EA Forum Podcast (All audio)

My median expectation is that AGI[1] will be created 3 years from now. This has implications on how to behave, and I will share some useful thoughts I and others have had on how to orient to short timelines.

I’ve led multiple small workshops on orienting to short AGI timelines and compiled the wisdom of around 50 participants (but mostly my thoughts) here. I’ve also participated in multiple short-timelines AGI wargames and co-led one wargame.

This post will assume median AGI timelines of 2027 and will not spend time arguing for this point. Instead, I focus on what the implications of 3 year timelines would be.

I didn’t update much on o3 (as my timelines were already short) but I imagine some readers did and might feel disoriented now. I hope this post can help those people and others in thinking about how to plan for 3 year [...]

---

Outline:

(01:16) A story for a 3 year AGI timeline

(03:45) Important variables based on the year

(03:57) The pre-automation era (2025-2026).

(04:56) The post-automation era (2027 onward).

(06:05) Important players

(08:03) Prerequisites for humanity's survival which are currently unmet

(11:22) Robustly good actions

(15:18) Final thoughts

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

---

First published:

December 22nd, 2024

Source:

https://forum.effectivealtruism.org/posts/tdfDfFyqbvBF9DdeY/orienting-to-3-year-agi-timelines

---

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.

The views expressed here are my own, not those of my employers or people who provided feedback.

Summary

• I investigate whether saving human lives globally, and in China, India and Nigeria may be harmful accounting for the meat-eater problem, i.e. the nearterm increase in farmed animal suffering caused by increasing income or population.
• I estimate a random person globally, and in China, India and Nigeria in 2022 caused 15.5, 34.6, 5.17 and 2.31 times as much suffering to poultry birds and farmed aquatic animals as the person's happiness. Moreover, I expect the meat-eater problem in those countries to become worse in the nearterm as their real gross domestic product (real GDP) per capita increases. So my results suggest extending human lives there is harmful in the nearterm.
• GiveWell has made 1.09 billion dollars of grants impacting people in the countries I mentioned. Ambitious Impact has incubated 8 organisations whose 1st target country was one I mentioned, and [...]

---

Outline:

(00:17) Summary

(02:59) Introduction

(03:21) Methods

(06:50) Results

(07:07) Discussion

(07:10) Meat-eater problem

(12:05) Suggestions

(15:41) Carbon and farmed animal welfare footprint

(18:15) Acknowledgements

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

---

First published:

December 22nd, 2024

Source:

https://forum.effectivealtruism.org/posts/FqioYEr97eoCQMxhk/givewell-may-have-made-1-billion-dollars-of-harmful-grants

---

Narrated by TYPE III AUDIO.

This past term EA Northeastern's London group revamped our Arete (ie. intro) fellowship in a few significant ways. This is how it went!

(tldr: We added a project into the intro fellowship and invested more time into personal relationships with fellows. We had lower drop out rates and higher engagement.[1] You can see this google sheet if you want the numbers.)

If you want to compare our experience to yours here is some context:

• Our group is 3 years old.
• There's appx. 2000 students at the London campus, all inclusive.
• Appx. 1250 of those at any given time are freshmen studying abroad from Boston.

The primary problems we had last year were:

1. Attendance drop off.
2. ‘Apprehensive’ participation in meetings.
3. Low engagement with the group (& EA) outside of the fellowship.

The following are changes that we made from last year to address them.

We Added a [...]

---

Outline:

(01:10) We Added a Project

(04:53) We Designed New Marketing Materials

(05:56) We Made more Friends (awww...)

(09:42) We Improved our Final Event

The original text contained 7 footnotes which were omitted from this narration.

---

First published:

December 20th, 2024

Source:

https://forum.effectivealtruism.org/posts/7ge6EPWJHXPyAGg62/running-a-project-based-intro-fellowship

---

Narrated by TYPE III AUDIO.

The Supervised Program for Alignment Research (SPAR) is now accepting mentee applications for Spring 2025! SPAR connects mentors and mentees for three-month AI safety and governance research projects. The Spring 2025 cycle of SPAR will occur virtually between February 10 and May 17, 2025. We have 46 exciting projects available across AI safety, policy, and strategy - you can click here to explore the projects!

Apply now

SPAR mentee applications are open until January 8th, but we strongly encourage applying early since many mentors review applications on a rolling basis.

As a SPAR mentee, you'll participate in a collaborative research project under an experienced mentor. You might be a good fit if you have a technical/STEM background or a governance/policy background. SPAR projects vary in how much specific knowledge they require.

We think SPAR's value add is that it enables mentors to increase their research output by working with a [...]

---

First published:

December 19th, 2024

Source:

https://forum.effectivealtruism.org/posts/zuQsEth6kgn6wHx6c/apply-now-to-spar

---

Narrated by TYPE III AUDIO.

This is a link post.

The AI safety community has grown rapidly since the ChatGPT wake-up call, but available funding doesn’t seem to have kept pace.

However, there's a more recent dynamic that's created even better funding opportunities, which I witnessed as a recommender in the most recent SFF grant round.[1]

Most philanthropic (vs. government or industry) AI safety funding (>50%) comes from one source: Good Ventures. But they’ve recently stopped funding several categories of work (my own categories, not theirs):

• Many Republican-leaning think tanks, such as the Foundation for American Innovation.
• “Post-alignment” causes such as digital sentience or regulation of explosive growth.
• The rationality community, including LessWrong, Lightcone, SPARC, CFAR, MIRI.
• High school outreach, such as Non-trivial.

In addition, they are currently not funding (or not fully funding):

• Many non-US think tanks, who don’t want to appear influenced by an American organisation (there's now probably more [...]

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

---

First published:

December 21st, 2024

Source:

https://forum.effectivealtruism.org/posts/s9dyyge6uLG5ScwEp/it-looks-like-there-are-some-good-funding-opportunities-in

---

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.

See livestream, site, OpenAI thread, Nat McAleese thread.

OpenAI announced (but isn't yet releasing) o3 and o3-mini (skipping o2 because of telecom company O2's trademark). "We plan to deploy these models early next year" (source). "o3 is powered by further scaling up RL beyond o1" (source); I don't know whether it's a new base model.

o3 gets 25% on FrontierMath, smashing the previous SoTA. (These are really hard math problems.) Wow. (The dark blue bar, about 7%, is presumably one-attempt; unfortunately OpenAI didn't say what the light blue bar is, but I think it doesn't really matter and the 25% is for real.[1])

o3 also is easily SoTA on SWE-bench Verified and Codeforces.

It's also easily SoTA on ARC-AGI, after doing RL on the public ARC-AGI problems + when spending $4,000 per task on inference (!).[2]

OpenAI has a "new alignment strategy"; looks like Constitutional AI (and just about [...]

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

---

First published:

December 20th, 2024

Source:

https://forum.effectivealtruism.org/posts/aNdg7ctFP9zFcowNd/o3

---

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 August 2023, we shared that:

• EAGs from 2022–2023 each cost around $2M–$3.6M USD at around $1.5k–2.5k per person per event.
  • These events typically cost a lot because the fixed costs of running professional events in the US and UK are surprisingly high.
• We’re aiming to get these event costs down to ≤$2M each moving forwards.
  • We’ve already started to cut back on spending and will continue to do so, whilst also raising default ticket prices to recoup more of our costs.

Since then, our team has cut the cost of running EA Global significantly. Our average gross cost-per-attendee (CPA) has fallen from ~$2,000 in 2023 to ~$1,600 in 2024. We also met our target of getting total event costs down: every EA Global this year cost <$1.5M.

We’ve cut costs by:

• Reducing the amount of food we serve. We cut dinners on [...]

---

Outline:

(06:01) Other 2024 updates and 2025 goals

(06:14) EA Global

(09:39) EAGx

(12:38) Partner Events

(14:27) Impact stories from our events

(14:56) Donate to support our work

(15:34) Appendix: Metrics

The original text contained 10 footnotes which were omitted from this narration.

---

First published:

December 20th, 2024

Source:

https://forum.effectivealtruism.org/posts/8mCsYwmZC7DCA7cYD/update-on-ea-global-costs-and-2024-cea-events-team

---

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.

The Long-Term Future Fund is pleased to invite applications to our Q1 2025 funding round. We are seeking to support individuals and organizations working to help humanity navigate existential risks or working on projects improving humanity's long-term flourishing.

→ Click here to apply ←

Deadline for applying is February 15th. We will respond to applications submitted before then by March 31st, and should have most funding disbursed by April 30th.

Why Grant Rounds?
After a period of accepting applications on a rolling basis, we're trialling returning to focused grant rounds. I am hoping this will:

• Allow us to recalibrate our funding thresholds given a rapidly changing giving environment
• Provide donors with clearer feedback on the impact of their contributions
• Offer applicants predictable response timelines

For particularly time-sensitive opportunities, we maintain a fast-track application process, though we expect the bar to be substantially higher for funding through this channel [...]

---

First published:

December 20th, 2024

Source:

https://forum.effectivealtruism.org/posts/aBkALPSXBRjnjWLnP/announcing-the-q1-2025-long-term-future-fund-grant-round

---

Narrated by TYPE III AUDIO.

This is a link post.

Executive summary

2024 has been a year of significant growth and consolidation for Observatorio de Riesgos Catastróficos Globales (ORCG), marked by impactful advancements across our four core workstreams. We've expanded our research initiatives and strengthened international collaborations, solidifying our role as a leading voice in understanding and mitigating catastrophic risks.

From the outset, ORCG has championed proactive and collaborative approaches to navigating the complex challenges of potential global catastrophes. This commitment fueled a diverse range of activities in 2024, with highlights including:

• Artificial Intelligence: We continued our work on the implementation of the EU AI Act, including by participating in the drawing-up of the GPAI Code of Practice, supported by resources from Longview Philanthropy. We also formulated the AI Risk Explorer (AIRE), securing funding from Open Philanthropy. Finally, we published the report “Compute for Good: The Potential of European Supercomputers for AI Research and Development”.
• Pandemic Preparedness [...]

---

Outline:

(00:12) Executive summary

(04:07) Products

---

First published:

December 18th, 2024

Source:

https://forum.effectivealtruism.org/posts/dgoZrPfveh3EqEbvB/2024-a-year-of-consolidation-for-orcg

---

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.

TLDR/Context: For some time, I donated 10% of my income anonymously, without telling my friends and family or sharing it on social media. I was worried that bringing up one's charitable donations would be perceived as bragging or virtue signaling. However, I also believe there are important sociological reasons why we should be open about our pro-social behavior. People are largely influenced by the behavior of others around them. So by being public about donations, one also provides ‘social proof’ of such actions, helping spread a wider norm of effective giving.

To deal with these conflicting issues, I wrote this essay outlining the reasons for giving effectively and for doing so publicly. I felt more comfortable being public about my donations if I could also explain the reason why I was public about them. A few weeks ago the essay was published on the [...]

---

Outline:

(02:14) 1. Why Altruism?

(08:02) 2. Societys Persecuted Minority: Do-Gooders

(11:30) 3. A Theory of Moral Sentiments and Amoral Behaviour

(15:27) 4. Normalise It!

The original text contained 24 footnotes which were omitted from this narration.

---

First published:

December 20th, 2024

Source:

https://forum.effectivealtruism.org/posts/3JScezgm5L3gdwsdB/the-virtues-of-virtue-signalling

---

Narrated by TYPE III AUDIO.