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: “Biological anchors” is about bounding, not pinpointing, AI timelines, published by Holden Karnofsky on November 18, 2021 on The Effective Altruism Forum.

I previously summarized Ajeya Cotra’s “biological anchors” method for forecasting for transformative AI, aka “Bio Anchors.” Here I want to try to clarify why I find this method so useful, even though I agree with the majority of the specific things I’ve heard people say about its weaknesses (sometimes people who can’t see why I’d put any stock in it at all).

A couple of preliminaries:

This post is probably mostly of interest for skeptics of Bio Anchors, and/or people who feel pretty confused/agnostic about its value and would like to see a reply to skeptics.

I don’t want to give the impression that I’m leveling new criticisms of “Bio Anchors” and pushing for a novel reinterpretation. I think the author of “Bio Anchors” mostly agrees with what I say both about the report’s weaknesses and about how to best use it (and I think the text of the report itself is consistent with this).

Summary of what the framework is about

Just to re-establish context, here are some key quotes from my main post about biological anchors:

The basic idea is:

Modern AI models can "learn" to do tasks via a (financially costly) process known as "training." You can think of training as a massive amount of trial-and-error. For example, voice recognition AI models are given an audio file of someone talking, take a guess at what the person is saying, then are given the right answer. By doing this millions of times, they "learn" to reliably translate speech to text. More: Training.

The bigger an AI model and the more complex the task, the more the training process [or “training run”] costs. Some AI models are bigger than others; to date, none are anywhere near "as big as the human brain" (what this means will be elaborated below). More: Model size and task type.

The biological anchors method asks: "Based on the usual patterns in how much training costs, how much would it cost to train an AI model as big as a human brain to perform the hardest tasks humans do? And when will this be cheap enough that we can expect someone to do it?" More: Estimating the expense.

...The framework provides a way of thinking about how it could be simultaneously true that (a) the AI systems of a decade ago didn't seem very impressive at all; (b) the AI systems of today can do many impressive things but still feel far short of what humans are able to do; (c) the next few decades - or even the next 15 years - could easily see the development of transformative AI.

Additionally, I think it's worth noting a couple of high-level points from Bio Anchors that don't depend on quite so many estimates and assumptions:

In the coming decade or so, we're likely to see - for the first time - AI models with comparable "size" to the human brain.

If AI models continue to become larger and more efficient at the rates that Bio Anchors estimates, it will probably become affordable this century to hit some pretty extreme milestones - the "high end" of what Bio Anchors thinks might be necessary. These are hard to summarize, but see the "long horizon neural net" and "evolution anchor" frameworks in the report.

One way of thinking about this is that the next century will likely see us go from "not enough compute to run a human-sized model at all" to "extremely plentiful compute, as much as even quite conservative estimates of what we might need." Compute isn't the only factor in AI progress, but to the extent other factors (algorithms, training processes) became the new bottlenecks, there will likely be powerful incentives (and multiple decades) to resolve them.

Things I agree with about the framework’s weaknesses/limitations

Bio Anchors “acts as if” AI will be developed in a particular way, and...

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: “Biological anchors” is about bounding, not pinpointing, AI timelines, published by Holden Karnofsky on November 18, 2021 on The Effective Altruism Forum.

I previously summarized Ajeya Cotra’s “biological anchors” method for forecasting for transformative AI, aka “Bio Anchors.” Here I want to try to clarify why I find this method so useful, even though I agree with the majority of the specific things I’ve heard people say about its weaknesses (sometimes people who can’t see why I’d put any stock in it at all).

A couple of preliminaries:

This post is probably mostly of interest for skeptics of Bio Anchors, and/or people who feel pretty confused/agnostic about its value and would like to see a reply to skeptics.

I don’t want to give the impression that I’m leveling new criticisms of “Bio Anchors” and pushing for a novel reinterpretation. I think the author of “Bio Anchors” mostly agrees with what I say both about the report’s weaknesses and about how to best use it (and I think the text of the report itself is consistent with this).

Summary of what the framework is about

Just to re-establish context, here are some key quotes from my main post about biological anchors:

The basic idea is:

Modern AI models can "learn" to do tasks via a (financially costly) process known as "training." You can think of training as a massive amount of trial-and-error. For example, voice recognition AI models are given an audio file of someone talking, take a guess at what the person is saying, then are given the right answer. By doing this millions of times, they "learn" to reliably translate speech to text. More: Training.

The bigger an AI model and the more complex the task, the more the training process [or “training run”] costs. Some AI models are bigger than others; to date, none are anywhere near "as big as the human brain" (what this means will be elaborated below). More: Model size and task type.

The biological anchors method asks: "Based on the usual patterns in how much training costs, how much would it cost to train an AI model as big as a human brain to perform the hardest tasks humans do? And when will this be cheap enough that we can expect someone to do it?" More: Estimating the expense.

...The framework provides a way of thinking about how it could be simultaneously true that (a) the AI systems of a decade ago didn't seem very impressive at all; (b) the AI systems of today can do many impressive things but still feel far short of what humans are able to do; (c) the next few decades - or even the next 15 years - could easily see the development of transformative AI.

Additionally, I think it's worth noting a couple of high-level points from Bio Anchors that don't depend on quite so many estimates and assumptions:

In the coming decade or so, we're likely to see - for the first time - AI models with comparable "size" to the human brain.

If AI models continue to become larger and more efficient at the rates that Bio Anchors estimates, it will probably become affordable this century to hit some pretty extreme milestones - the "high end" of what Bio Anchors thinks might be necessary. These are hard to summarize, but see the "long horizon neural net" and "evolution anchor" frameworks in the report.

One way of thinking about this is that the next century will likely see us go from "not enough compute to run a human-sized model at all" to "extremely plentiful compute, as much as even quite conservative estimates of what we might need." Compute isn't the only factor in AI progress, but to the extent other factors (algorithms, training processes) became the new bottlenecks, there will likely be powerful incentives (and multiple decades) to resolve them.

Things I agree with about the framework’s weaknesses/limitations

Bio Anchors “acts as if” AI will be developed in a particular way, and...

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: We need alternatives to Intro EA Fellowships, published by Ashley Lin on November 19, 2021 on The Effective Altruism Forum.

Thanks to Kuhan Jeyapragasan, Akash Wasil, Olivia Jimenez, and Lizka Vaintrob for helpful comments / conversations.

TLDR: I think group organizers have become anchored on the idea of Intro EA Fellowships as 8-week small-group things, which might actually be a sub-optimal way to introduce promising students to the EA world. We need new alternatives that are exciting, immersive, and enable EA-interested students to move as quickly as they’d like through the EA funnel.

The Intro EA Fellowship (also known as the Arete Fellowship) is a program where a small group of fellows and a facilitator meet multiple times to learn about some aspect of effective altruism. Stanford EA’s virtual Intro EA Fellowship was the first structured EA program I was part of and when I realized EA was a thriving community with real humans in it. I don’t think my experience is unique. In uni EA groups around the world, the Intro EA Fellowship is one of the core programs offered to students.

Some context on Intro EA Fellowships:

Fellowships are usually structured in cohorts of 2-5 fellows and a facilitator who meets weekly for 6-10 weeks.

Each week, fellows do some amount of readings and participate in a discussion with their cohort.

There are often also social activities that allow fellows to get to know each other better and cultivate friendships.

Uni group organizers often use Intro EA Fellowships as an early/mid part of the funnel for highly-engaged EA students.

Intro EA Fellowships exist in so many places because they have some upsides, some of which I list below.

That said, I also think there are some strong downsides to the Intro EA Fellowship model as it currently exists. As a participant, I wasn’t particularly impressed by my Intro EA Fellowship experience -- I’m not sure if my fellowship cohort actually finished (people got busy) and think there’s a chance I would’ve bounced off the EA community had I not attended a summer EA retreat for high school students a couple months later. Now, as I help organize Penn EA’s Intro Fellowship cohort, I’m noticing how my frustrations as an Intro Fellowship participant weren’t unique to me.

In this post, I share some of those frustrations (downsides to the Intro Fellowship as I see them). I try to steelman the argument for why Intro EA fellowships should exist. Finally, I introduce some Intro EA Fellowship alternatives that I'd be excited to see uni group organizers prototype. I’m personally planning to try some of these -- if you’d like to coordinate, please reach out [email protected]!

Downsides of Intro Fellowships

Much of this is observed through my own experience being part of / facilitating fellowships. I also think Intro Fellowship experiences are highly variable depending on one’s cohort and facilitator, and I think there’s a good chance I’ve had a relatively worse Intro EA Fellowship experience than others:

The standard 8-week fellowship timeline is too slow for people who are really excited early on and want to move faster. In these scenarios, the fellowship might actually slow people down and cause their excitement to fade (an hour-long conversation and some readings each week is a pretty sluggish pace) -- worse case scenario, it might cause some promising people to lose interest.

I want to cultivate a fast-paced vibe among fellows, instead of a discussion-group like vibe. (To me “fast-paced,” when applied to something that seems interesting, cultivates genuine excitement and deep curiosity). For example, when I first found 80,000 Hours, I pulled an all-nighter reading it and was absolutely ecstatic that people had spent so much time thinking about triaging problems and how one can do the most good in the world...

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: We need alternatives to Intro EA Fellowships, published by Ashley Lin on November 19, 2021 on The Effective Altruism Forum.

Thanks to Kuhan Jeyapragasan, Akash Wasil, Olivia Jimenez, and Lizka Vaintrob for helpful comments / conversations.

TLDR: I think group organizers have become anchored on the idea of Intro EA Fellowships as 8-week small-group things, which might actually be a sub-optimal way to introduce promising students to the EA world. We need new alternatives that are exciting, immersive, and enable EA-interested students to move as quickly as they’d like through the EA funnel.

The Intro EA Fellowship (also known as the Arete Fellowship) is a program where a small group of fellows and a facilitator meet multiple times to learn about some aspect of effective altruism. Stanford EA’s virtual Intro EA Fellowship was the first structured EA program I was part of and when I realized EA was a thriving community with real humans in it. I don’t think my experience is unique. In uni EA groups around the world, the Intro EA Fellowship is one of the core programs offered to students.

Some context on Intro EA Fellowships:

Fellowships are usually structured in cohorts of 2-5 fellows and a facilitator who meets weekly for 6-10 weeks.

Each week, fellows do some amount of readings and participate in a discussion with their cohort.

There are often also social activities that allow fellows to get to know each other better and cultivate friendships.

Uni group organizers often use Intro EA Fellowships as an early/mid part of the funnel for highly-engaged EA students.

Intro EA Fellowships exist in so many places because they have some upsides, some of which I list below.

That said, I also think there are some strong downsides to the Intro EA Fellowship model as it currently exists. As a participant, I wasn’t particularly impressed by my Intro EA Fellowship experience -- I’m not sure if my fellowship cohort actually finished (people got busy) and think there’s a chance I would’ve bounced off the EA community had I not attended a summer EA retreat for high school students a couple months later. Now, as I help organize Penn EA’s Intro Fellowship cohort, I’m noticing how my frustrations as an Intro Fellowship participant weren’t unique to me.

In this post, I share some of those frustrations (downsides to the Intro Fellowship as I see them). I try to steelman the argument for why Intro EA fellowships should exist. Finally, I introduce some Intro EA Fellowship alternatives that I'd be excited to see uni group organizers prototype. I’m personally planning to try some of these -- if you’d like to coordinate, please reach out [email protected]!

Downsides of Intro Fellowships

Much of this is observed through my own experience being part of / facilitating fellowships. I also think Intro Fellowship experiences are highly variable depending on one’s cohort and facilitator, and I think there’s a good chance I’ve had a relatively worse Intro EA Fellowship experience than others:

The standard 8-week fellowship timeline is too slow for people who are really excited early on and want to move faster. In these scenarios, the fellowship might actually slow people down and cause their excitement to fade (an hour-long conversation and some readings each week is a pretty sluggish pace) -- worse case scenario, it might cause some promising people to lose interest.

I want to cultivate a fast-paced vibe among fellows, instead of a discussion-group like vibe. (To me “fast-paced,” when applied to something that seems interesting, cultivates genuine excitement and deep curiosity). For example, when I first found 80,000 Hours, I pulled an all-nighter reading it and was absolutely ecstatic that people had spent so much time thinking about triaging problems and how one can do the most good in the world...

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: [Linkpost] Don't Look Up - a Netflix comedy about asteroid risk and realistic societal reactions (Dec. 24th), published by Linch on November 18, 2021 on The Effective Altruism Forum.

Starring Leonardo DiCaprio and Jennifer Lawrence.

Judging from the trailer alone, I don't think I've seen a better fictional movie about xrisks to date.

The movie is scheduled to be released in theatres on December 10, 2021, prior to streaming on Netflix on December 24, 2021 (H/T Eliezer Yudkowsky).

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: [Linkpost] Don't Look Up - a Netflix comedy about asteroid risk and realistic societal reactions (Dec. 24th), published by Linch on November 18, 2021 on The Effective Altruism Forum.

Starring Leonardo DiCaprio and Jennifer Lawrence.

Judging from the trailer alone, I don't think I've seen a better fictional movie about xrisks to date.

The movie is scheduled to be released in theatres on December 10, 2021, prior to streaming on Netflix on December 24, 2021 (H/T Eliezer Yudkowsky).

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: Goodhart: Endgame, published by Charlie Steiner on November 19, 2021 on The AI Alignment Forum.

(This work was supported by CEEALAR. Thanks to Michele Campolo and others for conversations.)

I - Digression, aggression

The perspective arrived at in this sequence is almost aggressively third-person. Sure, I allow that you can have first-person stories about what you value, but these stories only rise in prominence to an outside observer if they're supported by cold, hard, impersonal facts. What's more, there can be multiple first-person stories that fit the data, and these different stories don't have to be aggregated as if we're under Bayesian uncertainty about which one is the One True story, they can all be acceptable models of the same physical system.

This can be unintuitive and unpalatable to our typical conception of ourselves. The assumption that there's a unique fact of the matter about what we're feeling and desiring at any given time is so convenient as to be almost inescapable. But if even knowing the exact state of the my brain leaves room for different inferences about what I want, just like we can impute different sets of preferences to a thermostat, then our raw intuition that we have just one set of True Values is no foundation to build a value learning AI on.

This sequence, which is nominally about Goodhart's law, has also been a sidelong way of attacking the whole problem of value learning. Contrary to what we might have intended when we first heard about the problem, the solution is not to figure out what and where human values Truly are, and then use this to do value learning. The solution is to figure out how to do value learning without needing True Values at all.

If you are one of the people currently looking for the True Values, this is where I'm calling you out. For your favorite value-detection procedure, I'd suggest you nail down how this procedure models humans, and why that way is good, and what the advantages and disadvantages are of the reasonable alternatives. I am happy to provide specific suggestions or argue with you in the comments.

II - Summing up

In the penultimate post we rephrased some mechanisms of Goodhart's law in terms of modeled human preferences rather than True Values. The old style arguments would give some reason why models give different recommendations in some circumstances, and then say "the choices get pushed away from those recommended by the True Values, which is Goodhart's law." We replaced these with new style arguments, optimized for the issues that comes up when trying to learn human values, that instead say "the choices recommended by different models get pushed away from each other, which is Goodhart's law."

While writing these posts, I've had a joke stuck in my head. A patient goes to the doctor and says "Doctor, it hurts when I do this." The doctor says "Then don't do that!"

The goal is to be in the situation of the patient. If the "non-obvious" sort of Goodhart's law kicks in when we go to a part of state space where otherwise-good models of us get pushed away from each other, can't we just... not do that?

Unfortunately we can't translate that into code, because I've stuffed the unsolved problems into that qualifier "otherwise-good" (we'll get back to that). But whatever route we take to end up with an AI that makes the future go well, it's got to "not do that" somehow. This is interesting to me because it seems like such a weird and kludgy property to ask for if you assume that humans have True Values. And yet it makes perfect sense if you treat us as physical systems that admit multiple interpretations.

Because we can't translate it into code, this sequence is a failure at sketching a solution to Goodhart's law. I hear you cry: "What have you been doing this whole time, then, if not sketching a solution?...

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: Goodhart: Endgame, published by Charlie Steiner on November 19, 2021 on The AI Alignment Forum.

(This work was supported by CEEALAR. Thanks to Michele Campolo and others for conversations.)

I - Digression, aggression

The perspective arrived at in this sequence is almost aggressively third-person. Sure, I allow that you can have first-person stories about what you value, but these stories only rise in prominence to an outside observer if they're supported by cold, hard, impersonal facts. What's more, there can be multiple first-person stories that fit the data, and these different stories don't have to be aggregated as if we're under Bayesian uncertainty about which one is the One True story, they can all be acceptable models of the same physical system.

This can be unintuitive and unpalatable to our typical conception of ourselves. The assumption that there's a unique fact of the matter about what we're feeling and desiring at any given time is so convenient as to be almost inescapable. But if even knowing the exact state of the my brain leaves room for different inferences about what I want, just like we can impute different sets of preferences to a thermostat, then our raw intuition that we have just one set of True Values is no foundation to build a value learning AI on.

This sequence, which is nominally about Goodhart's law, has also been a sidelong way of attacking the whole problem of value learning. Contrary to what we might have intended when we first heard about the problem, the solution is not to figure out what and where human values Truly are, and then use this to do value learning. The solution is to figure out how to do value learning without needing True Values at all.

If you are one of the people currently looking for the True Values, this is where I'm calling you out. For your favorite value-detection procedure, I'd suggest you nail down how this procedure models humans, and why that way is good, and what the advantages and disadvantages are of the reasonable alternatives. I am happy to provide specific suggestions or argue with you in the comments.

II - Summing up

In the penultimate post we rephrased some mechanisms of Goodhart's law in terms of modeled human preferences rather than True Values. The old style arguments would give some reason why models give different recommendations in some circumstances, and then say "the choices get pushed away from those recommended by the True Values, which is Goodhart's law." We replaced these with new style arguments, optimized for the issues that comes up when trying to learn human values, that instead say "the choices recommended by different models get pushed away from each other, which is Goodhart's law."

While writing these posts, I've had a joke stuck in my head. A patient goes to the doctor and says "Doctor, it hurts when I do this." The doctor says "Then don't do that!"

The goal is to be in the situation of the patient. If the "non-obvious" sort of Goodhart's law kicks in when we go to a part of state space where otherwise-good models of us get pushed away from each other, can't we just... not do that?

Unfortunately we can't translate that into code, because I've stuffed the unsolved problems into that qualifier "otherwise-good" (we'll get back to that). But whatever route we take to end up with an AI that makes the future go well, it's got to "not do that" somehow. This is interesting to me because it seems like such a weird and kludgy property to ask for if you assume that humans have True Values. And yet it makes perfect sense if you treat us as physical systems that admit multiple interpretations.

Because we can't translate it into code, this sequence is a failure at sketching a solution to Goodhart's law. I hear you cry: "What have you been doing this whole time, then, if not sketching a solution?...

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: 2021 ALLFED Highlights, published by Ross Tieman on November 17, 2021 on The Effective Altruism Forum.

Summary

The Alliance to Feed the Earth in Disasters (ALLFED) team recently connected with the EA community in London at EA Global. Thank you for your interest in and support of our work at the intersection of food systems resilience and global catastrophic risks, such as nuclear winter and tail risks from climate change.

Our work is inherently interdisciplinary, and our goal of global food systems resilience requires wide-ranging expertise and perspectives. Whether you’ve been following our work for a while or are new to what ALLFED does, we hope you dive in by checking out some of our biggest accomplishments from 2021 (here are the 2020 highlights).

We have nearly completed our prioritization of the most promising resilient foods (formerly alternative foods) that could be scaled up quickly and would be low cost. We are now ready to facilitate pilot demonstrations of scaling up these foods quickly, such as repurposing a paper factory to produce food. Therefore, our room for more funding is ~$100 million USD over the next 5 years, while smaller amounts would support our efforts to find financial mechanisms to fund pilots, small-scale demonstration of leaf protein concentrate from nettle leaves in Nepal, country-by-country resilient food analyses, and research on backup plans for scenarios that could disrupt electricity/industry, such as an extreme pandemic causing people to be too fearful to show up to work at critical industries. Our work has been estimated to be highly cost effective from both the near-term and long-term perspectives.

If you value our work and want to help us contribute to global food system resilience, please consider supporting ALLFED! We need additional funding to maintain and grow our projects. In particular, we encourage you to consider participating in two matching opportunities:

The Every.org Fall Giving Challenge is currently underway and won’t last for much longer (see a realtime dashboard of remaining funds here). Please message David if you would like to double your $100 donation to almost any charity, by donating to ALLFED in exchange for David donating to the charity of your choice.

Facebook’s Giving Tuesday match is coming up on November 30, 2021 (for instructions and important updates, please sign up at EA Giving Tuesday).

New resources on our work

We have submitted or are about to submit 10 articles to peer-reviewed journals this year.

Our new website puts forward our latest research on resilient food solutions, and is a growing repository of learning resources for catastrophic risks to food.

We have a new video on resource-efficient resilient foods that could not only remove barriers facing astronauts undertaking deep-space missions, but could also make refuges more economical and help feed more people on Earth today or in a catastrophe.

David Denkenberger returned to the 80,000 Hours Podcast (episode forthcoming) to revisit ALLFED’s work after his 2018 interview with Robert Wiblin, which is featured in the new 80,000 Hours Podcast series Effective Altruism: 10 Global Problems.

On our Twitter we regularly shared updates on our work, alongside new research and articles on global catastrophic risks, existential threats, resilient foods, and food security.

The latest Global Challenges Foundation report featured our work on resilient foods as part of its exploration into the challenges posed by several catastrophic threats.

Our work is explored in a forthcoming episode of the Clearer Thinking podcast with Spencer Greenberg.

Operations and internal resilience

Operationally, we are in the final stages of a major systems overhaul and a new CRM implementation.

In 2021, we have further built our organizational resilience and response capaci...

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: 2021 ALLFED Highlights, published by Ross Tieman on November 17, 2021 on The Effective Altruism Forum.

Summary

The Alliance to Feed the Earth in Disasters (ALLFED) team recently connected with the EA community in London at EA Global. Thank you for your interest in and support of our work at the intersection of food systems resilience and global catastrophic risks, such as nuclear winter and tail risks from climate change.

Our work is inherently interdisciplinary, and our goal of global food systems resilience requires wide-ranging expertise and perspectives. Whether you’ve been following our work for a while or are new to what ALLFED does, we hope you dive in by checking out some of our biggest accomplishments from 2021 (here are the 2020 highlights).

We have nearly completed our prioritization of the most promising resilient foods (formerly alternative foods) that could be scaled up quickly and would be low cost. We are now ready to facilitate pilot demonstrations of scaling up these foods quickly, such as repurposing a paper factory to produce food. Therefore, our room for more funding is ~$100 million USD over the next 5 years, while smaller amounts would support our efforts to find financial mechanisms to fund pilots, small-scale demonstration of leaf protein concentrate from nettle leaves in Nepal, country-by-country resilient food analyses, and research on backup plans for scenarios that could disrupt electricity/industry, such as an extreme pandemic causing people to be too fearful to show up to work at critical industries. Our work has been estimated to be highly cost effective from both the near-term and long-term perspectives.

If you value our work and want to help us contribute to global food system resilience, please consider supporting ALLFED! We need additional funding to maintain and grow our projects. In particular, we encourage you to consider participating in two matching opportunities:

The Every.org Fall Giving Challenge is currently underway and won’t last for much longer (see a realtime dashboard of remaining funds here). Please message David if you would like to double your $100 donation to almost any charity, by donating to ALLFED in exchange for David donating to the charity of your choice.

Facebook’s Giving Tuesday match is coming up on November 30, 2021 (for instructions and important updates, please sign up at EA Giving Tuesday).

New resources on our work

We have submitted or are about to submit 10 articles to peer-reviewed journals this year.

Our new website puts forward our latest research on resilient food solutions, and is a growing repository of learning resources for catastrophic risks to food.

We have a new video on resource-efficient resilient foods that could not only remove barriers facing astronauts undertaking deep-space missions, but could also make refuges more economical and help feed more people on Earth today or in a catastrophe.

David Denkenberger returned to the 80,000 Hours Podcast (episode forthcoming) to revisit ALLFED’s work after his 2018 interview with Robert Wiblin, which is featured in the new 80,000 Hours Podcast series Effective Altruism: 10 Global Problems.

On our Twitter we regularly shared updates on our work, alongside new research and articles on global catastrophic risks, existential threats, resilient foods, and food security.

The latest Global Challenges Foundation report featured our work on resilient foods as part of its exploration into the challenges posed by several catastrophic threats.

Our work is explored in a forthcoming episode of the Clearer Thinking podcast with Spencer Greenberg.

Operations and internal resilience

Operationally, we are in the final stages of a major systems overhaul and a new CRM implementation.

In 2021, we have further built our organizational resilience and response capaci...