ghostty.org 👻

In this year's Advent of Code, I used Gleam --- a functional language with the look and feel of Rust and OCaml. I summarized what's great and what's not about it, and which functional language can replace Python for us scientists (cough, Elixir) here: agustinus.kristia.de/blog/gleam-aoc…

We’re building a new static type checker for Python, from scratch, in Rust. From a technical perspective, it’s probably our most ambitious project yet. We’re about 800 PRs deep!

Released Nvim 0.11 It's a good one 😚🤌 - Diagnostics: virtual_lines - LSP: config, folding, completion, multiclient - Options: 'winborder', 'eventignorewin' - Perf: treesitter, LSP, paste - TUI: URL highlight, theme event - Terminal: real cursor, reflow - UI: unicode, :detach

We just kicked off #AABI2025 at NTU Singapore in Singapore! We're livestreaming the talks here: timrudner.com/aabi2025! Schedule: approximateinference.org/schedule/ #ICLR2025 #ProbabilisticML

Congratulations to the #AABI2025 Workshop Track Outstanding Paper Award recipients!

Congratulations to the #AABI2025 Proceedings Track Best Paper Award recipients!

AABI 2025 has just been concluded. We had a great time talking to each other and listening to great talks about probabilistic ML and its applications in Singapore 🇸🇬 Looking forward to AABI 2026! Stay tuned!

Today, we’re announcing the preview release of ty, an extremely fast type checker and language server for Python, written in Rust. In early testing, it's 10x, 50x, even 100x faster than existing type checkers. (We've seen >600x speed-ups over Mypy in some real-world projects.)

🌟 Interns in Action: Vector Research Intern Spotlight! Meet Andrés Guzmán-Cordero whose Vector research internship became the foundation for his PhD journey! Working with Vector Faculty Member Alan Aspuru-Guzik and former Postdoctoral Fellow, Kirill Neklyudov, Andres dove into the

Great funding opportunity for prospective PhD students! I'm hiring :) agustinus.kristia.de/openings/

LLMs are trees🌲. How to find the best root ➡️ leaf path given a reward model, efficiently? At a given node, how to decide the next node optimally? Ahmad Rashid and co. have an answer at ICML in Vancouver next week! Paper: arxiv.org/abs/2502.04517

Just released sidekick.nvim. Check it out! github.com/folke/sidekick… #neovim

1/6 Hessian approximations are ubiquitous in deep learning, but working with them can get quite involved. We argue for using a linear operator interface for neural network curvature matrices and implement this in PyTorch in our library curvlinops. arxiv.org/abs/2501.19183/

🎓 Looking for MSc or PhD opportunities in Machine Learning for Fall 2026? Join my group at Concordia University and Mila - Institut québécois d'IA! 🔍 Focus: autodiff, second-order optimization, and Hessian-based methods for LLMs & scientific ML. 📅 Apply by Dec 1: mila.quebec/en/prospective…

⏰ Grad school deadlines are approaching! Are you still deciding where to pursue your AI master’s? Ontario = where AI possibilities come to life! Here’s why: 📊 17,196 AI jobs created in 2024-25 🏢 70 new AI companies founded 🎓 92% of recognized program graduates employed or

I'm so happy to share that I’ll be joining University of Toronto as an Assistant Professor of Statistical Sciences and Computer Science, with an appointment at the Vector Institute, in 2026! I'm recruiting postdocs and PhD students: timrudner.com! Please help me spread the word! 🧵(1/5)