Join free with your email
Free. No spam, ever — we'll never share your email address and you can opt out at any time. Already a subscriber? Log in
Researcher with public evidence across AI research, NLP & language.
A general purpose (nothing special) internal OpenAI model solved one of the most famous previously unsolved problems in discrete geometry The solution involved far too many decisions for a human to feasibly explore openai.com/index/model-...
“It's literally the gulag,” one of the employees claims. “You have zero purpose in life all of a sudden, you barely interact with anyone, you just have these tasks every week." www.wired.com/story/mark-z...
wow, the Huggingface guys made an annotated version of Magnifica Humanitas huggingface.co/spaces/socie...
2) Characterizing Narrative Content in Web-scale LLM Pretraining Data, by @teagrjohnson.bsky.social, @elliottash.bsky.social, @andrewpiper.bsky.social, @mariaa.bsky.social arxiv.org/abs/2606.19468 Why: annotation and analysis of narrative features across the pretraining data (…
jacobin.com/2026/07/ai-n... “AI is a machine that takes good jobs and makes them precarious. It’s a machine for turning stable employment into gig work”
An incomplete list: 1) Explaining Attention with Program Synthesis, by Amiri Hayes, @belindazli.bsky.social and arxiv.org/abs/2606.19317 Why: LMs can generate code that can explain/replace attention heads! 🤯 (Program synthesis 🤝 Interpretability)
Does physical modeling matter or should we just do data-driven stuff? I feel like this paper characterizes things perfectly: arxiv.org/abs/2410.23179. Physics-based modeling is always helpful but the relative gain closes at scale
4) Why Larger Models Learn More: Effects of Capacity, Interference, and Rare-Task Retention arxiv.org/abs/2605.29548
Does anyone know why MASS (arxiv.org/pdf/1905.02450) didn't take off, while BART did? (OK maybe no one remembers BART today, but it was popular around 2019-2022)
recommended reading von @mitsuhiko.at man kann den KI aspekt dabei vollends ignorieren, btw. lucumr.pocoo.org/2026/6/13/am...
I think mathematics at this level is a human endeavor, and solving these problems is an exercise in creativity like working on a puzzle. We’re just losing our humanity by letting AI do all the fun stuff while we toil at our desks doing tedious stuff like filling out expense fo…
aclanthology.org/2025.finding...
Tip on reviewing ARR papers in 2026: - Check the references FIRST. I wasted 40 minutes reviewing a paper before realizing the references were LLM-generated.
I'm so tired of seeing LLM-generated reviews. They just wasting all our time. If this continues, "peer review" will mean nothing
It was a pleasure to moderate the panel at the ACL 2026 SRW: "The next big questions in NLP: What should students work on?" Huge thanks to our panelists, @boydgraber.bsky.social, @igurevych.bsky.social, Preslav Nakov and Yulia Tsvetkov for their thoughtful answers and helpful advice! Takeaways 👇
I asked GPT 5.6 Sol to solve the Collatz conjecture. Why is it searching the NOAA fisheries website??
Google AI: "oh I thought you meant Neuro-Linguistic Programming" I really thought we were over this! Apparently not
"Specification and Removal of Demons." Not what I expected to read in an NLP book! (Charniak and Wilks, Computational Semantics)
Claude is confidently telling me that Fable is a non-reasoning model “because it’s an Anthropic model” lol
What is the best open weights model+harness for mathematics? How far ahead are GPT 5.6 and Opus models from open models here?
Miyu Oba on “CxMP: A Linguistic Minimal-Pair Benchmark for Evaluating Constructional Understanding in Language Models”
Google AI: "I heavily hallucinated in that last response, completely made up details about the logo design and social events, and I am very sorry for spreading misinformation." I miss the old Google search
Center = Juan Diego Rodriguez. Left = members they follow (green edges). Right = members who follow them (blue edges). Top = mutual follows (orange edges, slightly larger). Drag any node to reposition; click to open that profile.