Research & grantmaking @coeff_giving • Leave me anonymous feedback: shorturl.at/7U0bz
New York
- Tweets 756
- Following 540
- Followers 2,635
- Likes 6,454
Pinned Tweet
AI has huge implications for global health and economic wellbeing.
So we @coeff_giving are putting out a $10-30M RFP to fund work that helps us prepare for and respond to the potential impacts of AI on economies, health systems, scientific research, and more. 🧵
1
28
90
19,238
In 2016 Geoffrey Hinton said “we should stop training radiologists now" since AI would soon be better at their jobs.
He was right: models have outperformed radiologists on benchmarks for ~a decade.
Yet radiology jobs are at record highs, with an average salary of $520k.
Why?
199
442
3,649
2,265,474
Inspired by this viral meme about DOGE: Research the US government has supported that can be made to sound silly, but that has contributed to human progress.
Valuable work can often be framed as absurd out of context. That doesn’t make funding research any less important.🧵
23
229
1,112
110,231
The USDA funded research on the sex life of the screwworm.
Screwworms were a cattle parasite in the southern US. That researcher developed the sterile insect release technique, and, by 1966, the U.S. was declared screwworm free -- saving ~$20B for US cattle producers. 3/
7
85
859
207,269
The NIH funded massages for infant rats.
The research found that tactile stimulation could improve health outcomes for premature babies -- leading to a change in the standard of care, and saving lives & ~$4.7B in healthcare costs. 2/
3
40
509
47,483
One reason for this: radiology AI doesn't read and form a holistic view on a scan like a doctor does.
Each model is built to answer a single narrow question:
– Is there pneumonia in this chest X-ray?
– Is there a lung nodule in this CT?
– What’s the calcium score in this scan?
2
15
402
51,817
There’s another catch: many models struggle outside the hospital they were trained in. In 2024, 38% of FDA-cleared radiology AIs were tested on data from just one hospital. Move them elsewhere, and accuracy can drop by up to 20 percentage points.
8
21
389
44,156
CheXNet, released in 2017, could detect pneumonia better than a panel of board-certified radiologists.
Since then, hundreds of models have outperformed humans on benchmarks. Today, over 700 FDA-cleared radiology AI models exist. They make up >3/4 of all medical AI devices.
6
13
368
56,963
If benchmark performance translated to reality, radiologists should be the canary in the coal mine of AI job loss.
Instead, radiology residency slots hit a record high in 2025.
2
9
342
54,350
The NSF funded researchers putting shrimp on treadmills.
The grant went toward studying how shrimp respond to environmental stressors, to understand maximum thresholds for exposure to bacteria commonly found in landfill waste, protecting animals and an important food source. 4/
3
12
361
29,727
DARPA funded research into why cats seem to land right side up so often.
Their findings could help make hardier and more agile robots for search-and-rescue missions, saving more lives in hard to reach places after natural disasters. 5/
4
10
358
23,766
That means a radiologist can’t just “hand off” a scan to AI. To cover a typical day, they’d need to pick from dozens of different models, run each one separately, and stitch the answers together.
5
13
308
46,493
Even platforms that bundle multiple models still spit out a list of disconnected yes/no answers.
But a good radiologist integrates findings, context, and patient history into a single coherent view of a person's health.
4
7
297
44,626
AI radiology today is powerful, but it consists of many narrow islands of automation that have failed to replace radiologists' time.
This isn't the full picture — there are regulatory and economic drivers too. Read about them in the @WorksInProgMag piece: worksinprogress.co/issue/the…
6
15
266
37,195
Interesting article, but notably presents ~no evidence for these lay-offs being driven by AI!
11
14
225
22,202
On top of this, datasets are still a barrier for some scans and patient populations. Even with 700+ FDA-approved imaging AIs, they cover only a fraction of real-world work; most cluster around a few conditions like stroke, breast cancer, or lung cancer.
3
6
206
38,527
If you enjoyed the article, I write about health, economic development, and scientific and technological progress here: newsletter.deenamousa.com/
8
6
140
29,820
If you found this thread interesting, I write regularly about health and development at deenamousa.substack.com/!
3
2
143
13,576
In 2003 an iron & folic acid supplement RCT in Zanzibar was shut down abruptly when researchers found the treatment increased mortality. Kids who got an iron dose were 12% more likely to die from severe illness vs the control. Why?
I explain in my latest for @WorksinProgMag! 🧵
4
19
98
16,225
100 years ago, the US introduced iodized salt and, in a major public health victory, ~eliminated iodine deficiency. Now it’s making a quiet comeback, with potentially devastating health effects.
What’s behind its return, and what can we do? I wrote about it in @TheEconomist 🧵
8
35
89
50,185
Excited to be part of this fellowship! I really admire the work @WorksInProgMag and @AsimovPress do.
I’ll be writing about AI x health & economic impacts - would love to chat about ideas, tips, or pieces you'd love to read.
Announcing our inaugural cohort for the @AsimovPress + @WorksInProgMag Writers’ Fellowship.
5 writers. 5 cities. Covering everything from pharmaceuticals to the history of science and AI’s impacts on the developing world.
Their writing will appear regularly in our magazines.🔻
7
6
89
8,122
Skepticism of government-funded scientific research is making the rounds again. In @asteriskmag, @notanastronomer and I write about why “weird” research pays off, and how the government is particularly well-positioned— better than the private sector—to fund it. 🧵
Inspired by this viral meme about DOGE: Research the US government has supported that can be made to sound silly, but that has contributed to human progress.
Valuable work can often be framed as absurd out of context. That doesn’t make funding research any less important.🧵
2
24
73
7,447
“It’s only a cost overrun if the cost you planned for was real.”
This morning's Statecraft is in my top 5 all-time interviews we've conducted.
As an environmental activist, Stephanie Pollack sued to stop transit projects.
As secretary of Massachusetts' DOT, she had to save them. statecraft.pub/p/how-to-salv…
4
7
62
12,786
I heard we're talking about air conditioning again... One year ago from @WorksInProgMag
1
7
61
21,298
Excited that @open_phil is expanding our work on economic growth and scientific progress.
The case for the Abundance & Growth Fund is simple: Economic growth isn’t just about GDP—it’s one of the most powerful forces for improving human lives. 🧵
Exciting update: @open_phil is doubling down on our YIMBY, innovation, and metascience success by launching a >$120m Abundance & Growth Fund to accelerate economic growth and boost scientific & technological progress. Funding from @GoodVentures, @patrickc, + others.
Why now? 🧵
3
8
58
7,635
Could we catch the next outbreak before anyone gets sick?
I wrote for @AsimovPress about airborne biosensors that can detect viruses in real time and why, despite their promise, we’re not using them yet.
In 2015, avian flu infected over 50 million birds and cost the U.S. poultry industry about $3.3 billion.
Today, the technology to almost insantly detect pathogens like H5N1 exists. What we lack is government action and investment to use it.
Our latest from @deenamousa.
2
13
56
9,431
It’s fascinating to me how much higher the standards are for autonomous vehicles to be considered safe than for human drivers
We recently began serving riders in Austin, and it’s great to see the safety benefits of the @Waymo Driver already playing out for both our riders and other road users.
2
3
50
3,299
I weigh in on the British air conditioning discourse in @thetimes with some commentary on how heat drags health and productivity.
A riff on a @WorksInProgMag piece I wrote last year about developing economy contexts.
4
4
51
3,469
I went on @TheEconomist's podcast, The Intelligence, to talk about the quiet re-emergence of iodine deficiency. Luckily, we know how to solve it! A few highlights: 🧵
4
9
48
45,616
If he takes notes from Alec / @IFP, that alone makes me feel some amount better about the world and about DOGE’s likely recommendations
43
1,907
We at @open_phil have been proud supporters of Baker since 2017. Extremely well deserved. Congratulations to everyone at @UWproteindesign!
Several other foundations provided support throughout the years too. Flexible philanthropy is great for people/problems who break NIH boxes.
1
42
1,601
From @mattsclancy in @WorksInProgMag -- it's easy to overlook how much technological progress has reduced the frictions of life and given us back time in incremental ways
2
3
41
4,466
Come work on accelerating economic & technological progress at Open Phil!
The Abundance and Growth Fund at Open Philanthropy is hiring! We’re looking for 2-4 people to help expand this new $120+ million program to accelerate economic growth and reduce the cost of living through strategic grantmaking and research. (1/4) openphilanthropy.org/researc…
1
40
1,718
New post! We celebrate breakthroughs—the internet, life-saving drugs, major peace treaties—but behind them are decades of overlooked investments. Now, those investments are on the chopping block. 🧵
2
2
32
1,941
Many inventions - like CRISPR and Ozempic - likely would not have come about without publicly-funded research. Orgs don't benefit from all of the positive externalities of early-stage research, leading them to structurally undervalue it.
1
4
29
2,085
The trajectory of the cost of solar over the past few decades is astounding (chart from @OurWorldInData)
2
3
29
3,722
How much is a year of your life worth?
In Denmark, the average answer is $24,000. In Japan, it’s about $67,000.
New post up on what we say when asked to put a price on life. 🧵
1
4
27
2,291
A recent paper claims AI passes the Turing test for poetry -- @mattsclancy wrote a blind test to sense-check. So far, results from a few of us at @open_phil suggest the paper is optimistic & AI has a way to go to beat human poets.
Link below if you'd like to take it yourself!
2
5
27
6,562
Striking op-ed highlighting @PureEarthNow's work:
"One estimate suggests that cognitive impairments from lead account for more than one-fifth of the education gap between the rich world and the poor world."
1
8
20
2,089
On how heat is hostile to human health, productivity, and social cohesion -
New from Works in Progress by @deenamousa:
How higher temperatures can lead to more violence, disorder, and poverty.
worksinprogress.news/p/heat-…
6
22
6,659
The link between iron supplementation and malaria risk is understudied but really interesting— I wrote a bit about this in @WorksInProgMag
worksinprogress.news/p/anemi…
- We've focused our iron grantmaking in India and may be missing opportunities in Sub-Saharan Africa. While there are concerns about iron increasing malaria risk, this issue may be less problematic when paired with malaria prevention programs we already fund.
4
21
2,353
Great post - becoming familiar with (and able to make the strong version) of arguments behind grievances with foreign aid is especially important right now
It's taken me a little while to collect my thoughts on DOGE. Here are 50 of them. I'd appreciate your feedback. statecraft.pub/p/50-thoughts…
4
22
3,503
New post about a few "big if true" health interventions that might be much more important than we think. They’re strange, they’re striking - and they need more research.
1
7
20
3,362
Internet access seems likely to get significantly more important over time given LLMs. Fortunately, the trend seems to have strongly pulled in that direction by default -
Internet access unlocks economic and educational opportunities.
Using (free) geospatial data, we can now identify which areas lack internet access at super granular levels.
Here's the breakdown:
4
20
4,927
Historically, happiness is highest among the young & old, dipping in middle age. One of the most robust findings in this literature - even replicating in chimps!
Young people (esp 18-25 y/o women) are much more unhappy today, breaking this relationship for the first time.
since this is getting traction, the most important thing here is that while gdp and happiness/life satisfaction are strongly correlated over time, there appears to be, especially in high income countries, a decoupling AND recent downward trend
1
3
20
2,599
My contribution to the AI job destruction discourse: Why are radiologists still around?
Radiology seems like one of the most intuitively automatable white collar professions: digital, pattern-recognition based, lots of existing data... I discuss the many predictions of the industry's demise, and what has actually happened since.
This is a precursor to a @WorksInProgMag article (out next week) in which I discuss *why* radiologists continue to be gainfully employed, and what lessons we might apply from that to the near future of other white collar professions.
3
2
19
1,601
I was recently surprised to learn how little we know about pain.
For example, Americans have been experiencing more chronic pain over time, and we're not entirely sure why 1/
2
4
18
2,629
Malaria parasites thrive on iron. In endemic areas without strong prevention and treatment, iron deficiency may ironically be keeping children safer from severe illness & death — even with the toll of anemia. 2/
2
1
19
1,398
Great paper on whether conditional unemployment benefits can encourage rural-urban migration in LMICs:
The treatment resulted in a 200% increase in the probability of moving to Nairobi vs the control (and fared better than unconditional cash transfers)
1
3
17
13,753
Great to be at the Abundance book launch with @ezraklein & @DKThomp in NYC last night. A core, striking idea from the book: The U.S. isn’t failing to solve housing, healthcare, and energy shortages because it can’t. It’s failing because it won’t. 🧵
2
1
18
1,143
Unironically, concrete is fascinating and deeply important to economic development and climate policy. A thread on why you should care. 🧵
my dad just called me to tell me he's at a concrete convention called concrete world and when I asked what he was doing there he said "learning about concrete"
1
2
17
998
Cool piece from @KelseyTuoc on whether LLMs are biased by the language the asker is using — mildly surprising to me that the answer was 'no'! (h/t @otis_reid)
1
2
16
3,691
Good episode from @DKThomp about AI in medicine - engages well with how tech interacts with existing healthcare incentives.
And with some nice discussion of my recent article on radiology, quoted below!
In 2016 Geoffrey Hinton said “we should stop training radiologists now" since AI would soon be better at their jobs.
He was right: models have outperformed radiologists on benchmarks for ~a decade.
Yet radiology jobs are at record highs, with an average salary of $520k.
Why?
2
3
16
24,936
Just published this piece about a topic that's been on my mind: The choices we're really making when we decide how to interact with technology - and how those stakes are getting higher with AI - using Nozick's philosophical thought experiment as a lens
A coincidental double anniversary: it's 50 years since Robert Nozick's 'experience machine' proposal and 25 years since The Matrix. @deenamousa on what Nozick got right and wrong about people's attitudes towards simulation vs reality bbc.com/future/article/20240…
2
3
14
2,481
Interesting new NBER working paper: Most people worry about outdoor air pollution, but, in many homes, indoor air is worse. A field experiment in London tested whether showing people their indoor pollution levels in real time might change behavior and finds a 17% reduction. 🧵
2
5
16
1,116
Days like this I'm really proud to be at @open_phil
There is no science without reliable data. Who pays the ongoing costs has been an issue from the beginning. This is a shared infrastructure problem but it gets treated like a “one-off” situation whenever it comes up. We need better long-term strategies. #Funding #DataScience
15
683
JP Morgan modeling when Labubus will peak looks like if the solar capacity predictions were right the first time
15
895
Consumerism final boss: lining up at 2am for the Starbucks bear-shaped cup and re-selling for $500 on eBay
2
15
2,577
New post: When does more automation mean more human workers?
One argument I made in a @WorksInProgMag piece is that, if AI made reading scans quicker, this might result in *more* jobs for radiologists, rather than fewer.
How does this apply to other professions? 🧵
2
3
14
2,299
One of the dangers of rising polarization and mistrust is that people start rejecting even the most well-documented facts—like that smoking is catastrophically bad for you.
There's no hidden agenda here; and the government has plenty of self-serving reasons to care too 🧵
1
3
12
1,132
A look back at some of the amazing things our grantees on the Global Health and Wellbeing portfolio achieved in 2024
From @OpenNYForAll helping to pass NYC's largest zoning reform in 50 years to @TheHumaneLeague improving welfare for 350M chickens, our grantees achieved incredible progress in 2024. Read more "cool things" from our Global Health & Wellbeing portfolio here: openphilanthropy.org/researc…
14
1,488
Article is here: worksinprogress.news/p/anemi…
And if you’re interested in more global health & development content, I write a regular digest of what I’ve been reading / writing / thinking about here: deenamousa.substack.com
1
13
1,063
Great thread (and report) about the health burden of addiction and why pharma fails to fill the gap
1
3
13
3,837
Highly recommend applying- @IFP does great work!
🚨 Today is the last day to apply for multiple open positions on our emerging technology team.
Come work with @fiiiiiist on AI, chips, and other cutting edge policy issues.
(And don’t forget about the $3,000 referral bonus!)
1
3
13
6,571
My prediction is that this will lead to a huge bifurcation in ability between students using AI to do assignments for them, and those using it as a tutor to learn more effectively. Over 4+ years, that difference likely compounds massively.
1
13
328
It’s giving season! 🥳 This post full of recommendations for individual donors from @open_phil staff is now out. Sharing a thread of a few of my favorite orgs on the list 🧵
2
3
11
1,510
The prevalence of sickle-cell anemia in malarial regions is a useful analogy. Having one copy of the gene is protective against malaria — a disease so deadly it’s ‘worth’ evolving other life-threatening diseases to avoid. 3/
1
12
1,232
Climate change gets $1.3 trillion in annual funding. But only 7% of that—$63B—goes to helping people adapt to rising temperatures, floods, and extreme weather. Why does adaptation funding lag so far behind mitigation? 🧵
2
2
11
499
New Substack post is out! I've been thinking about how much systems shape what’s possible—whether in nature or policy. From rove beetles mimicking ants to the chaos of foreign aid stop-orders, here’s what I learned this month 🧵
1
1
12
845
We treat QALYs as a uniform unit of measure, but every QALY may not be the same. We may care more about QALYs when we're sickest. This makes sense, given how we think about income increases, but isn't how we evaluate health.
I write about why in a new post, linked below.
1
2
11
1,094
Just out in @techreview: I look at the companies using AI to measure how much pain patients are in based on everything from involuntary facial movements, to heart rate, to peripheral temperature changes.
Will this oust the classic self-reported 1-10 scale?
2
2
9
3,951
Good to have it empirically confirmed that people do walk faster these days (group encounters also declined). Though not clear to me why.
Are people busier? Making different time trade-offs? Less likely to bump into people they know? Seeing public spaces as less appealing?
A neat new NBER paper finds that the average pedestrian walking speed in NY, Boston, and Philadelphia was 15% faster in 2010 than in 1980.
3
10
1,786
One of the most interesting things I've had the chance to learn about recently: How Rwanda is tackling the explosive potential of Lake Kivu by converting its trapped methane into a power source.
nationalgeographic.com/envir…
3
10
5,213
Great article in the latest @WorksInProgMag on how gene drives can be used to eliminate malaria. Remarkably, a decade after the tech’s discovery, we’re still waiting for a wild release
worksinprogress.co/issue/the…
2
4
11
3,921
Snakebite kills tens of thousands of people a year and disables hundreds of thousands more. Anti-venom research has barely progressed in decades.
This study, using antibodies from a man who deliberately let snakes bite him 200 times, suggests a path to a universal anti-venom. 🧵
1
1
11
1,141
Just learned that in NYS there's a paranormal activity disclosure rule, requiring sellers to disclose if their home is allegedly haunted
1
10
2,453
Wrote a companion article to my @techreview piece about companies using AI to measure how much pain people are in.
This one focuses on the implications: what happens when we quantify something we don’t fully understand, and what clinical judgment might we lose in the process?
Just out in @techreview: I look at the companies using AI to measure how much pain patients are in based on everything from involuntary facial movements, to heart rate, to peripheral temperature changes.
Will this oust the classic self-reported 1-10 scale?
1
1
10
1,220
A recent article by @jburnmurdoch suggests the discourse about unemployment among college grads has missed an important nuance: equivalently-experienced non-graduates are faring even worse
This complicates the claims from some that the job market woes are AI-driven
3
3
9
2,124
Great article on human challenge trials featuring
@1daysooner & @joshcmorrison's work aiming to accelerate treatment and vaccine development - including, potentially, future Hep C trials
1
2
10
758
I foolishly assumed they would say something that supports the idea that AI is disappearing jobs in the article about how AI is disappearing jobs
1
10
504
Could not be more excited for Works in Progress to be on my coffee table
Works in Progress is now available as a print magazine. I think it's the most beautiful and readable magazine I've ever seen.
Subscribe today for $100/£75 to receive six beautiful, 120-page issues a year.
worksinprogress.co/print
10
828
Fortunately, pairing iron supplements with malaria control—like bed nets, rapid diagnostics —eliminates the risk, and new vaccines like R21 provide reason for optimism. But we need more evidence to resolve this question and make global health efforts more effective. 4/
1
9
1,222
Very excited to release the findings of my thesis on the effect of language fluency on nudges in Harvard Business Review Arabia. Endless thanks to @makkifadi and @Helena_V_K for their collaboration and support.
مفتوح للجميع، مقال خاص لـ منصة هارفارد بزنس ريفيو العربية، بعنوان: كيف يمكن للغة أن تقف عائقاً أمام التدخلات السلوكية المثبتة؟، لـ قلم: فادي مكي @makkifadi ، وهيلينا فلادينجا كلاوزنييسر @Helena_V_K ، ودينا موسى @deenamousa
hbrarabic.com/?p=299023
2
1
8
Just found out about @TryBoredCow - which makes a product similar to cow's milk, but without the cows. They use microflora to produce the whey protein instead. Extremely similar process to how we make medical insulin today!
1
9
319
New York City just launched the first congestion pricing program in the U.S. Drivers entering Manhattan south of 60th Street now pay up to $15 per trip. It’s controversial, but cities that have tried it have seen less traffic, better air quality, and more funding for transit. 🧵
1
8
894
Excited for this absolutely stacked @ea_global
Really excited to be welcoming @iqbaldhali (@JPAL), @rachelbnfield (@Cgdev) and Zac Hatfield-Dodds (@AnthropicAI) to EA Global Boston 2024.
🚨 Applications close Sunday! 🚨
More info and how to apply:
effectivealtruism.org/ea-glo…
1
9
904
Dairy is another key source of iodine — but many are switching to alternatives like soy or oat milk, which often aren’t fortified. Women are making that swap at a higher rate. And 40% of prenatal vitamins lack iodine, in spite of the mineral’s importance during pregnancy. 5/
1
3
8
854
Useful to keep in mind when you see these kinds of headlines: usually, funders — including and especially the US gov — have a reason for supporting a piece of research. If one isn’t referenced, you should try to find out what it is before passing judgement.
Research that can seem frivolous when described uncharitably (i.e "studying the sex lives of bees" ) often has crucial applications (bee population are key for agricultural productivity!).
8
394
Computational protein design is one of the clearest early examples of AI accelerating fundamental scientific progress. Cool (and potentially extremely impactful) work happening at @UWproteindesign and @KingLabIPD
openphilanthropy.org/researc…
8
732
Any look into college students’ flight to corporate needs to start here: the high and rising levels of risk aversion among recent graduates from top schools
A journey into my college-years existential crises for @TheEconomist Christmas special issue, where I head to the manicured lawns of the Ivy League — and find a lot more resume-polishing than wokery.
Take a look here:
economist.com/christmas-spec…
8
609
Turns out ~30% of pipes in NYC may contain lead - @nylcv made an interactive map so you can see where they are, using using Department of Environmental Protection data
2
2
8
412
Collected my merch & formal certificate of membership in @nycgov's Rat Pack — part of the DoH's recent focus on sanitation interventions and reducing the rat population in the city. Link below if you're interested in the program!
2
7
447
