Ryan J. Gallagher
@ryanjgallag.com
Applied scientist trying to make the internet a little better. PhD. Trust & safety, platform manipulation, networks, fingerstyle guitar. I use my hair to express myself. He/they
Reposted by Ryan J. Gallagher
“Should we follow the Geneva Conventions?” Pete Hegseth asked in 2024.
It’s almost as if there were signs!
bsky.app/profile/anna...
It’s almost as if there were signs!
bsky.app/profile/anna...
November 29, 2025 at 3:05 AM
“Should we follow the Geneva Conventions?” Pete Hegseth asked in 2024.
It’s almost as if there were signs!
bsky.app/profile/anna...
It’s almost as if there were signs!
bsky.app/profile/anna...
"I'm a methods they/them" as I reflect and realize all my favorite papers are either theoretical or qualitative
November 26, 2025 at 11:35 PM
"I'm a methods they/them" as I reflect and realize all my favorite papers are either theoretical or qualitative
Reposted by Ryan J. Gallagher
Here it is, the most damning academic figure in the history of all figures: Russian IRA accounts deeply embedded in both sides of the 2016 Twitter BLM "discussion" (by @katestarbird.bsky.social & team, 2018).
THIS is what our information ecosystem supports. Twitter just made it 1% more visible.
THIS is what our information ecosystem supports. Twitter just made it 1% more visible.
November 25, 2025 at 1:30 PM
Here it is, the most damning academic figure in the history of all figures: Russian IRA accounts deeply embedded in both sides of the 2016 Twitter BLM "discussion" (by @katestarbird.bsky.social & team, 2018).
THIS is what our information ecosystem supports. Twitter just made it 1% more visible.
THIS is what our information ecosystem supports. Twitter just made it 1% more visible.
This is so reflective of my experience with failed machine learning initiatives
No data at all, no willingness to invest in annotation, no budget for the compute, no support integrating into the backend, no capacity to surface results in the UI, no tolerance for monitoring quality, and so on
No data at all, no willingness to invest in annotation, no budget for the compute, no support integrating into the backend, no capacity to surface results in the UI, no tolerance for monitoring quality, and so on
November 25, 2025 at 11:30 AM
This is so reflective of my experience with failed machine learning initiatives
No data at all, no willingness to invest in annotation, no budget for the compute, no support integrating into the backend, no capacity to surface results in the UI, no tolerance for monitoring quality, and so on
No data at all, no willingness to invest in annotation, no budget for the compute, no support integrating into the backend, no capacity to surface results in the UI, no tolerance for monitoring quality, and so on
Reposted by Ryan J. Gallagher
We also provided the first-ever full data archive of @bsky.app complete with hydrated objects at the point of ingest alongside a dozen ML classifications, for the cheapest possible rate, to other vetted builders and researchers in the ecosystem
Announcing the Graze Archives
A brief tour of the S3 requestor-pays archives of the Graze turbostream and freshly-announced megastream
www.graze.social
November 22, 2025 at 7:59 PM
We also provided the first-ever full data archive of @bsky.app complete with hydrated objects at the point of ingest alongside a dozen ML classifications, for the cheapest possible rate, to other vetted builders and researchers in the ecosystem