Hello!
Our favourite links this month include:
|
|
|
Also – Probably Good has relaunched their 1-1 career advising service. EAGxNigeria has opened applications, and EAG London is closing its applications on May 25th. Longview Philanthropy is building an AI grantmaking team — they are hiring a programme director, and two grantmakers. Thanks for reading, and I hope you enjoy the articles!
If you’ve found this newsletter valuable, let us know how, and if you have any other feedback, reply to the email, or fill in our form.
— Toby, for the EA Newsletter Team
|
|
Articles
How to stop half a million children dying of sepsis
Each year, between 400,000 and 700,000 newborns die of sepsis — a reaction to bacterial infections. Most of those deaths could be prevented by well-timed antibiotics, if only we could test for sepsis well.
Sepsis is hard to diagnose, and the current test — blood culturation — takes 2–3 days, is expensive, and has a high rate of false negatives. This leads to delays in treatment for babies who need it. In those countries without well-equipped neonatal intensive care units, this means death.
NeoTest, a new global health initiative, wants to raise funds to offer a guaranteed market — via an “advance market commitment” — to manufacturers that can develop a rapid, affordable sepsis test suitable for low-income countries. By ensuring demand upfront (specifically, paying $5 — plus $3 from recipient countries — per test for the first 24 million produced), the initiative will incentivise companies to develop and supply the tests.
This model has worked before: a larger advance market commitment ($1.5 billion) helped create pneumococcal vaccines, saving around 700,000 lives. NeoTest aims to repeat the success on a smaller scale ($120 million), with funding from donors and the countries that will use the tests.
If you’d like to learn more about advance market commitments, and the other crucial technologies they could accelerate, I highly recommend Rachel Glennerster’s appearance on the 80,000 Hours podcast.
|
|
We're on track to farm trillions of insects this decade
Most people worried about factory farming think of chickens or pigs. A few consider fish. Almost none think of insects. But, by 2033, among the 6 trillion animals slaughtered each year, the majority will be insects.
A new report from Rethink Priorities suggests insect farming is likely to explode — from hundreds of billions of individuals farmed today to several trillion within a decade.
This isn’t because of consumer demand for cricket burgers (fun word: “entophagy”, the practice of eating insects). Most insects are raised to feed other farmed animals—chickens, pigs, and fish. Because insects are so tiny, farming them efficiently and providing food for billions of animals means producing staggering numbers.
There’s growing evidence that some insects — like flies and bees — may experience pain. Yet industry practices remain unregulated: freezing, shredding, and boiling them by the billion, with no welfare standards.
The lesson from earlier factory farming is clear: if we wait until an industry is enormous to consider animal welfare, it’s already too late.
If you’d like to support future research into farmed animal welfare, consider donating to Rethink Priorities, or the Arthropoda Foundation. For more information about insect farming, and the science behind insect sentience, it’s worth listening to another 80,000 Hours podcast, with Meghan Barrett.
|
|
AI Welfare at Anthropic
Anthropic recently launched a first-of-its-kind research programme to explore AI model welfare, which will ask: Are present-day AI systems conscious? Could they become conscious? And crucially: How should companies like Anthropic act under this uncertainty?
In an interview with Stuart Ritchie, Kyle Fish — who leads the initiative — discusses the uncertainties involved (we don’t know if current models are conscious, or even how we’d tell) and the reasons for taking the possibility seriously. The conversation covers the scientists and philosophers who take this issue seriously, criticisms of the field, and some early practical suggestions — such as giving AI models the option to opt out of distressing prompts. That’s a speculative idea: we don’t yet know whether models find anything distressing, or even what “distress” would mean for a digital mind. But it’s a small, reversible step that could generate useful data.
Rob Long, a researcher at Eleos AI, summarised the interview and links to more in-depth material and some measured pushback.
As with insect suffering, the case for early work on AI welfare doesn’t depend on certainty. There may soon be millions (or billions) of digital minds. We have the opportunity to set humane defaults while the cost of change is still low.
You can support further research by donating to Eleos AI.
|
|
In other news
- Asterisk Magazine forecasted the effects of the US’s cuts to USAID programmes. Over 5 key areas, they estimate a death toll of between 483,000 and 1.14 million over the next year.
- In more positive news — even in rich countries, child mortality has fallen substantially since 1991. Our World in Data explores the stats.
- The biggest welfare problem facing chickens might be their genetics — genetics manipulated by humans.
- The Wild Animal Initiative explores how AI might impact their work over the next two decades.
- Open Philanthropy tells the story of a new device that could cheaply and reliably improve water quality across the developing world.
- Trump signed an executive order aimed at curtailing dangerous gain of function research. The responses, as expected, are mixed, with Science taking a more sceptical stance, and On reading, a column in the Biosafety Now substack, broadly lauding the move.
- The Shrimp Welfare Project, which recently featured on the US’s Daily Show, outlines its plans through to 2030.
- On AI
For more stories, try these email newsletters and podcasts.
|
|
Resources
Links we share every time — they're just that good!
|
|
Jobs
Boards and resources:
Selection of jobs
Bluedot
The Good Food Institute
Lead Exposure Elimination Project
Longview
- AI Programme Director (London, Washington DC or remote (San Francisco), USD $176K–$214.5K)
- AI Grantmaker (London, Washington DC or remote (San Francisco), USD $115.5K– $154K)
Open AI
Open Philanthropy
|
|
Announcements
Events and Conferences
- EAG London returns June 6–8. Applications are open until May 25th.
- EAGxNigeria (July 11–13, Abuja) will bring together EA-aligned individuals from across Africa. Free to attend. Apply by June 30th.
- Rethink Priorities is launching its Strategic Animal Webinars series with a session on Including Animal and AI Welfare in AI Alignment, featuring Adrià Moret. The webinar will explore integrating nonhuman and AI welfare into AI alignment efforts. Register to join on May 28th at 5pm BST.
Fellowships and Courses
- CBAI’s Summer Research Fellowship (Cambridge, MA) offers $8K, office space, and mentorship from top AI safety researchers. Open to students and early-career professionals in technical or governance AI safety. Apply by May 21st.
Prizes and Funding
- Prêmio Artigo Eficaz (Effective Article Prize) by Altruísmo Eficaz Brasil seeks impactful research on global challenges from Brazilian researchers. Six winners across three themes (Global Risks, Animal Welfare, Development) will receive R$1,500 each. Submit published articles or unpublished essays with Brazilian authorship by June 30th.
- Foresight Institute is funding projects at the intersection of AI safety, neurotech, coordination, and forecasting. Grants range from $10K to $300K+, with quarterly deadlines. Open to individuals and orgs tackling one of four focus areas — the next round closes on June 30th.
- Tarbell is offering $1K–$15K grants for original reporting on AI’s societal impacts — from frontier lab accountability to global policymaking. Open to freelancers and staff journalists. Apply by May 31st.
|
|
Monthly Organizational Updates
You can see this month's updates from a wide range of organisations on the EA Forum.
|
|
Timeless Classic
This month's timless classic – Cheerfully – from Julia Wise, is a reflection on how to live well, while focusing on doing the most good with your life. Julia has thought long and hard about these trade-offs, and has reached some real wisdom on how to balance them. (And I am proud to call her a colleague).
|
|
We hope you found this edition useful!
If you’ve taken action because of the Newsletter and haven’t taken our impact survey, please do — it helps us improve future editions.
Finally, if you have any feedback for us, positive or negative, let us know!
– Toby, for the Effective Altruism Newsletter Team |
|
|
|
|
