Sign in

Nathan Bartley

@nbartley.bsky.social
156 followers 898 following 126 posts

CS PhD from USC / social networks, games, recommender systems, audits, ⚽️

PostsRepliesMedia
Reposted by Nathan Bartley
YY Ahn @yyahn.bsky.social · 22/09/2026
As people have predicted, there are quite a few open-source alternatives to Jev, including: - Laya, this was the original according to him: github.com/NandhaKishor... - on Apple Core ML / neural engine: github.com/mizorewww/la... - with Qwen: github.com/jaredpalmer/...
github.com
GitHub - NandhaKishorM/laya
Contribute to NandhaKishorM/laya development by creating an account on GitHub.
1144
Reposted by Nathan Bartley
Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 23/09/2026
By the way, bad software engineering does regularly kill people, just not in an easily traceable way.
512215
Reposted by Nathan Bartley
Jay Van Bavel, PhD @jayvanbavel.bsky.social · 18/09/2026
People randomly assigned to a liberal or conservative group chat groups shifted their expressed stance in the direction of the group, demonstrating that social norms alters political expression This effect emerges after controlling for baseline ideology www.sciencedirect.com/science/arti...
86827
Reposted by Nathan Bartley
Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 23/08/2026
Really good overview of how arxiv is thinking about handling the incoming LLM wave and the considerations going into it
3509
Nathan Bartley @nbartley.bsky.social · 14/08/2026
The transparency is great — on the research side it’s great for reasoning about how the system works, but AFAICT there are no model weights in this release (understandable, but still). It’d be great to have even outdated weights to make stronger empirical claims about the effect of the feed
000
Nathan Bartley @nbartley.bsky.social · 31/07/2026
Feeling big #ic2s2 fomo this year, seems like a great conference (like always!)
020
Reposted by Nathan Bartley
Eamon Duede @eduede.bsky.social · 21/07/2026
New preprint with @carlbergstrom.com, Kevin Gross, and @mjcrockett.bsky.social. See Carl's thread. Surprise! Even if LLMs work perfectly, they'll make scientists refine their work less deeply, because they raise the opportunity cost of time, pushing researchers to do more, less well.
A figure describing the results.
1238
Nathan Bartley @nbartley.bsky.social · 09/07/2026
Very cool! A good reminder that there are other very useful network libraries besides just networkx and igraph
010
Reposted by Nathan Bartley
Olúfẹ́mi O. Táíwò @olufemiotaiwo.bsky.social · 04/07/2026
Cape Verde defeats Argentina via a 2-3 loss in extra time
7756221035
Nathan Bartley @nbartley.bsky.social · 04/07/2026
Argentina barely deserved the win
000
Nathan Bartley @nbartley.bsky.social · 04/07/2026
The World Cup is full of magic moments, loving Cabo Verde today
000
Nathan Bartley @nbartley.bsky.social · 02/07/2026
that was a good match to watch. The US was favored coming in but they definitely had to earn that win. I’m honestly fired up for a USA-BEL rematch #USAvBIH
100
Reposted by Nathan Bartley
Jake Gold @jacob.gold · 26/06/2026
Hoping Bluesky Communities launches before Musk manages to clean up X, which may take a while given the incentives at work, but maybe not. Even people who remain loyal to Musk know it has become a dumpster fire. It’s a big opportunity, but only if Bluesky provides space for many more communities.
creenshot of an X/Twitter thread in dark mode. Andrej Karpathy writes: “I’ve been on Twitter for almost 2 decades now so I can say with confidence that it has never been this toxic and Reddit-like. I think the alg actively encourages it and people get RL’d by it. Main reason I’ve been coming here less and posting less.” His post shows about 571 replies, 689 reposts, 7.4K likes, and 877K views.

Below it, Elon Musk replies: “We need a complete overhaul of the algorithm.” The reply is timestamped 10:49 PM, 6/24/26, and shows 1.5M views, about 1.9K replies, 1.3K reposts, 10K likes, and 635 bookmarks.
4463
Reposted by Nathan Bartley
the beggars out of here getter @milesklee.bsky.social · 26/06/2026
this Türkiye win was for Eric Adams 🇹🇷
1277874
Nathan Bartley @nbartley.bsky.social · 26/06/2026
That was a great game to watch, honestly that momentum change in the second half gave me a lot more faith in the US squad than anything from the first two matches #TURvUSA
010
Reposted by Nathan Bartley
Ruo Shui @ruoshuiresearch.bsky.social · 21/06/2026
Figure 2. Main Results

Displays three panels (Panel A, Panel B, and Panel C), each containing a density distribution plot on the left and an event-study/trend plot on the right. The graphs compare academic metrics across three groups or periods: Post AI (red/maroon), Pre AI (green outline), and Never AI (blue outline).

Panel A: Homework Score Left (Density Plot): Shows the distribution of homework scores. The "Post AI" distribution is shifted significantly to the right, showing a bimodal distribution with a major peak around 120, indicating higher overall homework scores after AI adoption. Right (Trend Plot): Plots homework scores against the "Months Since Start of Gen AI Use." The "Never-AI Students" (blue dots) remain flat at a baseline score of around 100. The "Ever-AI Students" (red red x's) match the baseline before month 0, but spike sharply upward after the start of AI use, leveling off at a score near 118 by month 5.

Panel B: Homework Completion Time (min.) Left (Density Plot): Shows the distribution of homework completion times. The "Post AI" distribution shifts sharply to the left, with a prominent peak around 45, indicating shorter completion times. Right (Trend Plot): Plots homework completion time in minutes against the months since AI use began. "Never-AI Students" stay consistent at around 64 minutes. "Ever-AI Students" show a rapid drop starting at month 0, falling and stabilizing around 45 minutes by month 5.

Panel C: Exam Score Left (Density Plot): Shows the distribution of exam scores. The "Post AI" distribution shifts significantly to the left, showing a broader distribution spreading toward lower scores (down to 60 or below).Right (Trend Plot): Plots exam scores against months since AI adoption. "Never-AI Students" remain steady near a score of 100. "Ever-AI Students" show a steep decline starting at month 0, dropping rapidly to a score of approximately 81 by month 5, where it remains flat through month 10.
2206
Nathan Bartley @nbartley.bsky.social · 20/06/2026
While this is a German court’s ruling, I imagine this adds fuel to the US debate on section 230
000
Reposted by Nathan Bartley
CJ Fogler @cjzero.bsky.social · 19/06/2026
absurd saves from Rangel with the replays
17538115
Reposted by Nathan Bartley
Greg Pak @gregpak.net · 19/06/2026
Rooting for South Korea but my god that was was an incredible double save from Mexico.
31218
Reposted by Nathan Bartley
Graze Social @graze.social · 18/06/2026
Announcing co/core: a social computing experiment from the Graze team
console.cocore.dev
co/core — an AI cooperative
co/core is a cooperative for AI inference — people pooling the Macs they already own to run open models for each other, instead of renting from the big clouds. An experiment in AI infrastructure we bu...
3035286
Nathan Bartley @nbartley.bsky.social · 18/06/2026
A few PDSes I haven't seen in awhile broke into the top ten in raw # of active repos (June 2026) 👀 Looking at repos on pds.ls suggests a fair amount of bsky spam. Are there any lists of 'spammy' PDSes (or do devs just roll their own)? I know @mackuba.eu tags some PDSes w/ high bot activity.
A table screenshot showing 6 top independent PDSes by number of active repos, led by brid.gy and eurosky.social
210
Nathan Bartley @nbartley.bsky.social · 10/06/2026
Very cool, helping people find people!
000
Reposted by Nathan Bartley
Dominik 🔮 @tkdodo.eu · 09/06/2026
I'd want the opposite of "mute thread" on @bsky.app as a feature: "subscribe to thread" à la slacks "get reply notifications". Really useful when someone asks a good question and I'm interested in seeing the answers.
510412
Nathan Bartley @nbartley.bsky.social · 08/06/2026
leaflet.pub
Who Needs Some Retention Attention?
A cohort-retention analysis of Bluesky: established cohorts are steady, new sign-ups go quiet after a day -- maybe fixable in onboarding.
161
Reposted by Nathan Bartley
Patrick Boueri @patrickboueri.com · 29/05/2026
So I've been digging into bluesky retention data to see what is preventing growth of the ecosystem. Using likes/posts/follows you can see who is active or dormant on ATProto. The headline is that only 5% of users are retained at 28d and 10% of current users churn. That's not pretty
Layer cake of cohorts who joined bsky that are still active -- dominated by Pre Q1 2025 influxRetention metrics showing new users are 5% likely to stay and churn is 10% of current user base (they get resurrected) A break down of growth that shows resurrections / new users / new churn / old churn. It's hovering around steady state
48212
Nathan Bartley @nbartley.bsky.social · 02/06/2026
Tezos has shot up the past couple of weeks with a ton of active repos 👀 presumably for this work tz2at.xyz
A picture depicting a table of the top 10 PDSes by total number of active repos. tz2at.store is relatively new and has rapidly become the second largest
010
Nathan Bartley @nbartley.bsky.social · 02/06/2026
Great to see pre-registered studies testing the effect of algorithms on perception!
010
Nathan Bartley @nbartley.bsky.social · 29/05/2026
an interesting counterpart to the X transformer-centered pipeline setup — just make the whole pipeline one big PyTorch model @spacecowboy17.bsky.social @patrickboueri.com engineering.fb.com/2026/05/26/m...
engineering.fb.com
SilverTorch: Index as Model — A New Retrieval Paradigm for Recommendation Systems
We’re introducing SilverTorch, a reimagining of recommendation systems that unifies all retrieval components for user generated content under a unified architecture.  SilverTorch shows up to 23.7x …
210
Reposted by Nathan Bartley
Bluesky @bsky.app · 28/05/2026
What is Standard.site? Think of it as a file format for publishing longer posts from blogs, news sites…wherever. If you’re using Leaflet, pckt.blog, or Offprint, you’re already using Standard.site! standard.site
standard.site
Standard.site - One schema. Every platform.
Standard.site provides shared lexicons for long-form publishing on AT Protocol. Making content easier to discover, index, and move across the ATmosphere.
1044958
Nathan Bartley @nbartley.bsky.social · 27/05/2026
a bird in tree is worth two in the bush
a bird scooter hanging in a tree
000
Nathan Bartley @nbartley.bsky.social · 27/05/2026
Made some updates to my PDS centralization dashboard atproto.barn.city following great feedback. Three things I'm hoping to show with this (to start): 1) Functional centralization 2) Where are repos being made? 3) Where are repos migrating to? 1/3
atproto.barn.city
ATProto Health
AT Protocol ecosystem health dashboard
111
Reposted by Nathan Bartley
spacecowboy @spacecowboy17.bsky.social · 17/05/2026
The X For You algorithm has been published: github.com/xai-org/x-al... Now we have the answer to how true this statement was: "Grok will literally read every post and watch every video" This was just marketing. There is no Grok reading any posts. 🧵
github.com
GitHub - xai-org/x-algorithm: Algorithm powering the For You feed on X
Algorithm powering the For You feed on X. Contribute to xai-org/x-algorithm development by creating an account on GitHub.
510113
Reposted by Nathan Bartley
kourge the jafnhár 🏳️‍🌈 @kourge.net · 12/05/2026
enough. i’m tired of github actions. it’s time for github consequences
292577665
Nathan Bartley @nbartley.bsky.social · 12/05/2026
Great resource — I was starting to build a follower graph from scratch and this makes that so much easier!
121
Nathan Bartley @nbartley.bsky.social · 08/05/2026
It’s D-day in sunny LA #trains
000
Reposted by Nathan Bartley
Bluesky @bsky.app · 05/05/2026
Keep it private! We’re adding more ways to connect with other Atmosphere apps. You can now launch Germ’s E2E encrypted messaging from Bluesky profiles. Check out @germnetwork.com for more.
Demonstration of where the Germ DM button appears on your Bluesky profile after you pair your accounts within the Germ DM app.
1782475405
Reposted by Nathan Bartley
Bryan (they/them) @chaosgreml.in · 04/05/2026
Okay, it’s here! pdscheck.dev is LIVE!!! Thanks to @bad-example.com and @pds.dad for all their work on debug.hose.cam, and thanks to Claude for doing most of the heavy lifting. License is CC0 with an MIT fallback, and source is at tangled.org/chaosgreml.in/pdscheck, PRs OH SO WELCOME!
pdscheck.dev
pdscheck — atproto PDS diagnostic
Drop in a handle, DID, or PDS URL. pdscheck runs probes against the PDS, the relay network, and PLC, then tells you what's broken and how to fix it. Runs entirely in your browser.
48717
Reposted by Nathan Bartley
Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 03/05/2026
The prevalence of acadamese in scientific writing is partly why writing overall in science has been devalued. The average scientific paper is boring to read and has a boilerplate style, a void into which LLMs jump
9603
Reposted by Nathan Bartley
Maria Antoniak @mariaa.bsky.social · 03/05/2026
“A structural change in research contribution and assessment is necessary to unshackle research communication from long-form narrative publication.”
2163
Reposted by Nathan Bartley
Anne Bertucio de Hollembeak @annebdh.bsky.social · 29/04/2026
TL;DR on AI-assisted or agentic coding: Some things change; some things stay the same. The same: - Code sent for review needs to be *reviewable* - Trust but verify, all the way through your SDLC - Getting far "off main" accrues technical debt - Risk tolerance informs where and how you experiment
3498
Reposted by Nathan Bartley
𝕮 @chrisshank.com · 29/04/2026
Next @atproto.la meetup is coming up on May 9th!
071
Nathan Bartley @nbartley.bsky.social · 27/04/2026
recently I’ve been more intentionally courteous to other drivers on the road (like letting people merge when it’s bumper-to-bumper and thanking those who let me merge), and I feel like my average road-stress levels have plummeted Maybe it’s spurious but would recommend
010
Reposted by Nathan Bartley
Internet Archive @archive.org · 24/04/2026
🧵 The web is disappearing 🕳️ From Pew: 26% of pages from 2013-2023 are ☠️ A new study in VANISHING CULTURE gives hope: 16% have been restored through the Wayback Machine. 56% are preserved before they disappear. 📚 Read VANISHING CULTURE free from Internet Archive: archive.org/details/vani... ⬇️
Infographic titled “URLs Preserved in the Wayback Machine” based on a dataset of 5.4 million URLs. It shows 26% of web pages are dead and 74% are still alive. Of the total, 10% have vanished completely, 18% are endangered (not yet preserved), 16% have been restored through the Wayback Machine, and 56% are preserved. The visual uses a flow-style diagram to show how preservation reduces permanent loss.
71027498
Reposted by Nathan Bartley
Joseph Seering @josephseering.bsky.social · 24/04/2026
Looks very interesting from a quick read of the first few pages, bookmarked to finish. Not an entirely new concept, but worth reminding the field (and those outside the field) where we really are.
021
Nathan Bartley @nbartley.bsky.social · 23/04/2026
IMO a pretty glaring false positive, this should be escalated to a human reviewer
051
Reposted by Nathan Bartley
Rude1 Haunted Badness. ⁂ @rude1.blacksky.team · 22/04/2026
Feeds should be portable. It would also allow the feed creator to delete or deactivate their account without taking a feed down.
514126
Reposted by Nathan Bartley
hailey @hailey.at · 19/04/2026
claude code is slowly becoming more and more unusable imo inside of serious code bases. not because its "broken" but because its deteriorating from where it was at even just 1 month ago, let alone 3. model iterations are (supposedly) much better but i am seeing less effectiveness with each model.
51126
Reposted by Nathan Bartley
Atmosphere Community @atmosphere.community · 17/04/2026
Yes, it's true! @brittanyellich.com crushed it in shipping a v0.1 version of a new community hub at atmosphere.community Features: @standard.site aggregation powered by @opensocial.community, group calendar, regional communities etc etc etc
Screenshot of atmosphere.community
57713
Nathan Bartley @nbartley.bsky.social · 17/04/2026
tvguide channel but for the atmosphere
000
Reposted by Nathan Bartley
Jerry Chen @jcsalterego.bsky.social · 16/04/2026
the Edit Button Now! Elite Digital Force has claimed responsibility for the ongoing disruption of bluesky and will continue until their demands are met
1017212