Karandeep Singh
banner
kdpsingh.bsky.social
Karandeep Singh
@kdpsingh.bsky.social
Faculty at UC San Diego. Chief Health AI Officer at UC San Diego Health. #rstats. Creator of Tidier.jl #julialang. #GoBlue. Views own.
Pinned
No amount of AI can overcome bad health policy.
Reposted by Karandeep Singh
👀
November 11, 2025 at 3:45 AM
Reposted by Karandeep Singh
80 layers — for those not paying attention, @dorialexander.bsky.social has been posting for weeks about how small models with deep rather than wide layers exhibit eerie emergent behavior

this one is worth checking out
Synthetic playgrounds enabled a series of controlled experiments that brought us to favor extreme depth design. We selected a 80-layers architecture for Baguettotron, with improvements across the board on memorization of logical reasoning: huggingface.co/PleIAs/Bague...
November 11, 2025 at 2:26 AM
Reposted by Karandeep Singh
Synthetic playgrounds enabled a series of controlled experiments that brought us to favor extreme depth design. We selected a 80-layers architecture for Baguettotron, with improvements across the board on memorization of logical reasoning: huggingface.co/PleIAs/Bague...
November 10, 2025 at 5:32 PM
November 10, 2025 at 11:34 PM
Reposted by Karandeep Singh
When Damon Landor, a Rastafarian with dreadlocks, showed a prison guard a court ruling protecting his right to keep his hair, the guard threw it away. Despite his pleas, officers restrained him and cut his hair.

Now, he's going to the Supreme Court.
A Rastafarian goes to the Supreme Court after a prison shaved his dreadlocks
The question is not whether Damon Landor was mistreated but whether he can sue the individual prison officials and guards for damages.
wapo.st
November 10, 2025 at 7:00 PM
Reposted by Karandeep Singh
journals.lww.com
November 10, 2025 at 12:03 PM
Reposted by Karandeep Singh
Hope everyone stays safe in that 65 mostly sunny weather.
November 9, 2025 at 5:56 PM
“Mostly”???
November 9, 2025 at 5:21 PM
Reposted by Karandeep Singh
Any published papers in the wild with a declared use of AI for analyzing the data yet? We know people are doing it. Probably quite often. So there must be one where use was declared?
November 9, 2025 at 3:44 PM
November 8, 2025 at 8:25 PM
Reposted by Karandeep Singh
There could be six seven prs
Gotta get this rizz into #julialang @notreallydaniel.bsky.social
Y’all these genzplyr PRs are off the hook: github.com/hadley/genzp...
November 8, 2025 at 7:00 PM
Reposted by Karandeep Singh
Y’all these genzplyr PRs are off the hook: github.com/hadley/genzp...
Pull requests · hadley/genzplyr
dplyr but make it bussin fr fr no cap. Contribute to hadley/genzplyr development by creating an account on GitHub.
github.com
November 8, 2025 at 12:29 AM
This is a San Diego thing.
November 8, 2025 at 3:27 AM
"I'm a bit of a large X model user, myself." - no one
November 4, 2025 at 4:10 AM
Reposted by Karandeep Singh
On halloween there’s nothing spookier than a future without CDC
October 31, 2025 at 8:31 PM
Next up: Interactive Python coding... brought to you by Zoom.
October 31, 2025 at 9:59 PM
Vibe-based ROI: where you go to Rick Rubin‘s house and he tells you the ROI
October 30, 2025 at 12:36 AM
Reposted by Karandeep Singh
Here’s a clip 🎻😋 Share if you want to send some love and swing to others ❤️🌎 #tommyemmanuel #jazzviolin #onetake
October 18, 2025 at 4:16 AM
Reposted by Karandeep Singh
reticulate v1.44 is on CRAN!

Positron integration is now enabled by default. Enjoy R and Python REPL consoles powered by concurrent threads in the same process, zero‑copy R↔Python data transfer, and full Variables Pane support.

Full changelog: github.com/rstudio/reti...

#rstats #python #pydata
github.com
October 25, 2025 at 3:37 PM
Sky (for OS integration) + Net (via ChatGPT Atlas) = ….
October 25, 2025 at 8:26 PM
Nope nope nope nope nope.
October 24, 2025 at 8:08 PM
Excited to launch our Universal AI Disclosure, a best practice for ensuring that use of approved AI tools is accompanied by standard language to inform patients. This comes alongside internal guidance setting the expectation that we are responsible for our words regardless of how they were drafted.
October 23, 2025 at 10:32 PM