100 years old and still the coolest person alive. Happy birthday, Sir David! https://t.co/dgxQb6LPZ6

Python, MCP, A2A, and a whole lot about how to build multi-agent systems. This goes well beyond building a single agent: it focuses on how multiple agents can collaborate towards a goal. Best part: The book will show you how to build everything from scratch. You start with a simple agent and then add complexity as the book progresses. Amazon link: https://t.co/GIIiIOKxgw

@Santandave1 “He’s known for changing a key like John Coltrane” https://t.co/6kjL98wfvE

@animeposuto https://t.co/j4RPTYOZ7I

@animeposuto https://t.co/j4RPTYOZ7I

🚨🚨| OFFICIAL: Morgan Gibbs-White has been named the Premier League Player of the Month for April! 🏴󠁧󠁢󠁥󠁮󠁧󠁿🎖️ https://t.co/VhSfCYkuBQ

Words to live by today—and every day. https://t.co/qFRIiQcCGH

Words to live by today—and every day. https://t.co/qFRIiQcCGH

NVIDIA is at SCSP AI EXPO in Washington D.C. Giving a talk with @Baxate about importance of American open source, from internet to Linux to AI, with a demo of OpenClaw paying my parking tickets 3:25pm ET @ Convention center. At Expo hall demo stage https://t.co/QYpIWKrNrF

Another packed night at @PubKey 🍻🤖 Loved hosting @WashingtonAINet with Tammy Haddad moderating a great conversation with @NaderLikeLadder and @Baxate on the future of AI, innovation, and all the weird exciting stuff ahead. Smart people, strong drinks, big ideas — exactly the kind of nights we love.

+1 more

Super busy week in DC talking to policy makers and think tanks on AI. The space is moving fast for technologists, even faster for people who recently had to learn what “open source” is for the first time https://t.co/3K3BnZLeIE

+2 more

Fun time on the hill explaining why Open Source is deeply strategic for America https://t.co/ilPDzCXe1u

+2 more

Fun time on the hill explaining why Open Source is deeply strategic for America https://t.co/ilPDzCXe1u

+2 more

Introducing the Printing Press, a CLI-factory and a CLI-library. Built with @trevin. 🏭🖨📚 Most APIs suck for agents. Most MCPs suck for agents. Most official CLIs suck for agents. They waste tokens and time. @steipete started making his own because of this. 📚 A Library of agent-native CLIs you install today (Linear, ESPN, Flight GOAT (Google Flights + Kayak nonstop), Contact Goat (LinkedIn + Happenstance + Deepline more) +30+ more) 🏭 A factory that prints new ones for any service - just type /printing-press <product name> CLIs are fast, local, SQLite-backed. Work in Claude Code, Codex, OpenClaw, Hermes. 🌐 https://t.co/GjnN9E9yTH

"I think the highest turnover job of any Silicon Valley executive position might be the CMO." Airbnb's @bchesky says he has a huge amount of respect for people in marketing because it's one of the hardest functions in a business. "Because once something works, it almost becomes stale. Because everyone does it." "Part of my theory is that what works in marketing changes every few years and you have to be adaptable. Your old playbook gets outdated." "It's this thing called 'banner blindness.' After you see something over and over, you tend to be blind to it, so you need a new tactic."

I wake up every morning to 3 emails summarizing 1) emails from last week, 2) emails from the weekend, and 3) emails from this morning It's so wild that we used to spend manual time curating emails to bring ourselves up to speed. https://t.co/LJOJO4TATk

NVIDIA’s Nader Khalil and Carter Abdallah joined Tammy Haddad and the @WashingtonAINet to discuss how open models, agents and tools like NemoClaw can help build safer, more transparent AI ecosystems. Open source AI is not just a developer story. It’s a U.S. leadership story. https://t.co/MIiz9hbZVc

The clock is ticking to join the program for #KubeCon + #CloudNativeCon + #OpenInfraSummit + #PyTorchCon China. Share your knowledge with the global ecosystem & submit to speak by 3 May: https://t.co/U0HIhZMoFE And don't forget early bird rates end in 1 week! Save now: https://t.co/bTRuC5Azau

Got a #PyTorch breakthrough to share? 🎤 The #CallForProposals for #PyTorchCon North America (Oct 20-21 | San Jose) is in full swing. Submit your technical session idea by June 7! Apply: https://t.co/hLlKK7WxLD https://t.co/3RbHVVvQvY

LightSeek Foundation recently released Shepherd Model Gateway (SMG). It came out of a production bottleneck in LLM serving: CPU-bound work sitting on the critical path. Light Seek moved all non-GPU work into a Rust-based gateway, with a minimal gRPC boundary around inference. The project is developed and maintained by @lightseekorg. The result: up to 3.5× throughput in long-context scenarios. 🔗 More details in our latest blog: https://t.co/S4OzoakVBr #PyTorch #LightSeek #OpenSourceAI #ShepherdModelGateway

🎉 Day-0 support for @deepseek_ai V4 Pro and Flash on vLLM — a new generation of DeepSeek model, purpose-built for tasks up to 1M tokens. Alongside the release, we're publishing a first-principles walkthrough of the new long-context attention and how we implemented it in vLLM. The new attention mechanism, in four moves: • Shared K/V + inverse RoPE → 2× memory savings • c4a / c128a KV compression → 4×–128× savings • DeepSeek Sparse Attention over compressed tokens • Short sliding window for locality across compression boundaries At 1M context, per-layer KV state is ~8.7× smaller than a DeepSeek V3.2-style 61-layer stack (9.62 GiB vs 83.9 GiB, bf16). fp8 attention cache + fp4 indexer cache shrink it further. vLLM side: • Unified hybrid KV cache — single logical block size (256 native positions) across all compression rates; compressor state folded into the SWA KV cache spec so prefix caching, disagg prefill, CUDA graphs and MTP reuse the same abstraction • Three page-size buckets for the full 5-way cache stack → no cross-kind fragmentation • Fused kernels: compressor + RMSNorm + RoPE + cache insert (1.4–3×), inverse RoPE + fp8 quant (2–3×), Q-norm + KV RoPE + K insert (10–20×) • Multi-stream overlap of indexer vs main-KV compression vs SWA insertion Disaggregated serving is supported out of the box and strongly recommended for best performance. Follow our recipes site for verified commands for @nvidia Blackwell (B200, B300, GB200, GB300) and Hopper (H100/H200/H20) systems. Thanks to the @deepseek_ai team for open-sourcing DeepSeek V4, and to @inferact for landing day-0 support 🤝 📝 Blog: https://t.co/Eh7vk6xVJy 📖 Recipes: https://t.co/jlWuzYyZeX 🤗 https://t.co/IA9qAysqJk

+2 more

FINAL CALL! 🚨 The #CFP for #KubeCon + #CloudNativeCon + #OpenInfraSummit + #PyTorchCon China, 8-9 Sept in Shanghai, closes TODAY at 23:59 China Standard Time. Don't miss your chance to speak on the premier #OpenSource stage. Submit NOW: https://t.co/PKdf9djzqR https://t.co/ELdAu3V4KR

Registration deadline alert! 🚨 Tomorrow is the last day to save ¥1,270 (USD$179) on your pass for #KubeCon + #CloudNativeCon + #OpenInfraSummit + #PyTorchCon China in Shanghai (8-9 Sept). Secure early bird rates by 6 May at 23:59 CST: https://t.co/9x7dLq0gVE https://t.co/ttriVff1dg

Relive the energy of 2025 as we prepare for San Jose! ⚡ #PyTorchCon North America returns to the Bay Area Oct 20-21. Check out last year’s highlights &amp; lock in your $400 savings today. 📺 Recap 2025: https://t.co/yLtZy5VrD9 🎟️ Register for 2026: https://t.co/AVHdaIFT20 https://t.co/uwx47Q9EXK

We’re excited to share In-Kernel Broadcast Optimization (IKBO)—a kernel-model-system co-design approach driving @Meta's large-scale recommendation system (RecSys) inference. It serves as the scalability backbone for the request-centric, inference-efficient framework underlying the Meta Adaptive Ranking Model (serving LLM-scale models in production). The core idea is simple but powerful: broadcast is a data layout concern, not a computational necessity. Instead of materializing broadcast at the service layer, IKBO pushes it all the way down into the kernel layer—precisely where request and candidate features interact. This enables the system, together with model co-design, to capture the full performance upside. By preserving request and candidate inputs in their native batch sizes and fusing broadcast directly into the kernel, IKBO eliminates replicated memory movement and redundant computation across the stack. From Attention in sequence modeling to GEMM in traditional workloads, the same idea delivers broad inference gains. Results from our co-designed deployments: ⚡ Up to a 2/3 reduction in compute-intensive net latency on co-designed models. ⚡ End-to-end deployment across Meta's RecSys stacks on both GPU and MTIA (Meta's in-house accelerator). ⚡ Up to 4x speedups for linear compression GEMM and up to 2.4x/6.4x (kernel only / kernel + broadcasting) throughput for IKBO attention compared to FlashAttention-4 on H100. 🔗 If you’re interested in how rethinking a primitive like broadcast can unlock system-level gains for modern AI inference, take a look: https://t.co/mdl0F1Zj7E By @jiao_jian62798, @liptds_11, Hongtao Yu, Yuanwei Fang, Zhengkai Zhang, Zoran Zhao, Yuxuan Chen, Sijia Chen, Yang Chen #PyTorch #OpenSourceAI #RecSys #AIInfrastructure

The second in the series of Simplifying Sparse Deep Learning with Universal Sparse Tensor in nvmath-python In a previous post, NVIDIA introduced the Universal Sparse Tensor (UST), enabling developers to decouple a tensor’s sparsity from its memory layout for greater flexibility and performance. We’re excited to announce the integration of the UST into nvmath-python v0.9.0 to accelerate sparse scientific and deep learning applications. Read the follow-on post: https://t.co/sARnNMbz1r #PyTorch #OpenSourceAI #AI #Inference #Innovation

🚀 Day-0 MTP support for Gemma4 now available at vLLM with ready-to-use docker image! ⚡️Enjoy up to 3x faster decoding performance to supercharge your development with zero quality degradation! Check out the full vLLM recipes for Gemma 4 model series👇 https://t.co/IrCaaa6SIo https://t.co/eFcAZRogLF

FINAL CALL for savings! ⏳ Today is the absolute last day for early bird rates for #KubeCon + #CloudNativeCon + #OpenInfraSummit + #PyTorchCon China (8-9 Sept). Join us in Shanghai and save ¥1,270 RMB if you register by 23:59 CST tonight: https://t.co/9x7dLq0gVE https://t.co/OmOVEzgTxR

Introducing TokenSpeed, a speed-of-light LLM inference engine. &gt; TensorRT LLM level performance &gt; vLLM level usability &gt; Built by a lean and mission-driven team in two months &gt; MIT license, open-source https://t.co/MJzhCEg7m8 https://t.co/anhoETwwS9 https://t.co/BWn4Me62x7

+1 more

🚀 Excited to be the exclusive day-0 launch partner for @lightseekorg's Tokenspeed project! We've integrated Tokenspeed's MLA library, optimized specifically for agentic workloads with long context and multi-turn, purpose-built for Kimi 2.5/2.6 and DeepSeek R1 on NVIDIA Blackwell hardware! Try it out today with our preview image - nightly support coming soon!

From #Kernel Engineering to Responsible #AI: We want your voice at #PyTorchCon North America in San Jose (Oct 20-21). 🔥 Submit your talk by June 7 &amp; join the world's leading AI innovators. #CallForProposal submissions: https://t.co/hLlKK7WxLD https://t.co/fLsCa4jaJX

AI-Powered Workflows to Build Fault-Tolerant Quantum Systems The fundamental challenge in quantum computing—qubits are inherently noisy. The best quantum processors make an error roughly once in every thousand operations. To become useful accelerators for scientific and enterprise problems, error rates must drop to one in a trillion or better. AI is the most promising path to closing that gap at scale. NVIDIA Ising is the world’s first family of open AI models for building quantum processors, launching with two model domains: Ising Calibration and Ising Decoding. 🔗 Read the full post: https://t.co/0m20o5x8FU #PyTorch #OpenSourceAI #AI #Inference #Innovation