EA Forum Podcast (All audio)

Even if we solve the AI alignment problem, we still face non-alignment problems, which are all the other existential problems[1] that AI may bring.

People have written research agendas on various imposing problems that we are nowhere close to solving, and that we may need to solve before developing ASI. An incomplete list of topics: misuse; animal-inclusive AI; AI welfare; S-risks from conflict; gradual disempowerment; risks from malevolent actors; moral error.

The standard answer to these problems, the one that most research agendas take for granted, is "do research". Specifically, do research in the conventional way where you create a research agenda, explore some research questions, and fund other people to work on those questions.

If transformative AI arrives within the next decade, then we won't solve non-alignment problems by doing research on how to solve them.

These problems are thorny, to put it mildly. They're the sorts of problems where you have no idea how much progress you're making or how much work it will take. I can think of analogous philosophical problems that have seen depressingly little progress in 300 years. I don't expect to see meaningful progress in the next 10.

[...]

---

Outline:

(02:54) Approach 1: Meta-research on what approach to use

(03:20) Approach 2: Pause AI

(05:02) Approach 3: Develop human-level AI first, then (maybe) pause

(06:40) Approach 4: Research how to steer ASI toward solving non-alignment problems

(07:57) Conclusion

The original text contained 6 footnotes which were omitted from this narration.

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

The Donation Election has begun!

You can vote here, find out more about the voting system here, and donate to the Donation Election fund here.

This post introduces our candidates. We’re splitting by cause area for easy reading, but you’ll be voting on them all together.

Note that:

• There will likely be a few more candidates added in the next few days (candidates who wrote marginal funding posts but haven't let me know whether they'd like to participate in the election yet).
• It is possible that, after eligibility checks, a few candidates may be removed from the list. I won’t post publicly about removing specific candidates because eligibility in the Donation Election shouldn’t affect your decision on whether to donate to them or not — eligibility is all about whether CEA can grant to them easily. 

This table was co-written with GPT-5 to improve clarity and readability for Forum readers. Any mistakes or oversights are our own, not the organisations mentioned. If you spot an error, please message me — I’ll donate $10 to the Donation Election Fund for everyone who spots an error.

Global Health

1Day Sooner

• [...]

---

Outline:

(01:20) Global Health

(01:32) Animal Welfare

(01:40) Global Catastrophic Risks

(01:49) Meta- or mixed

(02:00) AI governance

(02:10) Other

---

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.

Summary

• So far in 2025, EAIF has made 34 grants for a total of $2.4M and we expect to grant a total of $2.8M by the end of the year. 
• For EAIF to continue at the same level of grantmaking in 2026, we need to raise $1.9M in funding. 

EAIF's Grantmaking

The Effective Altruism Infrastructure Fund (EAIF) aims to increase the impact of projects that use the principles of effective altruism, by increasing their access to talent, capital, and knowledge.

EAIF is part of EA Funds, which has recently merged with the Centre for Effective Altruism. The current Fund Managers at EAIF are Harri Besceli (author of this post) and Jamie Harris.

EAIF makes grants across a range of different areas, though typically these are aimed at the following:

• Awareness: Increasing awareness of the ideas of effective altruism (eg. EA related content and EA Groups).
• Careers: Increasing the number of people who are making career decisions based on the ideas and recommendations of effective altruism (eg.  EA Groups and Cause/ Path specific community building).
• Donations: Increasing the amount of money donated to causes and areas recommended by effective altruism (eg. effective giving projects).
• Community Support: Supporting members of the [...]

---

Outline:

(00:10) Summary

(00:33) EAIF's Grantmaking

(03:36) Marginal Funding

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

TLDR:

• Apart Research incubates talent and research in AI Safety. We use high-volume weekend hackathons to filter for talent, then support them through the Studio and Fellowship programs.
• Track Record Highlights include placements at leading AIS orgs (e.g. METR, CAIF), alumni starting their own AIS orgs (e.g. Geodesic Resesarch, Asymmetric Security) and impactful research output, including an ICLR 2025 Oral (top 1.8%) on AIS Evals. Our work has also been used by Anthropic, UK AISI and the EU AI Office.
• We are currently operating at capacity. While our remote infrastructure is scalable with decreasing marginal costs, we cannot support the volume of qualified applicants without additional funding.
• Donations between $1k and $12,5k will enable us to support more top talent, produce more research and host additional global hackathons in AI Safety. Donations between $80k and $100k will enable us to make key hires for 2026.

What is Apart Research

Apart Research is a remote-first AI Safety organisation for field-building and research. We focus on converting technical talent into published AI safety researchers, ready to contribute to impactful organizations. We serve top talent globally, whether they're new to AI safety or looking to deepen their engagement.

We believe we [...]

---

Outline:

(00:13) TLDR:

(01:26) What is Apart Research

(02:16) How Apart Research creates impact

(02:52) en-US-AvaMultilingualNeural__ Flowchart showing Apart Research Programs timeline: Studio, Sprint, and Fellowship stages.

(03:05) What makes our approach distinctive:

(05:04) Aparts pipeline creates three types of value:

(06:34) Track Record

(06:37) Track Record (Inception - June 2025)

(07:32) Our People: Case Studies from success stories in our programs

(09:36) Our Research

(10:46) Recent Progress (June 2025 - Present)

(12:29) What we will be able to do with more funding

---

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.

I’ve spoken with hundreds of entomologists at conferences the world over. While there's clearly some self-selection (not everyone wants to talk to a philosopher), my experience is consistent: most think it's reasonable to care about the welfare of insects. Entomologists don’t regard it as the last stop on the crazy train; they don’t worry they’re getting mugged; they don’t think the idea is just utilitarianism run amok. Instead, they see some concern for welfare as stemming from a common-sense commitment to being humane in our dealings with animals.

Let's be clear: they embrace “some concern,” not “bugs have rights.” Entomologists generally believe it's important to do invasive studies on insects, to manage their populations, to kill them to document their diversity. But given the choice between an aversive and a non-aversive way of euthanizing insects, most prefer the latter. Given the choice between killing fewer insects and more, most prefer fewer. They don’t want to end good lives unnecessarily; they don’t want to cause gratuitous suffering.

It wasn’t always this way. But the science of sentience is evolving; attitudes are evolving too. These people work with insects every day; they constantly face choices about how to catch insects, how [...]

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

TL;DR

Despite operational setbacks, EA Poland made solid progress on the goals outlined in our December 2024 post:

• In February, we published and promoted The Life You Can Save in Polish, generating 2,153 downloads.
• In August, we launched Dobra Kariera - a local high-impact career initiative.
  • Published and promoted the 80,000 Hours Guide in Polish (80 000 Godzin), generating 2,052 downloads.
  • Conducted 25 career 1-1s. 
• In October, we got AI Safety Poland to a new level, with a strong team, dedicated career advisory (8 advisees so far), and regular events attracting 30+ attendees. 
• Więcej Dobra, the Polish effective giving platform, is now formally a standalone charity run by Daniel Wyrzykowski, and it has recently received its second grant from an EA org. 

We see the above as a promising foundation. However, a path from the initial interaction with EA and its content to exerting a direct impact might take years. To increase the chance of such an impact crystallizing, we need time to convert the generated interest into deeper engagement. That is how we help keep the drive afloat and lead to so-called Significant Actions (e.g., career shifts into EA cause areas or 10% Pledges).

To enable it, we are now seeking $37,000 [...]

---

Outline:

(00:14) TL;DR

(02:20) Progress and goals

(03:06) AI Safety Poland

(07:27) Dobra Kariera: 80,000 Hours Guide / 80 000 Godzin and Career 1-1s

(15:47) The Life You Can Save (TLYCS)

(17:34) Community

(23:00) Więcej Dobra

(23:40) Funding Diversification

(26:29) How can you help?

(27:46) Thanking spree (in random order):

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

Last year I gave my reasoning on cause prioritization and did shallow reviews of some relevant orgs. I'm doing it again this year.

Cross-posted to my website.

Cause prioritization

In September, I published a report on the AI safety landscape, specifically focusing on AI x-risk policy/advocacy.

The prioritization section of the report explains why I focused on AI policy. It's similar to what I wrote about prioritization in my 2024 donations post, but more fleshed out. I won't go into detail on cause prioritization in this post because those two previous articles explain my thinking.

My high-level prioritization is mostly unchanged since last year. In short:

In the rest of this section, I will cover:

What I want my donations to achieve

By donating, I want to increase the chances that we [...]

---

Outline:

(00:22) Cause prioritization

(01:25) What I want my donations to achieve

(04:32) There is no good plan

(05:23) AI pause advocacy is the least-bad plan

(06:22) How Ive changed my mind since last year

(06:27) Im more concerned about non-alignment problems

(07:02) Im more concerned about AI-for-animals

(08:38) How my confidence has increased since last year

(08:43) We should pause frontier AI development

(09:03) Peaceful protests probably help

(10:22) I have a high bar for who to trust

(11:35) My favorite interventions

(13:04) Organizations (tax-deductible)

(13:50) AI-for-animals orgs

(15:18) AI Safety and Governance Fund

(15:54) Existential Risk Observatory

(17:02) Machine Intelligence Research Institute (MIRI)

(18:58) Palisade Research

(19:41) PauseAI US

(21:25) Video projects

(21:50) Non-tax-deductible donation opportunities

(22:28) AI Policy Network

(23:21) Americans for Responsible Innovation (ARI)

(24:18) ControlAI

(25:17) Congressional campaigns

(26:17) Encode

(26:54) Where Im donating

The original text contained 16 footnotes which were omitted from this narration.

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

A message on OpenAI's internal Slack claimed the activist in question had expressed interest in “causing physical harm to OpenAI employees.”

OpenAI employees in San Francisco were told to stay inside the office on Friday afternoon after the company purportedly received a threat from an individual who was previously associated with the Stop AI activist group.

“Our information indicates that [name] from StopAI has expressed interest in causing physical harm to OpenAI employees,” a member of the internal communications team wrote on Slack. “He has previously been on site at our San Francisco facilities.”

Just before 11 am, San Francisco police received a 911 call about a man allegedly making threats and intending to harm others at 550 Terry Francois Boulevard, which is near OpenAI's offices in the Mission Bay neighborhood, according to data tracked by the crime app Citizen. A police scanner recording archived on the app describes the suspect by name and alleges he may have purchased weapons with the intention of targeting additional OpenAI locations.

Hours before the incident on Friday, the individual who police flagged as allegedly making the thread said he was no longer part of Stop AI in a [...]

---

First published:

Source:

Linkpost URL:
https://www.wired.com/story/openai-office-lockdown-threat-san-francisco/

---

Narrated by TYPE III AUDIO.

Overview

Strategic Alliances for a Viable Earth (SAVE) aims to reduce the risk of crossing ecological and social tipping points by having a systemic approach on the interconnected crises that are present today. This is done by connecting actors who have the power to reduce these risks with experts and organizations who know how to reduce these risks by using the networks in SAVE.

SAVE is composed of diverse, highly networked, impact-driven, experienced policy advisors; climate activists; management consultants; representatives from large and small/medium enterprises (SMEs); lobbyists; journalists; and employees from NGOs, ministries, parliaments, think-tanks, and leading scientific institutions. SAVE is connected to hundreds of heterogeneous actors from all relevant societal spheres: politics, science, (social) media, civil society, economy.

At the moment the CEO, Johnny Stengel, is the only one with a paid position at SAVE. For 90 000 USD, SAVE would hire two people for one year, who would connect experts and organizations in global existential risks with politicians, policymakers and other people with the ability to reduce those risks. These people use their networks and experience to connect the dots that can make evidence-based ideas become reality.

Why SAVE?

• No existing actor combines sector-crossing strategic [...]

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

TLDR: Hundreds of millions of mice and rats are bred and killed each year as feeder animals for pet snakes, in conditions that resemble the worst of factory farming. We are developing a convenient, cost-competitive, and nutritionally complete snake food made from beef meat, organs, and bones to displace frozen feeder rodents. Cattle to mice is up to 1 to 100,000 replacement and a very strong welfare tradeoff. Products will be on the market in 2026, and will start displacing feeder rodents immediately on a 1:1 basis. We are anonymous on the Forum, but are an ACX Grants project raising 75k in marginal funding for early product launch, plus 9k for a university study. You can donate via Manifund or reach us at [email protected].

1. The overlooked problem: hundreds of millions of mammals in poor conditions

Each year, on the order of hundreds of millions of mice and rats are bred, farmed, and killed as feeder animals for pet snakes. These mammals are typically:

• raised in crowded plastic tubs
• kept in barren environments
• exposed to high disease and mortality
• killed using methods that would be alarming by farmed-animal standards (freezing, blunt-force trauma, sometimes live feeding)

Despite the [...]

---

Outline:

(01:01) 1. The overlooked problem: hundreds of millions of mammals in poor conditions

(01:41) 2. The approach: a convenient, cost-competitive, complete diet for snakes

(02:49) 3. Short-timeline and 1:1 replacement at scale

(03:37) 4. Where we are now

(04:09) 5. Why this post is anonymous

(04:44) 6. Funding gaps

(04:48) A. $75k for early-stage launch

(05:05) B. $9k for a published university veterinary study

(05:26) C. Pre-seed funding + potential European build-out (impact investing interest)

(06:05) 7. How to help

---

First published:

Source:

---

Narrated by TYPE III AUDIO.