Bram Moolenaar, creator of #Vim, just passed away today. RIP Bram

In the last two years, Xinyu has been working with Darko Marinov and other PL/FM/SE at Illinois faculty to integrate the research on Configuration Testing (Ctest) into splash grad-level courses, including CS 527 (Software Engineering) and CS 591SE. Ctest is now a course project in 523.

With Hao Wang and Chenyi Zhang, we study quantum algorithms for minimizing the maximum of convex functions, and also prove quantum lower bounds showing near-optimality. The paper was accepted by ICLR 2024 with arXiv link arxiv.org/abs/2402.12745 See you in the conference!

🤝 New work on LLM for code optimization with @allen_a_nie, Rohan Yadav, and amazing teams at Stanford, Intel, NVIDIA, and Visa Research. We show how LLMs can generate system code that outperforms human experts in parallel matrix multiplication. 🧠🔧 🔗 arxiv.org/abs/2410.15625

The next frontier in AI: autonomous app creation! BaxBench evaluates LLMs on 392 app backend creation tasks, highlighting the gap between correctness and security. SRI Lab LogicStar AI baxbench.com

Humanity's Last Exam's stated goal is PhD-level questions that can not be easily looked up and undergrads shouldn't even understand. Literally one of the examples on their homepage is solvable by the respective wikipedia entry. Did anyone check the % of this?

[1/8] 🚀 New preprint: Climbing the Ladder of Reasoning: What LLMs Can—and Still Can’t—Solve after SFT? We provide a systematic study on the potential and limitations of Supervised Fine-Tuning (SFT) on math reasoning tasks.

Excited to present my upcoming PLDI paper at the ICML Workshops DL4C and VerifAI! Type systems are useful at preventing bugs - so why not leverage them for LLMs? Using constrained decoding, we reduce reduce compiler errors of TypeScript code by over 50%! More details in the 🧵

1/🧵Introducing VERINA: a high-quality benchmark for verifiable code generation. As LLMs are increasingly used to generate software, we need more than just working code--We need formal guarantees of correctness. VERINA offers a rigorous and modular framework for evaluating LLMs

1/ 🔥 AI agents are reaching a breakthrough moment in cybersecurity. In our latest work: 🔓 CyberGym: AI agents discovered 15 zero-days in major open-source projects 💰 BountyBench: AI agents solved real-world bug bounty tasks worth tens of thousands of dollars 🤖

Join us at Agentic AI Summit 2025 — August 2 at UC Berkeley, with ~2,000 in-person attendees and the leading minds in AI. Building on the momentum of the 25K+ LLM Agents MOOC community, this is the largest and most cutting-edge event on #AgenticAI. As 2025 emerges as the Year of

1/N 🚀 Launching LEANN — the tiniest vector index on Earth! Fast, accurate, and 100% private RAG on your MacBook. 0% internet. 97% smaller. Semantic search on everything. Your personal Jarvis, ready to dive into your emails, chats, and more. 🔗 Code: github.com/yichuan-w/LEANN 📄

🚀Ever wondered how to make RL work on impossible hard tasks where pass@k = 0%? 🤔 In our new work, we share the RL Grokking Recipe: a training recipe that enables LLMs to solve previously unsolvable coding problems! I will be at #CoLM2025 next week so happy to chat about it!

🎓 Amazon launches AI PhD Fellowship program, providing $68 million over two years to fund PhD students at 9 universities pursuing research in machine learning, computer vision, and natural-language processing. #AmazonAIFellowship amzn.to/4o6uPlu

Amazing! 10 @BerkeleyEECS @SkyCompLab grad students are Amazon AI PhD Fellows! Congrats! Learn more about our fellows here: bit.ly/48Hl88q #AmazonAIFellowship UC Berkeley Sky