasting that a global pandemic will start in November of 2019.

Moravec: You should read my book, my friend, into which I have put much effort. In particular - though it may sound impossible to forecast, to the likes of yourself - I have carefully examined a graph of computing power in single chips and the most powerful supercomputers over time. This graph looks surprisingly regular! Now, of course not all trends can continue fo...

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: Biology-Inspired AGI Timelines: The Trick That Never Works, published by Eliezer Yudkowsky on December 1, 2021 on LessWrong.

- 1988 -

Hans Moravec: Behold my book Mind Children. Within, I project that, in 2010 or thereabouts, we shall achieve strong AI. I am not calling it "Artificial General Intelligence" because this term will not be coined for another 15 years or so.

Eliezer (who is not actually on the record as saying this, because the real Eliezer is, in this scenario, 8 years old; this version of Eliezer has all the meta-heuristics of Eliezer from 2021, but none of that Eliezer's anachronistic knowledge): Really? That sounds like a very difficult prediction to make correctly, since it is about the future, which is famously hard to predict.

Imaginary Moravec: Sounds like a fully general counterargument to me.

Eliezer: Well, it is, indeed, a fully general counterargument against futurism. Successfully predicting the unimaginably far future - that is, more than 2 or 3 years out, or sometimes less - is something that human beings seem to be quite bad at, by and large.

Moravec: I predict that, 4 years from this day, in 1992, the Sun will rise in the east.

Eliezer: Okay, let me qualify that. Humans seem to be quite bad at predicting the future whenever we need to predict anything at all new and unfamiliar, rather than the Sun continuing to rise every morning until it finally gets eaten. I'm not saying it's impossible to ever validly predict something novel! Why, even if that was impossible, how could I know it for sure? By extrapolating from my own personal inability to make predictions like that? Maybe I'm just bad at it myself. But any time somebody claims that some particular novel aspect of the far future is predictable, they justly have a significant burden of prior skepticism to overcome.

More broadly, we should not expect a good futurist to give us a generally good picture of the future. We should expect a great futurist to single out a few rare narrow aspects of the future which are, somehow, exceptions to the usual rule about the future not being very predictable.

I do agree with you, for example, that we shall at some point see Artificial General Intelligence. This seems like a rare predictable fact about the future, even though it is about a novel thing which has not happened before: we keep trying to crack this problem, we make progress albeit slowly, the problem must be solvable in principle because human brains solve it, eventually it will be solved; this is not a logical necessity, but it sure seems like the way to bet. "AGI eventually" is predictable in a way that it is not predictable that, e.g., the nation of Japan, presently upon the rise, will achieve economic dominance over the next decades - to name something else that present-day storytellers of 1988 are talking about.

But timing the novel development correctly? That is almost never done, not until things are 2 years out, and often not even then. Nuclear weapons were called, but not nuclear weapons in 1945; heavier-than-air flight was called, but not flight in 1903. In both cases, people said two years earlier that it wouldn't be done for 50 years - or said, decades too early, that it'd be done shortly. There's a difference between worrying that we may eventually get a serious global pandemic, worrying that eventually a lab accident may lead to a global pandemic, and forecasting that a global pandemic will start in November of 2019.

Moravec: You should read my book, my friend, into which I have put much effort. In particular - though it may sound impossible to forecast, to the likes of yourself - I have carefully examined a graph of computing power in single chips and the most powerful supercomputers over time. This graph looks surprisingly regular! Now, of course not all trends can continue fo...

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 publication: A guide to the changing landscape of high-impact climate philanthropy, published by jackva on December 1, 2021 on The Effective Altruism Forum.

We -- Luisa Sandkuehler, Violet Buxton-Walsh and myself -- just published a new piece “A guide to navigating the changing landscape of high-impact climate philanthropy”.

The full guide is here and the Executive Summary is reproduced below.

This guide is part of ongoing research, and many of the considerations mentioned will be stated more formally in the future. We also do a lot of data aggregation work to be able to state things more quantitatively.

Some further contextualization for the Forum:

What we try to do with this guide:

(1) Lay out a fairly complete statement of our thinking on high-impact climate philanthropy. We do not do any cause prioritization (“why and how much should you care about climate?”) but rather answer the question “If you want to give to climate, what are the features of the climate space that are most informative in shaping what the highest-impact giving opportunities are likely to be?”

(2) It is intentionally called a “guide” not a “research report” as it is intended to be more informal, discursive, question / response addressing the most typical questions we encounter when explaining our work.

(3) We try to provide a conceptual toolbox around impact multipliers and theories of change that is more modular than our prior work. The basic underlying mental model is to map the climate philanthropy space, to identify the factors that should leave us to expect higher or lower impact from different possible interventions and organizations. Some of those multipliers are generic in EA, such as a focus on neglected solutions and maximizing counterfactual additionality as well as risk neutrality and patience (low discounting), while others -- such as advocacy, trajectory changes, catalytically growing small promising organizations, global diffusion of technological change -- are more specific to the climate action and/or funding space. We’ll try to give more quantitative treatment of those multipliers in future work, for now we use that to illustrate what pushes us towards different theories of change as local optima.

Main updates:

(1) We analyze up-to-date on climate philanthropy, including major commitments from COP26, for detailed estimates of climate philanthropy by sector and region, building on and extending the foundational work of ClimateWorks. We also contextualize these funding allocations in light of future emissions and other factors that should affect our judgment on relative neglectedness/crowdedness.

(2) We describe four theories of change we consider the most plausible candidates for high-impact climate philanthropy, beyond (i) advocacy for neglected tech innovation (which motivates much of our prior recs and grantmaking), we also include (ii) avoiding carbon lock-in in emerging economies, (iii) catalytically growing promising organizations and (iv) promoting important policy ideas that might cascade (policy leadership and paradigm shaping).

(3) We explain why climate policy (and philanthropy) in OECD countries should be primarily judged by its global long-run effects rather than domestic short-term effects and why the two might often be negatively correlated (many of the most successful policies for global decarbonization, such as solar subsidies, have barely saved emissions in the originating jurisdictions).

(4) We lay out how we think about uncertainty in the context of non-linear climate risk, describing a criterion we call “robust diversification” as a response, combining strategies with negatively correlated uncertainties and preparing for the worst worlds that contain most climate damage (according to our current estimates of emissions futures and climate damage, this p...

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 publication: A guide to the changing landscape of high-impact climate philanthropy, published by jackva on December 1, 2021 on The Effective Altruism Forum.

We -- Luisa Sandkuehler, Violet Buxton-Walsh and myself -- just published a new piece “A guide to navigating the changing landscape of high-impact climate philanthropy”.

The full guide is here and the Executive Summary is reproduced below.

This guide is part of ongoing research, and many of the considerations mentioned will be stated more formally in the future. We also do a lot of data aggregation work to be able to state things more quantitatively.

Some further contextualization for the Forum:

What we try to do with this guide:

(1) Lay out a fairly complete statement of our thinking on high-impact climate philanthropy. We do not do any cause prioritization (“why and how much should you care about climate?”) but rather answer the question “If you want to give to climate, what are the features of the climate space that are most informative in shaping what the highest-impact giving opportunities are likely to be?”

(2) It is intentionally called a “guide” not a “research report” as it is intended to be more informal, discursive, question / response addressing the most typical questions we encounter when explaining our work.

(3) We try to provide a conceptual toolbox around impact multipliers and theories of change that is more modular than our prior work. The basic underlying mental model is to map the climate philanthropy space, to identify the factors that should leave us to expect higher or lower impact from different possible interventions and organizations. Some of those multipliers are generic in EA, such as a focus on neglected solutions and maximizing counterfactual additionality as well as risk neutrality and patience (low discounting), while others -- such as advocacy, trajectory changes, catalytically growing small promising organizations, global diffusion of technological change -- are more specific to the climate action and/or funding space. We’ll try to give more quantitative treatment of those multipliers in future work, for now we use that to illustrate what pushes us towards different theories of change as local optima.

Main updates:

(1) We analyze up-to-date on climate philanthropy, including major commitments from COP26, for detailed estimates of climate philanthropy by sector and region, building on and extending the foundational work of ClimateWorks. We also contextualize these funding allocations in light of future emissions and other factors that should affect our judgment on relative neglectedness/crowdedness.

(2) We describe four theories of change we consider the most plausible candidates for high-impact climate philanthropy, beyond (i) advocacy for neglected tech innovation (which motivates much of our prior recs and grantmaking), we also include (ii) avoiding carbon lock-in in emerging economies, (iii) catalytically growing promising organizations and (iv) promoting important policy ideas that might cascade (policy leadership and paradigm shaping).

(3) We explain why climate policy (and philanthropy) in OECD countries should be primarily judged by its global long-run effects rather than domestic short-term effects and why the two might often be negatively correlated (many of the most successful policies for global decarbonization, such as solar subsidies, have barely saved emissions in the originating jurisdictions).

(4) We lay out how we think about uncertainty in the context of non-linear climate risk, describing a criterion we call “robust diversification” as a response, combining strategies with negatively correlated uncertainties and preparing for the worst worlds that contain most climate damage (according to our current estimates of emissions futures and climate damage, this p...

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: FDA Votes on Molunpiravir, published by Zvi on December 1, 2021 on LessWrong.

While Paxlovid Remains Illegal and is expected to remain illegal for at least several weeks, the FDA did manage to finally meet to discuss whether or not to legalize the other Covid-19 treatment pill, Merck’s Molunpiravir. While later data reduced effectiveness estimates from 50% to 30%, that’s still much better than 0% and it uses a unique mechanism that can probably be profitably combined with other treatments, so one might naively think that after sufficient stalling for appearances this would be easy.

One would be wrong. The vote was 13-10, was restricted to those at high risk, and could easily have failed outright.

I may want to later refer back to this, so it’s splitting off into its own post.

High Level Summary

Usually we get live-blogs from Helen Branswell and Mattew Herper. They held off and only issued a summary post later on this time, perhaps because the meeting was too painful. Luckily, my commenter Weekend Editor is excellent at summarizing such meetings, so I’ll quote their summary in full, link goes to their full post. After the summary, I’ll note the salient other details in the full post.

Today the FDA’s AMBAC meeting voted to recommend molnupiravir get an emergency use authorization. But just barely: 13 yes, 10 no, 0 abstentions. And with a lot of caveats among the yes votes.

Some issues:

(a) The efficacy vs hospitalization was “wobbly”: the interim report had 48.3% (CL: 20.5% – 66.5%) efficacy, but when they added the rest of the data it was only 30.4% (CL: 1.0% – 51.1%). People thought this might mean there were responder/nonresponder populations, and nobody knew a biomarker to distinguish them.

This is an interesting mix of good and bad thinking. Yes, it’s possible that the 50% vs. 30% thing wasn’t random, but why is that an issue here?

If you have a drug with 30% efficacy, that’s good. If you have a drug that is 50% effective half the time and 10% effective the other half, depending on who it is used on, and you can’t tell who is who, then you still have a 30% effective drug.

The difference is that because of the way you are now stacking your coin flips, it feels like you are now giving a 10% effective drug to some people without knowing it? Or there’s the ‘problem’ that if you knew more you could differentiate between the two populations, so now you have an imperfect procedure and it wouldn’t be ‘ethical’ somehow to proceed, or you’re blameworthy when the drug doesn’t work in a particular case?

I’m glad this didn’t cause a veto for enough voters, but consider that if this wasn’t an emergency situation, it might have caused one, despite being good news since it opens up the possibility of doing better once we know more.

Then again, perhaps the argument is that experiments are illegal except for getting drugs approved, so if we approved the drug we’d never run the experiment? Which has a certain kind of dystopian logic to it, I suppose.

Of course, same as it ever was, we stopped the study when the early results looked so good, and now we’re saying the results are at most barely good enough.

(b) While the mechanism and efficacy calculations from Merck were quite convincing, the FDA showed there were some issues with mutagenicity, particularly in the first trimester of pregnancy. (And I have to apologize: I got bored, and went grocery shopping at this point, stocking up in case the Omicronomicon gets loose. So I missed most of this, and am too tired to go back and listen. Maybe tomorrow.) Any woman will only get Molnupiravir after a negative pregnancy test.

Points for Omicronomicon, that’s great, and given the timing I hope it was all dry goods that will keep for a long time. In any case, a pregnancy test is quick and cheap, and damaged babies are considered quite bad, s...

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: FDA Votes on Molunpiravir, published by Zvi on December 1, 2021 on LessWrong.

While Paxlovid Remains Illegal and is expected to remain illegal for at least several weeks, the FDA did manage to finally meet to discuss whether or not to legalize the other Covid-19 treatment pill, Merck’s Molunpiravir. While later data reduced effectiveness estimates from 50% to 30%, that’s still much better than 0% and it uses a unique mechanism that can probably be profitably combined with other treatments, so one might naively think that after sufficient stalling for appearances this would be easy.

One would be wrong. The vote was 13-10, was restricted to those at high risk, and could easily have failed outright.

I may want to later refer back to this, so it’s splitting off into its own post.

High Level Summary

Usually we get live-blogs from Helen Branswell and Mattew Herper. They held off and only issued a summary post later on this time, perhaps because the meeting was too painful. Luckily, my commenter Weekend Editor is excellent at summarizing such meetings, so I’ll quote their summary in full, link goes to their full post. After the summary, I’ll note the salient other details in the full post.

Today the FDA’s AMBAC meeting voted to recommend molnupiravir get an emergency use authorization. But just barely: 13 yes, 10 no, 0 abstentions. And with a lot of caveats among the yes votes.

Some issues:

(a) The efficacy vs hospitalization was “wobbly”: the interim report had 48.3% (CL: 20.5% – 66.5%) efficacy, but when they added the rest of the data it was only 30.4% (CL: 1.0% – 51.1%). People thought this might mean there were responder/nonresponder populations, and nobody knew a biomarker to distinguish them.

This is an interesting mix of good and bad thinking. Yes, it’s possible that the 50% vs. 30% thing wasn’t random, but why is that an issue here?

If you have a drug with 30% efficacy, that’s good. If you have a drug that is 50% effective half the time and 10% effective the other half, depending on who it is used on, and you can’t tell who is who, then you still have a 30% effective drug.

The difference is that because of the way you are now stacking your coin flips, it feels like you are now giving a 10% effective drug to some people without knowing it? Or there’s the ‘problem’ that if you knew more you could differentiate between the two populations, so now you have an imperfect procedure and it wouldn’t be ‘ethical’ somehow to proceed, or you’re blameworthy when the drug doesn’t work in a particular case?

I’m glad this didn’t cause a veto for enough voters, but consider that if this wasn’t an emergency situation, it might have caused one, despite being good news since it opens up the possibility of doing better once we know more.

Then again, perhaps the argument is that experiments are illegal except for getting drugs approved, so if we approved the drug we’d never run the experiment? Which has a certain kind of dystopian logic to it, I suppose.

Of course, same as it ever was, we stopped the study when the early results looked so good, and now we’re saying the results are at most barely good enough.

(b) While the mechanism and efficacy calculations from Merck were quite convincing, the FDA showed there were some issues with mutagenicity, particularly in the first trimester of pregnancy. (And I have to apologize: I got bored, and went grocery shopping at this point, stocking up in case the Omicronomicon gets loose. So I missed most of this, and am too tired to go back and listen. Maybe tomorrow.) Any woman will only get Molnupiravir after a negative pregnancy test.

Points for Omicronomicon, that’s great, and given the timing I hope it was all dry goods that will keep for a long time. In any case, a pregnancy test is quick and cheap, and damaged babies are considered quite bad, s...