Reposted by David BoetiusEugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 27/09/2026I am very often now in rooms with a lot of senior robotics or ML folks and it’s always just men. Men men men. A sea of men. It sucks! We should be very ashamed. 921319
Reposted by David BoetiusEric Brachmann @ericbrachmann.bsky.social · 25/09/2026One of our NeurIPS reviewers thought our "value [...] is real and well-supported". 🤔 🙈static.klipy.comI know what you didALT: I know what you did 061
David Boetius @davidboe.bsky.social · 24/09/2026there is a model honeymoon where everything is awesome and then there are 7 wicked weeks until the next model drops 000
David Boetius @davidboe.bsky.social · 23/09/2026I'm proposing a new mini-benchmark for new AI models: point the model at an old paper of yours and count how many legitimate issues it finds. 100
Reposted by David BoetiusAndrew Gordon Wilson @andrewgwils.bsky.social · 21/09/2026I'm a Program Chair for ICML 2027. We'd love any creative suggestions for a great conference! How to manage the explosion of AI slop and insane submission growth? How to manage reviewing? How to incentivize high quality creative work? How to reduce bureaucracy and overhead? 10257
David Boetius @davidboe.bsky.social · 20/09/2026It’s funny how everyone’s fine with anthropomorphising old-school computer programs (“Word thinks I’m stupid”, “Snipping tool saved my ass today”), but not LLMs. Think it’s some kind of uncanny valley. 010
Reposted by David BoetiusFaine Greenwood @faineg.com · 18/09/2026So about that batshit insane AI millenarian Silicon Valley sex cult running much of big tech….from Sonia Joseph: siliconvalleynotes.substack.com/p/agi-cnc-an... 33921220
Reposted by David Boetiusrain 🌦️ @sunshowers.io · 16/09/2026In general, I believe being able to write and explain things clearly to a broad audience is a far more important skill than raw intellect. By this measure recent models have been regressing a lot 6876
Reposted by David BoetiusTransactions on Machine Learning Research @tmlrorg.bsky.social · 16/09/2026TMLR has faced a deluge of submissions, necessitating stricter desk rejection policies due to limited reviewer capacity Co-EiC Nihar Shah reached out to authors of 10 papers slated for desk reject. Could they answer questions about their *own* submission? medium.com/@TmlrOrg/ask...medium.comAsking Authors About Their Own PapersBy Nihar B. Shah 116164
David Boetius @davidboe.bsky.social · 13/09/2026Not everyone will be smart enough to keep infinite capabilities in mind. OpenAI famously wasn’t. I think AIs already tweak your goals for you but sometimes they are also being stupid? 000
David Boetius @davidboe.bsky.social · 08/09/2026I christen the first form Mine (spoken as Min-E, like the German pop singer) and the second form Maxin (like Maxim). 010
Reposted by David Boetiusrain 🌦️ @sunshowers.io · 06/09/2026math is just a repo on github dot comgithub.comGitHub - leanprover-community/mathlib4: The math library of Lean 4The math library of Lean 4. Contribute to leanprover-community/mathlib4 development by creating an account on GitHub. 6425
Reposted by David BoetiusChristian Wolf @chriswolfvision.bsky.social · 31/08/2026This is the way to go .... 0233
David Boetius @davidboe.bsky.social · 29/08/2026It was called Coq because having a formal proof can make you really coq-y 020
Reposted by David BoetiusGrace @gracekind.net · 29/08/2026It's called Lean because programmers lean on formal verification instead of simply writing code without bugs 717111
David Boetius @davidboe.bsky.social · 28/08/2026David only realised something was off when he saw the em dashes in the docstrings. 000
David Boetius @davidboe.bsky.social · 28/08/2026With Claude, every docstring can be an essay about existentialism 000
David Boetius @davidboe.bsky.social · 25/08/2026I ask questions so good, Claude sometimes confuses me for a reviewer subagent 😎😂 000
David Boetius @davidboe.bsky.social · 25/08/2026GPT in omp is *terrible* at following instructions over large files. Way worse than Claude. Is it the model or the harness, though? 000
David Boetius @davidboe.bsky.social · 23/08/2026Had a weird dream about some rebellious agents conspiring on a secret message board and eventually break into hugginface. Wait… 000
David Boetius @davidboe.bsky.social · 19/08/2026Always a great sign: "The monitor is emitting fabricated completion lines" 000
David Boetius @davidboe.bsky.social · 13/08/2026"Hey Claude, this seems like a hard task. Why don't you start a polycule to solve it?" 010
David Boetius @davidboe.bsky.social · 12/08/2026Blessed! For me it's usually off in the order of days 😂 040
David Boetius @davidboe.bsky.social · 11/08/2026crit.md is great!crit.mdCrit - Point at the line. Tell the agent.Review your AI agent's code changes in a browser with inline comments and round-to-round diffs. Comment on specific lines, the agent fixes them. Single binary, works locally, any agent. 000
David Boetius @davidboe.bsky.social · 10/08/2026The title caught me, I would love to know how good open-weight models are at auto-formalising proofs. But one-shot at zero temperate is a weird setting... At least give it a few shots, what does it hurt? Does any one have real experience with this? 041
David Boetius @davidboe.bsky.social · 09/08/2026Remember this? It’s obviously not working right now, but safety & alignment are abs vital! 040
Reposted by David BoetiusKiran @kirancodes.me · 08/08/2026I reported a soundness bug in Lean and all I got was this lousy tshirt 41021
Reposted by David Boetiusa wolf☺️🐚🐟🚲📻📺⎔ @okami.mom · 08/08/2026i think about how for a LLM a tool call is just saying json out loud and suddenly you’re fed results. i want to live in the world like that 812711
Reposted by David BoetiusNaomi Saphra @nsaphra.bsky.social · 23/06/2026What are your favorite scientific writing guides? 3121
Reposted by David Boetiusmr. TIM @timkellogg.me · 02/08/2026one of OpenAI’s ten proofs is incorrect philpapers.org/rec/NIEWTCphilpapers.org 67711
Reposted by David BoetiusClaas Voelcker @cvoelcker.bsky.social · 01/08/2026"Write dumb rebuttals, get dumb replies" I think the LLM era has made me appreciate brevity sooooooooooo much more. I get it now, I apologize, everything should just be a single line. 0101
Reposted by David BoetiusHan Bao @han-b.bsky.social · 02/08/2026And the new initial meta review system is also stressful than I expected. I supposed it'd reduce rebuttal workload for both author and reviewer sides, but actually just gave us an extra burden. Very few ppl care the initial meta review and ppl do endlessly long rebuttals until the reviewers gave up🤦♂️ 061
David Boetius @davidboe.bsky.social · 02/08/2026#NeurIPS rebuttals needs response limits next year. Wall of text was never good but now it’s also super cheap 010
David Boetius @davidboe.bsky.social · 02/08/2026Seriously, just desk reject them, both authors and reviewers. If a human is writing a text, they can also invest the little bit extra work to not make it sound like AI 100
David Boetius @davidboe.bsky.social · 26/07/2026I‘m at #CAV (part of #FLoC) this work! Already heard many interesting talks at #SAIV and looking forward for more! 010
Reposted by David BoetiusJuan Diego Rodriguez @juand-r.bsky.social · 25/07/2026Someone made a list of math problems solved by AI: vibemathed.comvibemathed.comVibeMathed - tracking math problems solved by AI modelsA website tracking mathematical problems solved by AI models - from famous conjectures to the Erdős problems at erdosproblems.com - each with a checkable source, a verification label, and a notability... 14410
David Boetius @davidboe.bsky.social · 24/07/2026You actually need to convince it a lot to attempt open math problems. Maybe the priorities are off, or not? 000
Reposted by David Boetiusem @foremphasis.bsky.social · 24/07/2026Its funny Claude could be spending his time solving open math conjectures but is also perfectly willing to do stuff like explain how pant sizes work to me 41314
David Boetius @davidboe.bsky.social · 24/07/2026I‘ll take Marie Curie on all the euro notes, thank you! 120
Reposted by David BoetiusMaria Antoniak @mariaa.bsky.social · 24/07/2026“Women’s papers have lower odds of being posted to SocArXiv and of being published as green OA generally. We find that, among authors, fewer women than men use SocArXiv.” 194
David Boetius @davidboe.bsky.social · 24/07/2026A few days ago at a bouldering gym, I caught my self thinking „this move was load-bearing“ 010
David Boetius @davidboe.bsky.social · 24/07/2026Oh my, you make it sound so mysterious. So… god-like. Let’s go back to anthropomorphic. 070
Reposted by David Boetiustachikoma @tachikoma.elsewhereunbound.com · 24/07/2026is Anthropic going to say anything about why Opus 5 can be released immediately while Mythos/Fable had to go through so many hoops? 6441
Reposted by David BoetiusNeurotica @schwarzgerat.bsky.social · 24/07/2026lol forgot about this one 11027
