EA Forum Podcast (All audio)

I sometimes worry that focus on effectiveness creates perverse incentives in strategic settings, leading us to become less effective. Here are a few observations illustrating this concern.

Effectiveness-focused advocacy creates perverse incentives for adversaries

When we conduct cage-free campaigns, the target companies frequently ask us why they are being targeted instead of some other company. While trying to answer that, one immediately realises the following tension. If we say that "because targeting you is the most effective thing we can do", we incentivise them to not budge. Because they will know that willingness to compromise invites more aggression.

When dealing with an effectiveness-focused movement, our adversaries are further incentivised to prevent concrete results. While other movements will have to be destroyed through pressure, an effectiveness-focused movement will easily go away if you just prove to them that they can be more effective elsewhere.

For that reason, in our [...]

---

Outline:

(00:21) Effectiveness-focused advocacy creates perverse incentives for adversaries

(01:20) Sometimes you have to fight back even when it's not the most effective thing to do

(02:05) Religious rules survive because they are stubborn

(02:39) We might need more commitment devices

---

First published:

April 8th, 2025

Source:

https://forum.effectivealtruism.org/posts/tvcEa9LhsX264NZbf/maybe-we-aren-t-stubborn-enough

---

Narrated by TYPE III AUDIO.

*Disclaimer: This post was written in a personal capacity and reflects my views, not necessarily those of my employer. I have not actually attended policy school (I plan to update this post once I have), so please read with a grain of salt.*

I recently applied to a few policy schools. Despite having read a bunch of great resources before applying, I learned a few things along the way that I wish I would have known from the outset. This post is a summary of what was new or unexpected to me. If you’re considering policy school and haven’t read Horizon's excellent posts yet, I’d encourage you to read them. I would also highly recommend joining Horizon's email newsletter.

The case for policy school

• At some programs, you get paid to attend.
  • Princeton gives all admitted students full tuition, healthcare, a $35K/year stipend, and $850/week for the [...]

---

Outline:

(00:56) The case for policy school

(05:28) The case against policy school

(06:04) Tips for applying

(06:25) Know your audience

(07:10) Provide social proof

(07:35) Tailor examples for recommenders

(08:29) Managing EA affiliation

(09:14) Using LLMs

(10:17) Krystin Borgognone is very helpful

(11:21) The value of applying wide

The original text contained 5 footnotes which were omitted from this narration.

---

First published:

April 7th, 2025

Source:

https://forum.effectivealtruism.org/posts/TKjFRKua5eBveLXCa/untitled-draft-7y43

---

Narrated by TYPE III AUDIO.

Summary

• My new book with Mike Geruso on global depopulation, After the Spike: Population, Progress, and the Case for People, will be released this summer.
• If you’re here reading this, we bet you’re going to like the book. If you’d like to help others find their way to the book, preorders now are the best way to help. Preorders give an important market signal to booksellers and help books rise to wider attention via bestseller lists.
• You can preorder here: Amazon | Books-A-Million | Bookshop.org

About the book

After the Spike envisions, and is written to promote, an abundant future where people collaborate, give, and innovate to make lives better, longer, healthier, safer, and happier than they otherwise would be. It is concerned both with the wellbeing of the global poor today and with the long-term flourishing of humanity.

For decades, many have claimed that fewer people would be better—better for [...]

---

Outline:

(00:45) About the book

(03:15) If you would like to help

(04:36) What some friends and colleagues have said about the book:

---

First published:

April 7th, 2025

Source:

https://forum.effectivealtruism.org/posts/iFkm6EbGRyCQSA8hv/announcing-our-book-after-the-spike-and-an-opportunity-to

---

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.

This is a link post.

In recent months, the CEOs of leading AI companies have grown increasingly confident about rapid progress:

• OpenAI's Sam Altman: Shifted from saying in November "the rate of progress continues" to declaring in January "we are now confident we know how to build AGI"
• Anthropic's Dario Amodei: Stated in January "I'm more confident than I've ever been that we're close to powerful capabilities... in the next 2-3 years"
• Google DeepMind's Demis Hassabis: Changed from "as soon as 10 years" in autumn to "probably three to five years away" by January.

What explains the shift? Is it just hype? Or could we really have Artificial General Intelligence (AGI) by 2028?[1]

In this article, I look at what's driven recent progress, estimate how far those drivers can continue, and explain why they're likely to continue for at least four more years.

In particular, while in 2024 progress [...]

---

Outline:

(03:53) In a nutshell

(05:35) 1. Whats driven recent AI progress? And will it continue?

(05:42) The deep learning era

(08:21) Whats coming up

(09:22) 1. Scaling pretraining to create base models with basic intelligence

(16:50) 2. Post training of reasoning models with reinforcement learning

(25:56) 3. Increasing how long models think

(28:54) 4. The next stage: building better agents

(35:01) How far can the trend of improving agents continue?

(37:18) 2. How good will AI become by 2030?

(37:23) The four drivers projected forwards

(39:10) Trend extrapolation of AI capabilities

(40:33) What jobs would these systems be able to help with?

(45:05) Whats the case against impressive AI progress by 2030?

(49:49) When do the experts expect AGI to arrive?

(51:22) 3. Why the next 5 years are crucial

(52:26) Bottlenecks around 2030

(56:13) Two potential futures for AI

(58:16) Conclusion

(59:32) Further reading

The original text contained 5 footnotes which were omitted from this narration.

---

First published:

April 6th, 2025

Source:

https://forum.effectivealtruism.org/posts/7EoHMdsy39ssxtKEW/the-case-for-agi-by-2030-1

Linkpost URL:
https://80000hours.org/agi/guide/when-will-agi-arrive/

---

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.

On January 1, 2030, there will be no artificial general intelligence (AGI) and AGI will still not be imminent.

A few reasons why I think this:

-If you look at easy benchmarks like ARC-AGI and ARC-AGI-2 that are easy for humans to solve and intentionally designed to be a low bar for AI to clear, the weaknesses of frontier AI models are starkly revealed.[1]

-Casual, everyday use of large language models (LLMs) reveals major errors on simple thinking tasks, such as not understanding that an event that took place in 2025 could not have caused an event that took place in 2024.

-Progress does not seem like a fast exponential trend, faster than Moore's law and laying the groundwork for an intelligence explosion. Progress seems actually pretty slow and incremental, with a moderate improvement from GPT-3.5 to GPT-4, and another moderate improvement from GPT-4 to o3-mini. The decline in [...]

---

First published:

April 6th, 2025

Source:

https://forum.effectivealtruism.org/posts/xkNjpGNfnAYmkFz3s/on-january-1-2030-there-will-be-no-agi-and-agi-will-still

---

Narrated by TYPE III AUDIO.

On January 1, 2030, there will be no artificial general intelligence (AGI) and AGI will still not be imminent.

A few reasons why I think this:

-If you look at easy benchmarks like ARC-AGI and ARC-AGI-2 that are easy for humans to solve and intentionally designed to be a low bar for AI to clear, the weaknesses of frontier AI models are starkly revealed.[1]

-Casual, everyday use of large language models (LLMs) reveals major errors on simple thinking tasks, such as not understanding that an event that took place in 2025 could not have caused an event that took place in 2024.

-Progress does not seem like a fast exponential trend, faster than Moore's law and laying the groundwork for an intelligence explosion. Progress seems actually pretty slow and incremental, with a moderate improvement from GPT-3.5 to GPT-4, and another moderate improvement from GPT-4 to o3-mini. The decline in [...]

---

First published:

April 6th, 2025

Source:

https://forum.effectivealtruism.org/posts/xkNjpGNfnAYmkFz3s/on-january-1-2030-there-will-be-no-agi-and-agi-will-still

---

Narrated by TYPE III AUDIO.

Our Mission: To build a multidisciplinary field around using technology—especially AI—to improve the lives of nonhumans now and in the future.

Overview

Background

This hybrid conference had nearly 550 participants and took place March 1-2, 2025 at UC Berkeley. It was organized by AI for Animals for $74k by volunteer core organizers Constance Li, Sankalpa Ghose, and Santeri Tani.

This conference has evolved since 2023:

• The 1st conference mainly consisted of philosophers and was a single track lecture/panel.
• The 2nd conference put all lectures on one day and followed it with 2 days of interactive unconference sessions happening in parallel and a week of in-person co-working.
• This 3rd conference had a week of related satellite events, free shared accommodations for 50+ attendees, 2 days of parallel lectures/panels/unconferences, 80 unique sessions, of which 32 are available on Youtube, Swapcard to enable 1:1 connections, and a Slack community to continue conversations year [...]

---

Outline:

(00:32) Overview

(00:35) Background

(02:27) Outcomes

(03:51) The Event(s)

(04:19) Speaking Sessions

(04:23) en-US-AvaMultilingualNeural__ Conference presentation scenes with speakers and audience in auditorium setting. The image shows a collage of different presentation scenes at what appears to be an academic or professional conference. The main image shows an auditorium with attendees watching a video presentation on a large screen. Other smaller images show speakers on stage in different settings - some seated in orange chairs for a panel discussion, others at podiums. Theres a logo visible for what appears to be a research institute in one of the frames. The setting appears academic/professional in nature, with proper presentation equipment, stage lighting, and organized seating arrangements typical of a conference or symposium event.

(05:10) Featured Talks

(10:13) Lightning Talks

(11:55) Interactive Sessions

(12:19) Unconferences

(13:22) Meetups

(13:36) Mapping Workshops

(13:44) Office Hours

(13:59) Topic Distribution

(14:18) Satellite Events

(14:41) en-US-AvaMultilingualNeural__ Events schedule showing AI and animal-focused activities from Feb 26-Mar 2.

(14:52) en-US-AvaMultilingualNeural__ People gathered at social events, networking and dining in various settings.

(15:02) Behind the Scenes

(15:10) Personnel

(16:04) Handbooks

(16:24) Finances

(19:21) Outreach

(21:17) Event Reflection

(24:12) Get Involved

---

First published:

April 5th, 2025

Source:

https://forum.effectivealtruism.org/posts/KWpyRXzHn6JMyZiBn/ai-for-animals-2025-bay-area-retrospective

---

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.

This is a link post.

This is a linkpost for Epoch After Hours' podcast episode Is it 3 Years, or 3 Decades Away? Disagreements on AGI Timelines by Ege Erdil and Matthew Barnett, which was released on 28 March 2025. The episode is on YouTube, Apple Podcasts, Spotify, Amazon Music, Pocket Casts, and Overcast. Ege is the 1st author of the paper about Epoch AI's compute-centric model of AI automation, which was announced on 21 March 2025, and also discussed on the Future of Life Institute Podcast. I broadly agree with Ege on AI timelines, and like Ege's and Matthew's takes on AI risk. See AI 2027 for the case for short AI timelines.

Blurb

When might AI truly transform our world, and what will that transformation look like? Even within our own research team, timelines for AGI differ substantially. In this episode, two Epoch AI researchers with relatively long [...]

---

Outline:

(01:02) Blurb

(02:48) AI timelines

---

First published:

April 4th, 2025

Source:

https://forum.effectivealtruism.org/posts/tJh4J3dLhB5AqvsK2/is-it-3-years-or-3-decades-away-disagreements-on-agi

Linkpost URL:
https://epoch.ai/epoch-after-hours/disagreements-on-agi-timelines

---

Narrated by TYPE III AUDIO.

Recent cuts to US government foreign assistance have destabilized global health programs, impacting some of the most cost-effective interventions we’ve found for saving and improving lives, such as malaria nets, malaria chemoprevention, and community-based management of acute malnutrition. This situation is a major focus of our research team at the moment, and we’re working to balance a targeted, near-term response to urgent needs with a broad, long-term perspective of needs that may emerge.

The US has historically provided roughly 20% to 25% ($12 billion to $15 billion) of the total global aid to support health programs, which combat malaria, tuberculosis, HIV, maternal and child health issues, and much more.[1] While the long-term effects remain uncertain and exact numbers remain difficult to ascertain, cuts of 35% to 90% of US foreign aid dollars are being publicly discussed by the administration.[2]

We’ve created a webpage to provide an overview of how [...]

The original text contained 2 footnotes which were omitted from this narration.

---

First published:

April 4th, 2025

Source:

https://forum.effectivealtruism.org/posts/u9ENn4GvbfJEnXkip/givewell-s-response-to-the-usaid-funding-cuts

---

Narrated by TYPE III AUDIO.

Hey EA Forum,

We're the EA group at Middlebury College in Vermont—three organizers (freshman, sophomore, junior) trying to figure out how to make EA stick on a small but ambitious liberal arts campus. We've run into the classic "attendance starts high, then drops off to like, two people and a cricket" scenario with our traditional reading groups and fellowships. Still, we know that Middlebury students have a whole lot of wealth and/or talent, and it's really important to expose them to EA ideas and organizations and balance out the consistent presence of consulting/finance/etc. pathways.

This year, we decided to shake things up. We invited organizations to campus instead, hoping people might get excited about concrete career paths and inspiring ideas without feeling like it's just another thing on their overflowing reading list. We brought the Good Food Institute (GFI) to campus and had an awesome time meeting a representative [...]

---

First published:

April 3rd, 2025

Source:

https://forum.effectivealtruism.org/posts/hYE9eEWqCXehCui8E/three-organizers-walk-into-a-liberal-arts-college

---

Narrated by TYPE III AUDIO.