EA Forum Podcast (All audio)

Disclosure: I work at GiveDirectly. This is a linkpost summarizing findings from a pilot we ran in Rwanda. I used AI to assist in writing this post, and it's likely that >30% is AI-generated text.

View our blog and watch a video of recipients using AI here:

https://www.givedirectly.org/the-robots-work-at-night

Last year, GiveDirectly tested whether unrestricted access to an AI chatbot could complement cash transfers for recipients living in extreme poverty. Alongside our usual ~$1,000 one-time transfers in rural Rwanda, we offered 832 recipients access to a ChatGPT-powered chatbot via WhatsApp - a platform most already used - with no restrictions on what they could ask.

What we expected

We anticipated questions about the GiveDirectly program, help planning how to spend transfers, and basic business advice. People did use it for all of those things.

What actually happened

The more revealing pattern was how quickly recipients moved beyond program-specific questions. Across 21,000 inbound messages between November 2025 and April 2026, people used the chatbot the way people use AI everywhere: for family conflicts, sick children, market prices, and questions they couldn't easily take to anyone else. A few examples, translated verbatim from Kinyarwanda:

• "I have conflicts with the [...]

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

Losing GPS isn’t an X-risk, but would create a huge disaster on the scale of Covid-19 or bigger.

Hi! From 2020 - 2023 I was one of the early employees at Xona Space Systems, a company working on essentially a next-generation version of GPS. I ended up learning a lot about how GPS works and what it's used for, and (due to my personal interest in effective altruism) ended up doing some research into what would happen if today's GPS systems suddenly failed. This post is the product of that research. I discuss:

• What could kill GPS: it would be a tempting early target in a war between superpowers, or it could possibly be taken down by superhuman AI's cyber-hacking capabilities.  On the bright side, it's probably safe from even very large solar storms.
• If GPS was destroyed, how bad would this be?  I describe all the major areas in civilian life that would be disrupted (mostly summarizing some government reports).
• Then I try to crunch some very rough, vague numbers.  It looks like losing GPS would (by itself) be an economic hit to the USA perhaps equivalent to a few billion dollars a day.  On a very [...]

---

Outline:

(03:05) "Whoa, GPS comes from space?? I thought it was just a thing in my car..."

(05:56) To be clear, there are multiple satellite-navigation constellations

(09:02) What could kill GPS?

(10:34) Great-power war

(13:18) Hacking by superhuman AI???

(15:02) An unprecedented solar storm, maybe

(18:03) Other stuff

(19:40) What would break in the aftermath of losing GPS?

(21:49) Impact on the cell network & smartphones

(23:08) Impact on the power grid

(25:20) Impact on maritime industries

(26:41) Impact on travel and logistics

(29:03) Impacts on other industries

(30:46) Reckoning an overall cost per day

(36:15) Hard to say, but this feels approximately Covid-19-scale-ish

(38:30) How long would an outage last?

(39:36) Remember: you're still at war with China, or Russia, or AI, or possibly the Sun!

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

---

Images from the article:

Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

Cross-posted from my website.

Prior discussion: niplav's shortform (2025); Planning for Extreme AI Risks (2025) by Joshua Clymer

A frontier AI company (any one, I don't care which) should close shop and make an announcement along the lines of:

Powerful AI could end the human race. We are too worried that we don't know how to make this technology safe. We have decided to shut down because we don't want to be responsible for building the thing that kills us all.

A common refrain among safety-conscious AI developers: "it doesn't matter if we stop building dangerous AI, because someone else will just build it instead." Is that really true, though? If a multi-hundred-billion-dollar company comes out and says "We've concluded that our product is horribly dangerous, nobody knows how to make it safe, and there's too high a risk that it leads to human extinction", this won't raise any eyebrows? This has no chance of spurring policy-makers into action?

Shutting down would make people say, holy shit, they are serious about this extinction risk thing. Shutting down sends a strong signal to governments that they should pay serious attention to AI x-risk.

It [...]

The original text contained 4 footnotes which were omitted from this narration.

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

TLDR: The animal welfare movement in Africa has made significant progress over the last 5 years, but due to increasing commercialisation of animals on the continent and foreign interest in animal agriculture, now more than ever, it needs to accelerate. Our retrospective look at Europe's trajectory shows that reform did not stem from a gradual shift in attitudes. Instead, it was sparked by interconnected events over about 20 to 30 years. Events like public health and food safety crises involving animals fueled advocacy groups. These events led to policy wins and a change in the retail market. We think Africa can compress this timeline to 5–10 years by: actively and strategically pushing for institutional and policy recognition now rather than waiting for organic growth in public demand, capitalizing on non-animal welfare framing, recruiting mid-to-late career professionals now while focusing on education of early-to-mid career professionals, and, building strategic partnerships with universities, government agencies, and industry bodies to embed welfare into curricula, regulation, and market standards.

In our work, we routinely reflect on the status and ongoing development of Africa's animal welfare movement. In this, we have seen how small actions can lead to big wins for animals across [...]

---

Outline:

(01:22) Introduction

(04:26) Where the Analogy Holds, and Where It Breaks Down

(07:17) How the Change Started in Europe

(11:32) How Africa Can Skip This

(21:06) Conclusion

(21:58) Acknowledgment

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

In Rethink Priorities' (2021-2022) Moral Weights Project, they attempted to compare the welfare ranges of different farmed animals (which basically means comparing how many suffering shrimp are equal to one suffering chicken.) This project is the most comprehensive attempt at doing inter-species welfare comparisons to date, and it makes a very valuable contribution to the field.

That said, I think that many EAs have excessive confidence in these welfare ranges.[1] As such, in this post, I'm going to explain the project's basic methodology and some reasons to be skeptical of it.

Note: I am not a biologist. Others have written about this before, but I thought I would make this post since I've been thinking about the topic, and I think this view is worth re-iterating.

Welfare Range Estimates Chart

For reference, this chart is the welfare ranges being discussed:

The Basic Methodology

To create these moral weights, they calculated welfare ranges using three models:

1. The Equality Model
   1. According to this model, all species have equal welfare ranges. (So, a chicken and a pig's most intense experiences would be equal in intensity.)
2. The Neurophysiological Model
   1. According to this model, an animal's welfare range can be determined [...]

---

Outline:

(00:58) Welfare Range Estimates Chart

(01:19) The Basic Methodology

(02:46) Reasons to Be Skeptical

(06:22) Conclusion

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

---

Images from the article:

Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

In September 2025, I'd become increasingly convinced that a fieldbuilding program for content creators could solve a long-standing bottleneck of expanding reach and trust beyond the AI safety and EA bubble.

I had graduated from UCLA a few months earlier when I came across the AI-2027 report which had a significant impact on me. I rejected my six-figure tech consulting offer, left my place in LA, and moved to the first place I could find in the Bay Area to contribute to AI Safety. Contrary to my expectations, only a small subset of people were working towards what seemed to be far more important than building yet another AI B2B SaaS.

I wanted to change this. The power of a good external-facing comms effort is underrated. AI in Context, If Anyone Builds It Everyone Dies, Rob Miles is the first interaction with AI Safety for many, including mine. It has the power to empower mainstream audiences to act by creating trust and relatability, and meeting them where they’re at- social media. Moreover, I felt that I had the right background to do this- my experience as an activist, ex-founder building products for clients Google & [...]

---

Outline:

(02:29) The Fellowship

(06:24) Other Features of the program (with pics!)

(09:39) What we learnt

(11:01) Looking Ahead

(12:15) How to Get Involved

(13:34) How to Get Involved

(13:44) Early bird Applications are OPEN! APPLY

(13:57) Join the Team

(14:20) Partner with us

(14:39) Consider Donating

---

First published:

Source:

Linkpost URL:
https://www.lesswrong.com/posts/ALBdBRas7G3kdbPSq/the-quest-to-find-the-next-big-communicators-in-ai-safety

---

Narrated by TYPE III AUDIO.

---

Images from the article:

Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

I was recently at “Skoll”, the biggest NGO/social entrepreneurship conference. From one conversation to the next, two topics popped up over and over……. and over

1. Scale, Scale, Scale
2. Scale through Government, Government, Government

We at OneDay Health grapple with these questions: how then Shall we scale? And shall we through Government? Right now we operate 88 one-nurse health centers which save lives in remote rural healthcare black holes. Our one nurse in one room model can get good primary care to people cheap and fast. But to get good healthcare to everyone, we’ll need to do more than “grow”, but “scale”.

If we continue to launch more health centers every year through donor money, we’ll do plenty of good but can only grow our impact in a linear-ish way. From 100 to 150 to 200 health centers. Each year we’ll have to increase our donations and budget to increase our impact. Even if we succeed we’ll never reach tens of millions of people. So what is this “scale” business and how could we get there? Kevin Starr has outlined two key principles of scale.

1. “Scale is about exponential [...]

---

Outline:

(02:14) Three Concerns about Scale through Government

(02:55) Not the only way

(05:21) Should governments pay for the best interventions?

(05:57) Morally Dodgy

(07:10) Its really hard

(08:39) Poor Taxpayers not Excess Western Money now pay for the impact

---

First published:

Source:

Linkpost URL:
https://substack.com/home/post/p-201440232

---

Narrated by TYPE III AUDIO.

---

Images from the article:

Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

TLDR: Sarah Bluhm and I are funding and mentoring ideas-first (as opposed to, e.g., careers-first) EA community builders. If you’re at one of these universities or know someone who is, we want to talk to you, especially this subset:

• University of Toronto
• University of Michigan
• UCLA
• USC
• NYU
• Columbia
• Claremont Colleges (Claremont McKenna, Scripps, Harvey Mudd, Pomona, Pitzer)
• Stanford
• Harvard
• MIT
• Cornell
• Georgetown
• Northwestern

**Contact us here or at [email protected]**

The Network for EA Support & Training (NEST)

We (Matt Reardon and Sarah Bluhm) founded NEST, whose mission is to build vibrant in-person EA communities by teaching organizers the craft of communicating EA ideas with courage and authenticity.

I (Matt) am a former lawyer and 80k advisor who blogs, tweets, and podcasts about EA. Sarah worked on CEA's university groups team and has a background in bioengineering.

NEST's philosophy is that EA communities should be platforms for individuals to form their own views on the central questions of Effective Altruism: what are the most important issues in the world, why those issues, and what should I do about them?

This contrasts with approaches to community building that act as recruiting funnels for specific organizations or fields. We think [...]

---

Outline:

(00:56) The Network for EA Support & Training (NEST)

(02:19) University groups

(03:44) What we provide

(05:46) The next step

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

Deadline: July 8, 2026

Tentative program dates: August 15 – October 30

Know someone doing serious work on AI, animals, or the long-term future of sentient beings?

That someone might be you.

As a mentor, you can:

• Bring your own project idea and get matched with motivated fellows (think: short-term research assistant)
• Shape the next generation of researchers and practitioners in this space
• Join a growing network of 40+ mentors pushing the frontiers of nonhuman welfare

What's involved:

• Weekly 30-60 min call + light async support
• 10 weeks, fully remote

Apply by July 8 or visit here to learn more.

Feel free to share with anyone who'd be a good fit!

---

Outline:

(00:32) Know someone doing serious work on AI, animals, or the long-term future of sentient beings?

(01:10) Apply by July 8 or visit here to learn more.

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

Thanks to Arepo, David Thorstadt, Zeshen, and Michael st Jules for looking over this article. Disclaimer: I am not a subject matter expert and this is not a rigorous scientific article. This post is entirely human-written.

Introduction

In a previous article, I wrote a general introduction to the optimizer's curse, the phenomenon where a combination of random errors and ranking of causes results in the overestimation of top causes and a biasing of results towards uncertain causes. In that article, I mainly focussed on areas such as global health charities, where estimates or errors can be derived from the results of reasonably strong empirical evidence such as randomised control trials.

In this post, I want to explore how the optimizer's curse might affect tasks that are in the domain of extreme uncertainty. Specifically, I will try and model how the curse could manifest in the ranking of estimates of existential risk, a subject where empirical evidence is thin and experts disagree by orders of magnitude.

A standard way of ranking existential threats is to take a bunch of threats that have been raised as potentially deadly to humanity, and then for each one of them [...]

---

Outline:

(08:09) Part 1: Building a model

(25:19) Part 2: Interrogating the model

(40:15) Part 3: Objections

(57:26) Summary and Conclusion:

---

First published:

Source:

Linkpost URL:
https://titotal.substack.com/p/are-top-existential-risk-estimates

---

Narrated by TYPE III AUDIO.

---

Images from the article:

Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.