Effective Altruism Forum

Published on May 25, 2026 5:48 PM GMT I've spent the last couple of years working on AI risk at Coefficient Giving, but it’s only in the last few months that I’ve started to feel more viscerally that transformative AI is likely to happen and potentially very soon. To be honest, it probably took me longer than it should have. This shift has been a really difficult emotional experience and I've bee…

Published on May 25, 2026 4:22 PM GMT Cross-posted from my website . Three categories of futures, depending on how AI goes: ASI timelines are long. ASI timelines are short, and we're on track to solving AI alignment. ASI timelines are short, and we're not on track to solving AI alignment. If we want to make a good future for all sentient beings, each of these futures has different implications fo…

Published on May 25, 2026 3:35 PM GMT Pope Leo XIV has released a new, 42k-word encyclical laying out the Vatican's position on many AI safety topics.  You can read the full thing here , or read the vatican's press release here , or coverage in the NY Times, or perhaps consider having an LLM read the whole encyclical, then chatting about whatever specifics you're interested in! Below is a portion…

Published on May 25, 2026 10:44 AM GMT I used an LLM to help draft this post and create images; I’ve rewritten it extensively and endorse it. Exec summary I think foundations should put more effort into having impact through their investments (confidence: high). I more tentatively think that doing so could lead them to have as much impact again as they have through their grants (confidence: lower…

Published on May 25, 2026 10:01 AM GMT TL;DR: Applications are now open for the Secure Program Synthesis Fellowship , powered by Apart Research and Atlas computing . Apply by Sunday the 31st of May . This fellowship offers part-time research opportunities on mentor-led projects at the intersection of formal methods, AI systems, and security . Participants work in small teams to tackle challenging…

Published on May 25, 2026 7:58 AM GMT August 3 to October 2, 2026 in London | Applications close June 2 (AoE) TL;DR: We're running our first Frontier Biodefense Fellowship at pivotal . Nine weeks, fully funded, 1:1 mentorship from Blueprint, SecureBio, SynX, Coefficient Giving, CLTR and more. Open to applicants from a wide range of backgrounds, including those without prior bio experience. Apply …

Published on May 25, 2026 5:22 AM GMT TL;DR GiveWell has published a podcast, looking back on their investments in iron fortification and their $8.2M grant to Fortify Health in 2021. When the grant was made, the projected cost-effectiveness was 5x cash transfers we delivered and estimated 12x. We want to share what this looked like from the inside. Over three years, our team scaled from reaching …

Published on May 25, 2026 1:52 AM GMT AI ‘time horizons’ are mostly not about time (I think it’s mostly ‘data’, but you’ll see where I’m unsure). One chart from 2025 has become perhaps the most (in)famous in modern AI commentary. For those in the know, ‘ the METR graph ’ [1] is unusually compelling because it achieves what so few measures of AI progress have achieved: a somewhat meaningful Y axis…

Published on May 24, 2026 5:06 PM GMT As AI systems advance, our interactions with them will inevitably increase. These advancements will have consequences none of us can fully anticipate. With increased power comes more pressing questions of responsibility and an increased concern about losing meaningful human oversight. What perspective can Stoicism offer on individual responsibility in this co…

Published on May 24, 2026 6:27 PM GMT This is a crosspost for " My Current Solution to the Repugnant Conclusion " by Ibrahim Dagher , which was originally published on Ibrahim's Substack " Uncommon Counsel " on 2 July 2025. I also created an interactive walkthrough of the piece using Claude. It's not very polished but I still found it helpful. One of the most interesting, and difficult, puzzles i…

Published on May 24, 2026 3:40 PM GMT For context, there are two ways models learn in reinforcement learning: exploration vs. exploitation. 1 Every action a model takes has probability p of being random (exploration), and probability 1 - p of being the best possible action among known actions (exploitation). When the model has not been trained at all, when it knows nothing, that first action is p…

Published on May 24, 2026 1:31 PM GMT Summary This semester, I started the AI Policy Accelerator (AIPA), a new student group at George Washington University dedicated to upskilling students on their AI policy careers. We hosted various lectures, workshops, guest events, site visits, and an AI policy writing fellowship. Through this program, dozens of students have become more informed about AI po…

Published on May 24, 2026 3:34 AM GMT Maybe this will sound impolite, but are we now in the vulture-feeding business? Nan Ransohoff's " The Third Wave of American Philanthropy " is the article I've seen shared most widely across influential EA profiles. In it, we read that astronomical figures — $37B to $100B in new philanthropic funding — may soon be on their way, almost all of it (very fortunat…

Published on May 23, 2026 11:32 PM GMT TL;DR: Anthropic restricted access to Claude Mythos Preview, citing a major leap in vulnerability discovery and exploitation capability. I review the 3 most common arguments from skeptics: (1) AISLE Security’s paper showing cheaper models can identify the same bugs as Mythos, (2) benchmark comparisons showing GPT-5.5 performs comparably, and (3) Mythos findi…

Published on May 23, 2026 9:21 AM GMT [Subtitle.] Overstating the evidence in support of animal consciousness could impede efforts to develop more accurate ways of assessing it. This is a crosspost for Premature declarations on animal consciousness hinder progress by Hakwan Lau , which was originally published on The Transmitter on 18 November 2024. Relatedly, I liked Hakwan's article The End of …

Published on May 22, 2026 9:12 PM GMT Disclosure: I wrote this with help from an LLM as a drafting and editing partner. The argument, the experience, the factual claims, and every editorial decision are mine; I directed the drafting, verified the sources, and revised heavily. But enough of the final wording originated in that collaboration that it likely crosses the 10% threshold, so I'm flagging…

Published on May 23, 2026 3:30 AM GMT I have tasted the perfect job, but I am asked to help where it’s needed. Some people say job fit is very important, so should I ignore the calls for help and instead chase my dream role? I have struggled with this tension. Tired of lack of clarity, I did like any good EA does: I made a simple spreadsheet to make this important impact and life choice. My quick…

Published on May 22, 2026 11:28 PM GMT Saturday, May 30, 10–11 AM China time Friday, May 29, 7–8 PM Los Angeles/10–11 PM New York Three speakers, 15 minutes each, group Q&A after. 60 minutes total. Zoom: https://duke.zoom.us/j/99487280430 🎙️ Katt Domingue | Community-Based Coaching for the Vegan-Curious Intern with Nett Kind. Started a small community with new activists from the April 18 Ridglan …

Published on May 22, 2026 11:15 PM GMT Abstract Several major technology companies have announced plans to operate AI data centers in orbit. Elon Musk recently claimed : “the lowest-cost place to put AI will be space […] within two years, maybe three.” If a meaningful fraction of new AI compute really is placed in space within a few years, that would be a fairly big deal for AI governance and str…

Published on May 22, 2026 11:11 PM GMT A lot of conversations evaluating EA — both internal and external — run on a slippage between two things that sometimes go by the same name. It seems quite important to me to separate the philosophical core of EA—to the extent that one can be identified and defined, and the social fact—as it exists as an empirical phenomenon. The points I want to make in thi…