EA Forum Podcast (All audio)

Headline finding: I audited 17 AI Safety Talent programmes. Zero of 17 have published any comparison group, rejected-applicant follow-up, matched control or randomisation. Not one. Every programme that mentions a counterfactual does it by asking participants to self-report.

Background

At least 70 million dollars has gone into AI safety talent programmes so far, but we still can’t really say much about their impact. Notably, Kairos has just raised 50 million dollars on a self-described "shallow retrospective" that looked at "around 80 people". I decided to do a deep dive into how AI Safety Talent programmes self-evaluate their impact.

...

Grouper (Epinephelus sp.). Groupers can host a wide range of parasites, including anisakids and myxozoans. Photo: LDLP.

This post was originally published on my Substack. I’m sharing it here because parasite–host interactions in wild fish may have substantial and largely neglected welfare implications, while we still know remarkably little about the experiences of the animals involved. The post is intended mainly to highlight this empirical gap and some potentially tractable directions for further research.

Introduction

Cataracts cover much of both eyes of a fish who is still feeding and moving normally. A bird appr...

[...]

“Nearly every day, there is a frightening new story about how Big Tech companies are losing control of the technology they are developing, with potentially cataclysmic results,” Sanders said. “The leaders of the major AI companies publicly acknowledge that they do not fully understand the technology and that it is escaping their control. It is irresponsible for society to allow them to move forward and make these products even more advanced. That's why I am introducing legislation to immediately pause the development of increasingly powerful AI and ban the creation of systems that humanity cannot fully control — at home an...

These AIs are getting pretty smart. Maybe smart enough to help you think about your career? I was skeptical when first pitched on the project before joining 80,000 Hours. Turns out AI progress is a good train to hook yourself onto and AI has gotten pretty good at career advice.

It has some advantages compared to a human advisor:

Me and @Lily Curry launched an EA group at Binghamton University (a SUNY school with no prior EA presence) in August 2026. In two weeks we went from zero to ~500 Instagram followers, got 100+ tabling sign-ups, 59 people in our GroupMe, and ~40 people at our first general interest meeting. Our Instagram is now larger than the established EA groups at Harvard, Brown, Columbia, Princeton, Cornell, and NYU. This post is primarily about the factors we think led to our success.

James C. Scott's has this idea called legibility. Essentially, the part relevant to recruiting is the reversal of Scott's idea...

I’ve been feeling pretty shaken since the METR report about the Hugging Face incident came out last week. Over the weekend, I wrote up some thoughts on how lonely the AI situation sometimes feels to me. It's more personal than what I’d usually share publicly, but I thought I’d post it here in case it resonates with anyone else.

I’m very grateful to the many readers of this forum who dedicate so much of their time and energy to these problems.

—

When I was 12 years old, I found out that many...

At a conference on post-AGI futures earlier this year, I proposed a session titled “What's next for impact markets?” The conference speakers included the godfather of impact markets Paul Christiano himself, and noted supporters like Scott Alexander, Oli Habryka, and Andrew Critch — so I was hopeful that there would be some interest.

…

Zero people signed up for my session.

Before everyone writes these off as some EA fad of the early 2020s, or a pie-in-the-sky dream requiring impossible levels of funder buy-in: I’d like to give my account of what we’ve learned abou...

The Frame Fellowship is back for round two! In just about 10 days we'll have a brand-new cohort of 14 awesome content creators, storytellers, marketers, YouTubers, tiktokers- educating the world on AI safety through social media.

Frame 2.0 received hundreds of applications from award-winning filmmakers, investigative video journalists, AI Safety professors, former UN staff, Adobe professionals- all transitioning to make compelling content that reaches millions of people.

Somethings we looked for included exceptional communication ability, agency and passion for content, strong backgrounds/expertise, past AI safety engagement, distinct content styles, and how much they'd benefit from being part...

TL;DR: Kairos has raised 50 million dollars from Coefficient Giving for two years of funding, one of the largest commitments they’ve made towards AI safety fieldbuilding to date. We’re using this to make an ambitious push for growing Kairos, broadening our portfolio of talent infrastructure projects and incubating new organizations. We’ve doubled in size in the last six months, and we plan to double again in the next six. We now have open hiring rounds for ten roles on our team across events, group support, incubation, and special projects.

Two years in

Kairos...

TL;DR

We’re launching Efektif Berbagi, an organization dedicated to introducing and advancing effective giving in Indonesia.

Despite ranking several years in a row as the “most generous country in the world”, Indonesia currently has virtually no public advocacy or infrastructure directed towards cost-effective & high-impact giving. We’re working to be the bridge between Indonesia's vast philanthropic energy and the world's most pressing problems.

Why Indonesia?

For several years, Indonesia has placed #1 on the Charities Aid Foundation (CAF) World Giving Index, having an estimated ~90% donation rate and a ~65% volunteering rate. While th...

TLDR: Take the population ethics quiz here: https://mdickens.me/pop-ethics/

Population ethics is an oft-overlooked subfield within ethics. Many people hold views that they don't realize contradict each other, or that have strange implications that they wouldn't endorse if they thought about it more.

Not just that—population ethics is a BIG DEAL. A lot of ethical decisions hinge on how you think about changes in future populations.

I'll give concrete examples below, but reading the examples might change your answers, so you should take the quiz first. It has 13 to 17 short qu...

This post was written while participating in Gen Stream in London and done based on inspiration and with some advice from Roman Ross based on his work on helping rejected applicants upskill.

Introduction

There are way more applicants for AI safety fellowships, courses and projects than organizations have the capacity to accept. That means that for every fellowship that runs, hundreds and sometimes thousands of people will receive rejection letters. There is probably a better way that we could formulate rejections to make sure they have the proper resources to upskill and be ready for...

Reflections from the Effective Giving & Careers team.

Earlier this year we ran our first open call for the effective careers side of our work — a request for proposals from organizations that help people move into high-impact careers. We’re excited to share the results: We’re awarding $13,075,301 in grants to 28 organizations via the Effective Giving & Careers Fund, and several other Coefficient Giving funds expect to fund up to 16 more.

That ~13 million dollars is well above the ~5 million dollars we originally estimated, which reflects both the volume and the quality of applications. The RFP drew 474 applications. From t...

Last week, the Forum grappled with cluelessness. The volume of entries (76 essays!)—from newcomers and Forum power users alike—was a happy surprise, and there were many thoughtful contributions beyond those mentioned below.

Essay winners

First place ($2,500):

Joint second place ($1,500 each):

Magnus Vinding (aka ‘unsharp’), for Defusing cluelessness with degrees of justification

The judges—Andreas Mogensen, Jesse Clifton and Anthony DiGiovanni—decided against awarding the up-to-50 thousand dol...

Note: This post was crossposted from Planned Obsolescence by the Forum team, with the author's permission. The author may not see or respond to comments on this post.

Subtitle: It's a major warning shot, and might be the last one we get

This week, METR and Redwood Research published the report on our independent investigation into agents’ behavior and motivations in the Hugging Face attack; I was one of the investigators. This was an absolutely wild incident — I encourage you to check out the full report, but METR's tweet thread packs in some of the high...

Note: This post was crossposted from Planned Obsolescence by the Forum team, with the author's permission. The author may not see or respond to comments on this post.

Subtitle: It's a major warning shot, and might be the last one we get

All opinions are my personal view, and don’t represent my employer or fellow investigators.

This week, METR and Redwood Research published the report on our independent investigation into agents’ behavior and motivations in the Hugging Face attack; I was one of the investigators. This was an absolutely wild incident — I encour...

A preliminary estimate, and a request for better ones.

Summary

I believe the standard literature estimates for the number of DALYs attributable to a case of stunting are too low, largely because they don’t account for the long term effects. This means that childhood nutritional interventions that reduce the prevalence of stunting may be substantially more cost-effective than previously believed.

Epistemic status

Exploratory and back-of-envelope, ~100 hours invested in an area well outside my expertise. I have moderate-to-low confidence in any single point estimate, but high confidence that the standard literature va...

Overview

This post summarizes a new preprint on alternative proteins from the Humane and Sustainable Food Lab: New alternatives, same orders. We investigated 19 restaurants that introduced plant-based alternative proteins throughout 2016 to 2023, including Beyond sausage, the Impossible burger, and many others. We find essentially no evidence here that alternative proteins reduce meat purchases or those of other animal products, even for the same individual customers over time.

Interactive plots here.

Why this matters

A central theory of change for alt proteins is substitution: match meat on taste, price, and convenience, and consumers...

At the local AI safety co-working space, there are ~two kinds of regulars.

There's the kind of regular who's been thinking seriously about AI safety and alignment since pre-2022, who have passing to intimate familiarity with the funding ecosystem, the Sequences, and various conferences that happen at Lighthaven. Let's call them rationalists.

Then there's the kind of regular who comes in with many years of impressive industry or government experience, who realized in the last few years that it is important and worthwhile to pivot their career towards making sure that this AI thing is...

This is a linkpost for Is Extinction Risk Mitigation Uniquely Cost-Effective? Not in Standard Population Models by Gustav Alexandrie and Maya Eden, which was published on 18 August 2025 as chapter 20 of the book Essays on Longtermism: Present Action for the Distant Future. I got Claude Opus 5 to write the summary and 2 objections below after some iteration. The objections came from me. Gustav said the summary "looks correct from very quickly skimming it, though I may well have missed if there are mistakes". I also think it is accurate based on my read of the article. Thanks to Gustav for feedback...

Epistemic status: We are quite confident that Biological AI models (BAIMs) safety requires further work, but uncertain about its scale. The apparent gap may be refuted by one experiment, filled by a few researchers working for a year, or prove to be substantial enough to call for an entire subfield.

Disclaimer: this post has been written with a colleague, that due to her current job can't post out of her own forum account

What we’re looking for: please poke holes in this. In particular, we’d value:

Epistemic status: Speculation from two decently informed advocates armed with anecdata.

Note on process: After having some version of this conversation several times and saying, “we should probably write about this publicly,” we took the less heroic route: we recorded one of our conversations, fed the transcript into an LLM, and then substantially revised the structure, substance, and framing ourselves. We will not be sharing the transcript, as it is interspersed with side comments about dogs.

We’ve been very slow to get this out. We wrote this and began receiving feedback on our draft before...

I'm Molly Kovite, General Counsel at Coefficient Giving. I joined in 2021 as the organization's first lawyer and have since grown the team to (soon to be) seven lawyers and three legal admins.

Over the past couple of months, a number of AI safety organizations have asked me about whether they should hire in-house counsel and if so, what they should look for. This post is my attempt to write down what I've told them: how to tell when you've outgrown outside counsel, what in-house counsel does and doesn't get you, how to scope the role, and how...

Michael Nielsen has a beautiful new essay on moral imagination: the ability humans have to 'develop and transmit new notions of good action, indeed, even new kinds of good'.

As examples, he gives:

This personal post is not part of the competition, as my intent is not to produce an original constructive response to cluelessness. However, Toby Tremlett encouraged me to share it insofar as it contributes to this week's conversation.

TL; DR: I work in animal welfare and am clueless. To avoid choosing actions at random, I decided to write a decision procedure for myself and to indicate the assumptions that accompany it. This does not constitute a new argument on the topic. Rather, this is a layman's attempt to apply it. For those who want to skim, you...

Why I favor fallibilism over cluelessness

I’ve long been struck by a structural commonality between radical skepticism and cluelessness worries. Since the EA Forum has been discussing cluelessness this week, I figured I’d take another stab at explaining why I don’t find it a credible concern. (Too late to qualify for the official contest, alas.)

Skeptical Burdens

Open-minded people may be tempted by a discourse norm on which we should treat our interlocutors as epistemic equals in a sense that requires us to refrain from holding to conclusions that we cannot...

This is a cross-post from Astral Codex Ten, Scott Alexander's Substack. Original post here.

There are ACX meetup groups in cities around the world. Every fall and spring, I make a Meetups Everywhere post to try to signal-boost and support them. This summer, I want to focus on college effective altruist groups in particular.

Many colleges have effective altruist meetup groups or clubs. They range from fun social gatherings of like-minded individuals, to serious organizations working hard to build a better world; most are both at once. They have to replenish their ranks each fall...

In May 2025 I met Yo Shavit, who was working on national security policy at OpenAI and was thinking about how to prepare for a future in which models could seriously assist attackers in creating pandemics. We had a call, and when I shared notes with my team their main response was: "maybe start with not making models that can do that?"

Which is, in many ways, fair: by continuing to push the frontier in biological capabilities, OpenAI's actions were making things worse on many of the problems SecureBio is trying to solve. But OpenAI stopping wouldn't have...

I just read an article from The Guardian, written in April 2025 (I'm a bit behind on my reading ...). It's a good read, but that's noy why I share it here. What I want to highlight is something I'm seeing more and more, which is very negative portrayals and stereotyping of EA.

Some EA's seem ok with this, or accept that it goes with the territory. I vehemently disagree. I would argue that bringing EA thinking to the centre of the Overton window would be the single most impactful achievement of the 21st century, and that it is...

Sentient Futures Summit (SFS) Bay Area 2027 is a three-day conference taking place on March 12 to 14, 2027 @ Mox, San Francisco, exploring the impact of AI and other emerging technologies on sentient non-humans, both biological and potentially artificial. This is the 7th conference of our series, which has served as a premier meeting place for those working to ensure that non-human sentient beings are not left behind in the long-term future.

What to expect:

Summary

This post will summarize the trajectory of Wild Animal Initiative's field-building for wild animal welfare science so far, along with patterns and indicators we want to see more of going forward. Beginning with an initial phase focused on establishing credibility and awareness of the field, through a second phase that has placed more emphasis on the distinctiveness and broad priorities of the field, we are now moving into a third phase designed to consolidate effort and encourage more projects to build on each other, enabling the field make faster progress towards answers to key questions and...

I am writing up two crazy-ish charity ideas I’ve wanted to investigate, but haven’t been able to, in the hopes someone will investigate them further. If you end up investigating either, I'd be most interested to see what conclusions you come to!

Vulture preservation as a global health intervention

Since the mid 1990's the vulture population in India has been rapidly declining. The reason is due to a few chemicals which are fed to livestock, which is highly poisonous to vultures.

A paper by two University of Chicago economists came to the...

Prepared by: Claudinéia Pereira Costa, PhD; Meghan Barrett, PhD

Note: This report has not been peer-reviewed and has only been generated for the Forum. Results should be interpreted with caution, though we do follow peer-reviewed methodologies.

Executive Summary

Disclaimer: This blog post focuses on a new piece of research from Rethink Priorities. I was not involved with the funding or execution of this research, and therefore write this piece merely as a consumer. However, I am the Executive Director of Giving Green and a board member of Rethink Priorities, and acknowledge that these affiliations may bias my interpretation of the work.

TL;DR

Ever felt like something's missing from how EA talks about “doing good”? Like the conversation stops just short of asking what's actually broken at the root? This opening post introduces a sequence born from exactly that itch: a group of EA-adjacent volunteers who spent three years exploring whether there are systems-level dynamics and interventions that the community should be paying more attention to.

Post 1: In this first post of our sequence, we set the stage for the empirical and conceptual work to come: we outline who we are, how our collaboration came about, why we believe systemic chan...

I love EA; I am really proud of what we have built at CEA, and I am excited about where CEA is heading- that is all still very true. And after taking time to step back and think about where I most want to direct my energy next, I made the bittersweet decision to move on from CEA. I have spent my entire (!) full-time career there. I so deeply admire my stellar coworkers, and I deeply love EA and the impact we have- that is a big part of what made this decision so hard. I also still think...

This competition entry has been selected for publication by the Forum team.

Summary

Kollin, Clifton, DiGiovanni, and Macé (2025) propose two “bracketing” theories for recovering altruistic action guidance under cluelessness, formalizing a response that DiGiovanni (2025) had earlier viewed sympathetically but raised doubts about. On bottom-up bracketing, we set aside beneficiaries for whom it is indeterminate which option is better, and then aggregate over the remaining beneficiaries. On top-down bracketing, we instead evaluate the largest subsets of beneficiaries for which the comparison is determinate. Although bottom-up bracketing yields Kollin et al.'s intended verdict in the motivating donat...

This competition entry has been selected for publication by the Forum team.

1. Introduction

Here's Anthony DiGiovanni's unawareness argument, quoted from his summary post (footnotes omitted):

Let's say that we c-prefer A over B if the reason we prefer A is an impartial altruistic comparison of the actions’ possible consequences.

P1. Normative premise: To justify c-preferring A over B, it's not enough to say (e.g.) that A seems heuristically good. Rather, we need to argue that A has higher “expected value” broadly speaking, meaning: In some sense we “expect” that, if we were id...

Epistemic status: Wrote this pretty fast, probably ~70% confident some version of the thesis claim is true, would welcome being proven demonstrably wrong.

Context: I’m a grantmaker focused on alt-protein and am fairly well-connected/moderately in the loop of what animal welfare grantmakers are doing.

This post was significantly inspired by “You (Yes You) Should Have A Feb 2020 Checklist For AI Policy,” which is more geared for the existing AI safety crowd but directionally a good push in general.

Thesis: I think EA non-GCR grantmakers should probably be doing more to prepare for AGI/TA...

Summary