Anand Shah @apscomp

Thinking of starting a community for people who want to learn hardware, work in hardware, or just enjoy tinkering with things in general. Reply to the tweet if you want in.

This is amazing. Paper → interactive website demo is exactly what I would have loved to have for some of the more complex long NLP papers.

Jeff Dean @JeffDean

2 weeks ago

2/ Check out how Gemini 3.5 Flash instantly digests dense academic papers and autonomously codes a fully interactive, visual website explaining the intricacies of the research. It's an incredible stress test that seamlessly merges massive long context, deep reasoning, complex

6 27 273 88K 167

When it comes to Mathematics, He is the BEST. Absolute beginner friendly.

Le #CERN rend open-source sa librairie de composants électroniques pour le logiciel KiCAD. Disponible via Gitlab, elle contient des données pour plus de 17000 composants électroniques, y compris des symboles schématiques et des empreintes de circuits. gitlab.com/ohwr/cern-kica…

Today is the day. The Linux Field Guide is officially live. This is the project I've been talking about for a while. A long-form publication for "upper beginner" Linux users - the people who installed Linux, are comfortable in a terminal, and want to understand WHY things work the way they do, not just how to type the commands. The first article is now up. It opens Series 01: The C Layer. lfg.popovicu.com/series/the-c-l… Title: Why C is the Linux userspace interface. Most writing about C defends it the same way - "it's fast," "it's close to the metal," "there's too much legacy code to replace it." All of these treat C as a tool you happen to be stuck with. This article makes a different argument: C isn't a language you pick on Linux. It is literally the operating system interface, as POSIX defines it. Working code throughout. Real assembly for x86_64 and RISC-V, the actual ld command line gcc hides from you, and Apple's own documentation as receipts. About a 15 minute read. This is article 1 of many. Six series planned, each ~10 entries. The C Layer is just the start - shells, /proc, signals, files-as-everything, and bootstrapping a Linux system from scratch are all coming. A newsletter is in the works for readers who want article roundups plus extra content. For now, follow here for updates. Thank you to everyone who followed along through the campaign this past week. Today is the payoff.

x.com/i/article/2040…

Man explains how Cos Θ learnt in school is used in real world

Would you like to learn more about SQL or are you more interested in DuckDB's internals? Either way, you're in luck: Professor Torsten Grust of the University of Tübingen published two sets of lectures (almost 300 slides in total) along with their supplementary material.

Modern DRAM is based on a brilliant design from IBM. But, we're still paying for a latency penalty that's existed since the 60s! In this video, I'm introducing my research project (Tailslayer) that immensely reduces p99.99 latency on traditional RAM! By implementing a hedged read strategy taking advantage of (undocumented!) channel scrambling offsets, I've gotten as much as 15x reductions in tail latency. The technique works across Intel, AMD, Graviton, DDR4, DDR5, x86, ARM, you name it. Check out the C++ lib I wrote, watch the video, and try it yourself!

I built everything from scratch, 100% inhouse in india, it was just me & my obsession of making technology accessible to everyone in need! It is a brain controlled robotic prosthetic hand! Learn more at: brhm.in

An absolute banger of a paper. “A Gentle Introduction to Matrix Calculus” by econometrics legend Jan Magnus — one of the clearest explanations of matrix derivatives ever written. If you work in econometrics, machine learning, statistics, or optimisation, this paper is pure gold.

see how 4 university students reverse engineered Google's most advanced AI chip tomorrow

Open source Systolic Array (TPU) at home arxiv.org/abs/2411.09543

If you work in Physics, Machine Learning, Engineering, or any field with a serious mathematical component, Space Mapping is one of those ideas worth adding to your toolkit.

Tom and I have finally finished a draft of Dynamic Programming Vol 2! Exhausting but satisfying. New approach to DP theory, advanced material, many applications... dp.quantecon.org

CS 6120: Advanced Compilers: The Self-Guided Online Course cs.cornell.edu/courses/cs6120…

From learning math as a child to coding on my first VTECH computer (gotta start with the BASIC!), applied math and software have always been my passions. I’ve been lucky to get to build a career where with these fields now drive modern AI. PyTorch has been a big part of making that possible, bridging theory and implementation, and making it available to all as open-source software. Today, I’m proud to announce that we are contributing back to that ecosystem: The NeuralOperators library is officially joining the PyTorch ecosystem! NeuralOperators brings together learning on function spaces and deep learning to enable AI for science and engineering. It’s now available for you to use directly in your existing PyTorch framework. Thank you to the @PyTorch community and our amazing team: @AnimaAnandkumar, Valentin Duruisseaux, and @davehpitt neuraloperator.github.io/dev/index.html #PyTorch #NeuralOperators #AI #OpenSource #DeepLearning #AppliedMathematics #AIforEngineering #AIforScience

PyTorch @PyTorch

6 months ago

NeuralOperator is now part of the #PyTorch Ecosystem, bringing a PyTorch-native library for learning neural operators and modeling mappings between function spaces for AI-driven science and engineering. 🔗 pytorch.org/blog/neuralope… @JeanKossaifi @AnimaAnandkumar @davehpitt

4 78 598 42K 208

Super excited to release TorchLean!! I’m happy to answer questions and would love to discuss verified NNs + theorem proving especially what it’ll take for the field to become widely usable in real ML systems. Blog post + codebase release soon!

Prof. Anima Anandkumar @AnimaAnandkumar

3 months ago

We’re excited to release TorchLean which is the first fully verified neural network framework in Lean. The Lean community has largely focused on pure mathematics. TorchLean expands this frontier toward verified neural network software and scientific computing. With the recent

26 246 2K 148K 944

ZK proofs are headed for real-world deployment — but writing correct ZK statements is tricky. Our new post introduces zkLean, a Lean library that defines a domain specific language for specifying and formally verifying ZK statements. Read more: galois.com/articles/zklea…

One of the best talks on perf - "Performance Matters" by Emery Berger!