Sign in

Michael Eddy

@michaeleddy.bsky.social
405 followers 338 following 138 posts

Helping put social science to use for society at Stanford Impact Labs Twitter: @michaeleddy 🏳️‍🌈

PostsRepliesMedia
Michael Eddy @michaeleddy.bsky.social · 18/03/2026
What do 81k people want from AI? Fascinating new report from Anthropic captures their hopes, fears and concerns. Interesting to see such stark differences between how people responded in developing countries vs high-income countries. ⤵️ www.anthropic.com/fe...
010
Michael Eddy @michaeleddy.bsky.social · 19/02/2026
AI has limits. It may shift shortlists. Distinguishing bias vs improvement is complex & we have more work to do. Our study is small & shouldn’t be overinterpreted But if we care about how scarce resources get allocated, we should test these tools openly Learn more 👇
impact.stanford.edu
Responsible AI to Accelerate Research Funding - Stanford Impact Labs
Stanford Impact Labs used AI to streamline grant reviews, matching human decisions faster and cheaper, providing insights while keeping humans in the loop for accountability.
000
Michael Eddy @michaeleddy.bsky.social · 19/02/2026
Responsible AI use mattered: • Applicant transparency + opt-out • Data security & privacy protections • Clear human accountability • Human-in-the-loop review • Explicit learning agenda AI supported decisions. Humans made them.
100
Michael Eddy @michaeleddy.bsky.social · 19/02/2026
In addition to backtesting, we also ran this in parallel with our live 2025–26 cycle (in a sandbox). Decision-makers only saw AI reports after our usual process was done--allowing us to test whether AI added information, not just speed. It did.
100
Michael Eddy @michaeleddy.bsky.social · 19/02/2026
It was also much faster and cheaper. ~Two days; 10x cheaper than our standard first round. And we surfaced expert-level insights about 4 months earlier than usual.
100
Michael Eddy @michaeleddy.bsky.social · 19/02/2026
We back tested on four years of data (209 applications; 67 shortlisted; 28 funded). On our strongest signal of quality—who ultimately got funded—AI performed on par with generalist reviewer rankings. It surfaced somewhat different shortlists, but matched humans on funding decisions.
100
Michael Eddy @michaeleddy.bsky.social · 19/02/2026
The output wasn’t just a ranked list. It generated structured, comparative explanations about why proposals were stronger or weaker relative to others. That kind of pool-wide comparison is hard to do consistently with humans alone
100
Michael Eddy @michaeleddy.bsky.social · 19/02/2026
Then we changed the core question. Instead of asking, “Is this proposal good?” We asked, “Which of these two is stronger, and why?” That led to 1,000+ head-to-head matchups across the full pool. Think March Madness but on steroids.
100
Michael Eddy @michaeleddy.bsky.social · 19/02/2026
Instead of “chatbot vibes,” we built a structured multi-agent system w/ Claude Code. Not just replicating what reviewers do, but extending it: literature reviews, landscape scans, backgrounders & structured rubric scoring--previously not possible for early-stage proposals
100
Michael Eddy @michaeleddy.bsky.social · 19/02/2026
At Stanford Impact Labs, we asked a straightforward question: Can AI help us accelerate science funding without lowering standards or cutting humans out of the loop? We put it to the test.
100
Michael Eddy @michaeleddy.bsky.social · 19/02/2026
Funders, public procurement, and journal editors all face similar challenges: how to allocate scarce resources most effectively Most applications are rejected. Decisions can take months. We've all been there. That delay is a huge invisible tax on science & impact Could AI reduce that tax?
100
Michael Eddy @michaeleddy.bsky.social · 12/12/2025
And relatedly, from Erica Goldman, at FAS fas.org/publication/...
fas.org
Universities should help shape the vision for new NSF Tech Labs
At a time when universities are already facing intense pressure to re-envision their role in the S&T ecosystem, we encourage NSF to ensure that the ambitious research acceleration remains compatible w...
010
Michael Eddy @michaeleddy.bsky.social · 12/12/2025
"Science Funding Goes Beyond the Universities" by @calebwatney.bsky.social, IFP www.wsj.com/opinion/scie...
wsj.com
Opinion | Science Funding Goes Beyond the Universities
A new NSF initiative will support teams of researchers, a departure from the traditional practice of giving grants to individual professors.
110
Michael Eddy @michaeleddy.bsky.social · 12/12/2025
What does it take to build a scientific field? New @RenPhilanthropy playbook out with lessons learned from successful cases below 👇 www.renaissancephila...
000
Michael Eddy @michaeleddy.bsky.social · 11/12/2025
We all need stronger BS filters. As signal gets harder to distinguish from noise, real rigor becomes rarer—and more valuable. 3/3
020
Michael Eddy @michaeleddy.bsky.social · 11/12/2025
In the conferences I'm attending & across social media, I'm seeing more claims that sound good, but collapse under basic scrutiny. The framing is slick; the rigor is missing. Bad ideas have always existed. What’s new is how effortlessly they can now masquerade as insight. 2/3
120
Michael Eddy @michaeleddy.bsky.social · 11/12/2025
In the age of AI, sounding right is easier than being right. It’s never been simpler to spin ideas that 𝘧𝘦𝘦𝘭 true—confident tone, polished narrative. AI can turn half-baked thoughts into viral TED-talk soundbites. But rhetorical sparkle isn’t substance. 1/3
110
Michael Eddy @michaeleddy.bsky.social · 24/11/2025
Dunning-Kruger effect in practice: In 53k forecasts across 100 projects, people who are more confident are actually much less accurate. New 📄 by @evavivalt.bsky.social & @sdellavi.bsky.social
011
Michael Eddy @michaeleddy.bsky.social · 17/11/2025
Check out the new Export Boom Atlas from @GrowthTeams–an interactive map of 80+ export booms from developing economies, showing how countries from Vietnam to Morocco to Costa Rica achieved rapid sector-level growth that created jobs and prosperity. 🌍 exportbooms.org”
exportbooms.org
Export Boom Atlas by Growth Teams
Explore the playbooks behind 82 export booms in emerging economies since 1995
000
Michael Eddy @michaeleddy.bsky.social · 06/10/2025
I love it when papers have accompanying websites! also, neat when i found out @stanfordimpactlabs.bsky.social had a small role in supporting this work! 📄 & 💻by @marshallburke.bsky.social & a impressive team! adaptationatlas.org/... www.nber.org/papers/...
0102
Michael Eddy @michaeleddy.bsky.social · 29/09/2025
@ukri.org has announced a £11.5m investment in AI-driven evidence synthesis (METIUS). Backed by a $126m global alliance including @wellcometrust.bsky.social, the project aims to make research more accessible for policymakers on climate, education, justice and development.
010
Michael Eddy @michaeleddy.bsky.social · 25/09/2025
Fascinating paper on a timely question... kudos!!
000
Reposted by Michael Eddy
Matt Clancy @mattsclancy.bsky.social · 25/09/2025
New research by Pierre Azoulay, Danielle Li, Bhaven Sampat and me. Earlier this year, the President’s budget proposed a 40% cut to the budget of the NIH. This motivated us to ask: what if the NIH had been 40% smaller?
22617
Michael Eddy @michaeleddy.bsky.social · 23/09/2025
With a few notable exceptions, I’m struck by how few funders are openly sharing what they’re learning from AI in decision-making. What am I missing?
010
Michael Eddy @michaeleddy.bsky.social · 23/09/2025
We’ve embedded this directly in our latest RFP. (screenshot 👇) This is about experimentation, transparency, and learning—together with the research & funding community.
100
Michael Eddy @michaeleddy.bsky.social · 23/09/2025
Here’s what “responsible” means to us: ➡️ AI augments—not replaces—human decision-making ➡️ New workflows that weren’t previously possible ➡️ Robust safeguards ➡️ A learning agenda to test key claims
100
Michael Eddy @michaeleddy.bsky.social · 23/09/2025
That’s why @StanfordImpact is piloting responsible AI use in our funding processes. Our goal: accelerate impact-focused science R&D. ✅ Faster decisions ✅ Lower costs ✅ Reduced burden on applicants
100
Michael Eddy @michaeleddy.bsky.social · 23/09/2025
Right now: ⚠️ Many funders ban AI outright. 🙈 Others ignore it. 🤖 And slick AI vendors make bold, untested claims. None of this felt right to me
100
Michael Eddy @michaeleddy.bsky.social · 23/09/2025
Ever applied for funding… and then had to wait months for a response? ⏳ What if funders could move faster and make better decisions—so applicants can secure funding & get straight to work? AI could help. But the current landscape is messy. 🧵
110
Michael Eddy @michaeleddy.bsky.social · 16/09/2025
After a disastrous launch, the Dept. of Education's new team fixed the FAFSA system, helping 14M+ students. But the GAO criticized their modern methods with an outdated playbook. It's high time we focus on outcomes not broken processes, argues @pahlkadot
eatingpolicy.com
GAO gets schooled by the Department of Education
The FAFSA team snaps back, then punches back
000
Michael Eddy @michaeleddy.bsky.social · 10/09/2025
I wonder why more academics don't publish their papers with companion websites? I find it so much more enjoyable reading an interactive site! "AI Agents for Economic Research" by @antonkorinek www.genaiforecon.org
genaiforecon.org
Generative AI for Economic Research
000
Michael Eddy @michaeleddy.bsky.social · 05/09/2025
I remember meeting Karthik & @singhabhi.bsky.social w/ the Mindspark/EI team when I was working at Global Innovation Fund back in 2017—and being blown away by how they were designing for impact @ scale from day 1. A stellar example of academics + do-ers partnering for scale 🚀📈👇
041
Michael Eddy @michaeleddy.bsky.social · 04/09/2025
Kudos @benjaminsmanning.bsky.social & @johnjhorton.bsky.social for this important contribution! Not to mention, aiming to put your paper to use at scale with @expectedparrot.bsky.social 🦜 www.expectedparrot.com
expectedparrot.com
Expected Parrot
010
Michael Eddy @michaeleddy.bsky.social · 04/09/2025
my main concern is that these lab-style games don't mirror interventions at scale.... But they do provide a controlled testbed to explore whether underlying mechanisms generalize, and if mechanisms don't generalize, unlikely you're intervention will generalize either.
100
Michael Eddy @michaeleddy.bsky.social · 04/09/2025
Impact-focused funders often ask: if it worked here, will it work there? This paper is a small, but impt step toward a generalized approach (grounded in theory + human data) to generalize research across settings—core to putting social science to use at scale.💪
100
Michael Eddy @michaeleddy.bsky.social · 28/08/2025
Science faces a replication crisis: billions wasted, progress slowed, policy misled A new idea from @i4replication.bsky.social Brodeur & Barbarioli in @ifp.bsky.social's #LaunchEngine imagines how to build automated replication infrastructure for *better* & *faster* science ifp.org/the-replicat...
ifp.org
The Replication Engine | IFP
How to build automated replication infrastructure for better, faster science
062
Michael Eddy @michaeleddy.bsky.social · 27/08/2025
Now, go read the @nber.org paper! www.nber.org/papers/w34164 christopherneilson.github.io/work/Applica...
nber.org
College Application Mistakes and the Design of Information Policies at Scale
Founded in 1920, the NBER is a private, non-profit, non-partisan organization dedicated to conducting economic research and to disseminating research findings among academics, public policy makers, an...
020
Michael Eddy @michaeleddy.bsky.social · 27/08/2025
Authors: @anaisfabre.bsky.social at IFS; Chris Nielson @yaleeconomics.bsky.social; Ignacio Rios @utdallas.bsky.social; & Tomas Larroucau @arizonastateuni.bsky.social with collaborators at Chile’s Min of Education @mineducchile.bsky.social Kudos to everyone involved!! 🎉
100
Michael Eddy @michaeleddy.bsky.social · 27/08/2025
This wasn’t a one-off: the policy scaled nationwide with persistent enrollment gains—evidence that “designed-for-scale” information can work in real systems.
100
Michael Eddy @michaeleddy.bsky.social · 27/08/2025
Results: previously unmatched students were 44% more likely to get an offer; placements into higher-ranked programs rose 20%
100
Michael Eddy @michaeleddy.bsky.social · 27/08/2025
The intervention: show applicants tailored admission probabilities + program info + smart nudges (e.g., "don’t forget a safety school"). Delivered inside the application portal.
100
Michael Eddy @michaeleddy.bsky.social · 27/08/2025
Few things I love more than when I see a paper report on how research was put to use at scale!!👇 New study on fixing college admissions. In Chile’s centralized system, personalized, in-platform info helped students avoid common app mistakes & improved outcomes🧵⬇️
100
Michael Eddy @michaeleddy.bsky.social · 18/08/2025
Link to @apoorvanyt.bsky.social NYT coverage: archive.ph/ct2o0 Link to study: nber.org/papers/w34152 Link to @give-directly.bsky.social summary: givedirectly.org/mortality2025/ www.nytimes.com/2025/08/18/h...
nytimes.com
A New Way to Reduce Children’s Deaths: Cash
072
Michael Eddy @michaeleddy.bsky.social · 18/08/2025
Policy takeaway: Pair income support with access to obstetric care. Cost-effective, scalable complement to maternal/child health programs—and a great example of research that fuels real-world impact
100
Michael Eddy @michaeleddy.bsky.social · 18/08/2025
Mechanisms: big jump in hospital deliveries (~45%); largest declines in neonatal/maternal causes; impacts concentrated among poorer households; effects fade after cash stops
100
Michael Eddy @michaeleddy.bsky.social · 18/08/2025
Was stunned when I saw these results...👇 In rural Kenya a $1k transfer cut infant & child mortality in ~ *half* with largest drops around delivery & near physician-staffed facilities 😲 Kudos to @tedmiguel.bsky.social @nshankar.bsky.social @eggerdennis.bsky.social & others involved in the study!🧵
182
Michael Eddy @michaeleddy.bsky.social · 08/08/2025
And I just found @daveevansphd.bsky.social on bsky… so I’m still figuring out my way around here too!!
000
Reposted by Michael Eddy
Alondra Nelson @alondra.bsky.social · 06/08/2025
Great 🧵 about our recent @science.org policy essay arguing we need an evidence base for effective AI policy—plus support and protections for the researchers doing this important work. Thank you @michaeleddy.bsky.social!
1208
Michael Eddy @michaeleddy.bsky.social · 05/08/2025
Authors w/ bsky: @profsanjeevarora.bsky.social @drfeifei.bsky.social @randomwalker.bsky.social @jurafsky.bsky.social @yejinchoinka.bsky.social @emmapierson.bsky.social @schwarzinexile.bsky.social nexile.bsky.social @gaelvaroquaux.bsky.social aux.bsky.social @geomblog.bsky.social . Go follow them!
log.bsky.social
Bluesky
051
Michael Eddy @michaeleddy.bsky.social · 05/08/2025
Check out the full article 👇 #AIGovernance www.science.org/doi/10.1126/...
science.org
131