Johan Ugander
banner
jugander.bsky.social
Johan Ugander
@jugander.bsky.social
Associate Professor, Yale Statistics & Data Science. Social networks, social and behavioral data, causal inference, mountains. https://jugander.github.io/
Link to that paper: Slaughter, Peytavin, Ugander, Saveski (2025) "Community notes reduce engagement with and diffusion of false information online", PNAS. www.pnas.org/doi/10.1073/...
Community notes reduce engagement with and diffusion of false information online | PNAS
Social networks scaffold the diffusion of information on social media. Much attention has been given to the spread of true vs. false content on onl...
www.pnas.org
November 26, 2025 at 7:37 PM
Does Claude do badges? Claude should do badges.
October 28, 2025 at 6:24 PM
I had not seen that before, thanks for the pointer!
October 27, 2025 at 1:58 AM
I don't agree that all these are examples where ethics let slip in a way "that wouldn't be acceptable for academics". Would love to chat in longer former about all this, best off social media :)
October 25, 2025 at 4:49 PM
FWIW, I am deeply sympathetic to the thesis of the paper—for over a decade I have encouraged students interested in industry research to read Milosz's The Captive Mind in preparation. So I want to help you strengthen the work by encouraging precise arguments.
October 25, 2025 at 2:57 PM
Huszár et al. wasn't an experiment, it was observational. And the UCSD and Cornell studies both went before IRB. So the journals didn't waive anything, as far as I know. I feel like precision here is quite important, because the history of these studies is full of muddled details.
October 25, 2025 at 2:56 PM
Thanks, it just didn't seem like any of the three I knew the details of were "examples" of the previous sentence (journals waiving IRB).
October 25, 2025 at 2:40 AM
Am confused…
- Bond et al [33] was approved by UCSD IRB.
- Thomas & Wahedi [39] both authors had only Meta affils; no university IRB had jurisdiction.
- Huszar et al [75] dunno, cc @inference.vc.
- Kramer at al [82] was exempted; much has been written about this case. See the PNAS ed statement.
October 24, 2025 at 6:34 PM
Reposted by Johan Ugander
Have you ever wondered how to ensure a loss in Candyland though?

github.com/mkiang/candy...
github.com
October 6, 2025 at 4:40 PM