Lionel @spiindoctor.bsky.social · 10hopenai dumping those 700 papers is identical to those data labeling "help us train ai models and we'll pay you between $50-$100/hr" firms except they're not even offering any money. It's lowkey disgusting. i've done data labeling in the pre-ml age for a google search contractor: it rots your brain. 010
Lionel @spiindoctor.bsky.social · 10/10/2026I didn't like "artificial intelligence scientist" as a moniker and I like "super intelligence scientist" even less 000
Reposted by LionelJoshua Grochow @joshuagrochow.bsky.social · 10/10/2026“We’re now expected to either (1) check all this garbage that OAI is producing or (2) concede that all of these conjectures have been proven,” one algebraic topologist wrote to me. “I believe in politics this is known as ‘flooding the zone.’ We simply don’t have the resources to correct the record.” 57328
Reposted by LionelTMLR Published Papers @tmlr-pub.bsky.social · 08/10/2026New #TMLR-Paper-with-Video: The Role of Feature Interactions in Graph-based Tabular Deep Learning Elias Dubbeldam, Reza Mohammadi, Marit Schoonhoven, Ilker Birbil tmlr.infinite-conf.org/paper_pages/… #attention #neural #networks 011
Lionel @spiindoctor.bsky.social · 07/10/2026I am wondering about what is meant by the observed data likelihood, in the ising model you have two discrete states but the N particles involved makes maximization of the systems likelihood a 2^N évaluations ordeal, that's if we don't use clever algorithms. 110
Lionel @spiindoctor.bsky.social · 06/10/2026human doctors for me and doctor agent for you. I'm sure proponents will argue that "shortages" mean we must consider alternatives to human doctors in places where the ratio of doctors to people is very low. It side-steps the real cause of doctor shortages and I am not sure this is a good solution. 030
Lionel @spiindoctor.bsky.social · 05/10/2026tradeoffs are a thing but I am curious about who will be on the receiving end of the "bad things" for everybody else to "benefit" from ai. 001
Lionel @spiindoctor.bsky.social · 05/10/2026Not having those expensive to produce battles mean they have to spend more time making this writing hit and I welcome that. I am more annoyed when it's the opposite with lots of explosions and bad writing. 110
Lionel @spiindoctor.bsky.social · 05/10/2026Interesting development. Someone should nudge him into joining bsky 000
Lionel @spiindoctor.bsky.social · 05/10/2026That "of sort" is fighting for its life carrying the argument. 010
Lionel @spiindoctor.bsky.social · 03/10/2026I am guessing that usage is growing at a pace which offsets the revenue loss due to price drops, which can explain how in aggregate they are increasing revenue. 100
Reposted by LionelSam Power @spmontecarlo.bsky.social · 03/10/2026Still true in the ways that matter! 0444
Lionel @spiindoctor.bsky.social · 03/10/2026interesting that I feel like the xbox 360 was massively successful and they somehow found a way to bottle their win and nobody at that company has figured out a way to center gamers again (as opposed to corporate interests) 010
Reposted by LionelTMLR Published Papers @tmlr-pub.bsky.social · 03/10/2026New #TMLR-Paper-with-Video: Explaining with trees: interpreting CNNs using hierarchies Caroline Mazini Rodrigues, Nicolas Boutry, Laurent Najman tmlr.infinite-conf.org/paper_pages/… #cnns #interpretability #explanations 042
Reposted by Lionellastpositivist.bsky.social @lastpositivist.bsky.social · 03/10/2026A small but very bad quality of life change: google scholar used to offer the cite button for every version of a paper where there are multiple versions linked. Now (as per picture) it only does it for the top option. And sometimes the top option is the wrong one, rendering it useless. 1587
Reposted by LionelPetter Holme @pholme.bsky.social · 03/10/2026Network of the day. From: JL Moreno's Psychodrama (1946). 0102
Lionel @spiindoctor.bsky.social · 03/10/2026I hope he doesn't actually make it back, he's been on my nerves 010
Lionel @spiindoctor.bsky.social · 02/10/2026I'm giving a talk on correlation matrix filtering and downstream machine learning use cases next week, if you are curious about this topic, let me know and I will post a bunch of tweets once it's done. 020
Lionel @spiindoctor.bsky.social · 02/10/2026They're slowly expanding Houston coverage area particularly around highways. Highways are super chaotic around here, I suspect Dallas is not too different. 110
Lionel @spiindoctor.bsky.social · 02/10/2026Unsure how I feel about this cause California, even with Texas weather, would fantastic. I so wish that was the direction we were headed toward 000
Lionel @spiindoctor.bsky.social · 02/10/2026I have tons of Google scholar alerts building up a list of researchers in my field(s), I guess it's strangely useful in the current environment. 120
Reposted by LionelHacker News Top Stories @hackernewsbot.bsky.social · 01/10/2026ArXiv's Updated Rate Limit Policy | Discussionblog.arxiv.orgFair Moderation, Equitable Access, and AI: arXiv’s Updated Rate Limit PolicyarXiv, and the scientific community at large, are facing a watershed moment. Scholarly publishing is currently changing at a rapid pace, and we are seeing a… 011
Lionel @spiindoctor.bsky.social · 01/10/2026(don't close it, submit the application).static.klipy.comDoja Cat: Kermit 'Do It' MemeALT: Doja Cat: Kermit 'Do It' Meme 0110
Reposted by LionelElizabeth Hobson @elizabethhobson.bsky.social · 01/10/2026I am recruiting new parakeet postbacs to join my research team! If you are (or know of) someone who recently graduated undergrad & is looking to build research experience prior to applying to grad school or a career path (and you like animal sociality and cognition) apply here! tinyurl.com/59rf8jcytinyurl.comRecruiting parakeet research contractorsRecruiting parakeet research technicians! The Hobson Lab at the University of Cincinnati is recruiting research contractors on our Monk Parakeet sociality and cognition project at our resear… 01421
Reposted by LionelZachary Neal @zpneal.bsky.social · 30/09/2026The special issue on 𝗖𝗹𝗮𝘀𝘀𝗿𝗼𝗼𝗺 𝗔𝗰𝘁𝗶𝘃𝗶𝘁𝗶𝗲𝘀 𝗳𝗼𝗿 #𝗧𝗲𝗮𝗰𝗵𝗶𝗻𝗴 #𝗡𝗲𝘁𝘄𝗼𝗿𝗸𝘀 is now available in @insna.bsky.social 𝘾𝙤𝙣𝙣𝙚𝙘𝙩𝙞𝙤𝙣𝙨. The editorial introduction contains short descriptions of each activity, with direct links to the open-access papers providing detailed lesson plans. doi.org/10.21307/con...doi.org 142
Reposted by LionelSteven Strogatz @stevenstrogatz.com · 30/09/2026Beautiful article about math and AI by Dan Rockmore. The connection to art and photography via Walter Benjamin was very fresh -- definitely worth reading, even if you're tired of reading about math and AI. 02010
Reposted by LionelBinghamton Center of Complex Systems (CoCo) @binghamtoncoco.bsky.social · 30/09/2026Jeffrey Ventrella's CoCo talk is now on Vimeo! vimeo.com/1231801249vimeo.comParticle Systems with AttitudeBinghamton Center of Complex Systems (CoCo) Seminar September 30, 2026 Jeffrey Ventrella (Independent Algorithmic Artist, Researcher, and Author) "Particle Systems with Attitude" 162
Lionel @spiindoctor.bsky.social · 01/10/2026There is a calculator which open Ai hides but the calculator somehow makes mistakes so what do we know. 2610
Lionel @spiindoctor.bsky.social · 30/09/2026why is everybody jumping in to offer a defense of habermas... 010
Lionel @spiindoctor.bsky.social · 30/09/2026statistical physics and mechanics sequence for every college student watch the enrolment rate go poof. 030
Lionel @spiindoctor.bsky.social · 29/09/2026I am also a bit skeptical that training on all the data in the universe is better than modeling and understanding signal processing (I also like that more). 021
Lionel @spiindoctor.bsky.social · 29/09/2026Evaluating time series foundation model claims robjhyndman.com/hyndsight/fo... similar issues for me with respect to tabular foundation models (compared to production decision tree models): Performance improvement at scale, which relies on scaling laws, is not there. 121
Reposted by LionelColin @colin-fraser.net · 29/09/2026Another day working in AI Safety at a frontier lab. Everyone asks me if they can use the model without it hacking into the UN. Buddy, they won’t even let me use the model without it hacking into the UN. 638442
Reposted by LionelPhillip Isola @phillipisola.bsky.social · 28/09/2026Students sometimes ask me if it still makes sense, in this accelerating age, to pursue a PhD in AI. Perhaps counterintuitively, I think it's a great time to do so. I wrote up some thoughts on this here: web.mit.edu/phillipi/www...web.mit.eduOn the Value of Doing a PhD in the Age of AI 18116
Lionel @spiindoctor.bsky.social · 28/09/2026if you work on boring stuff (like me) they will never steal that 000
Reposted by LionelAl Merose @al.merose.com · 27/09/2026If there is one contribution that I hope to make to computer science, it’s that relational algebra is as important as linear algebra or calculus is to machine learning. 051
Lionel @spiindoctor.bsky.social · 28/09/2026They quickly got rid of the semi automated system put in place during the Biden administration to file your taxes directly with the IRS if I remember correctly. The issue is only partially technical, it's mostly political. 010