Satya Nadella

AI’s money loop is starting to look… circular. 🔄 WSJ points out a growing trend: ✅ Nvidia invests in OpenAI → OpenAI buys Nvidia chips. ✅ OpenAI signs a $300B compute deal with Oracle → Oracle buys more Nvidia GPUs. ✅ CoreWeave, AMD, Microsoft — all cross-invested and interlinked. Analysts are asking: is this ecosystem synergy or déjà vu from the dot-com bubble? “Circularity can be virtuous on the upside, and vicious on the downside.” 💬 What do you think: healthy collaboration or risky self-funding spiral? #AI #semiconductors #GPU #Datacenters Follow me on X: https://lnkd.in/grmUdMbD Image Source: WSJ

157

13 Comments

In NVIDIA’s Q2 earnings this week, Jensen Huang highlighted physical AI—alongside reasoning AI—as a key growth driver. Analysts echoed this, noting breakthroughs in robotics and autonomous systems as the next stage of AI adoption. It’s a welcome acknowledgment that AI’s impact won’t be confined to digital spaces. The real opportunity lies in the physical world—transportation, logistics, energy, construction—where safety, efficiency, and sustainability gains can be profound. That’s the frontier many of us are working on: bringing AI to the systems and workers that keep the real economy moving.

72

New work on Privacy Leakage Evaluation for Autonomous Web Agents: Paper: https://lnkd.in/e9iSkd-k LLM-powered AI agents hold great potential for enhancing productivity but risk inadvertent privacy leakage when given access to sensitive information. To address this, AI agents can be trained to better satisfy the privacy principle of data minimization, i.e. sharing private information only when strictly necessary for a specific task. This work also develops AgentDAM, a new benchmark that evaluates how well AI agents can limit unnecessary processing of potentially private information in realistic web scenarios. Our findings show that existing agents can often misuse sensitive data, but a prompting-based approach can substantially reduce this risk. With Arman Zharmagambetov, Chuan Guo, Ivan Evtimov, Maya Pavlova, and Kamalika Chaudhuri

57

1 Comment

🚀 Big week ahead for the Quantum + AI team at IEEE Quantum Week 2025! We’re thrilled to share our latest work at the intersection of Quantum Computing and Artificial Intelligence. If you're attending, here’s where you can find us: - 📅 Sunday, August 31 – 3:00PM 🧠 Paper Presentation Juan Cruz-Benito presents, "Quantum Processing Unit (QPU) Processing Time Prediction with Machine Learning" - How ML can improve QPU time prediction for quantum jobs—enabling smarter quantum workload scheduling. - 📅 Monday, September 1 – 10:00AM 🛠️ Tutorial: TUT13 — AI Methods for Quantum Circuit Optimization David Kremer and Juan Cruz-Benito will demonstrate how Reinforcement Learning can be used to transpile and synthesize quantum circuits, using qiskit-ibm-transpiler and other tools developed by our team. - 📅 Tuesday, September 2 – 12:00PM💡 Live Demo at IBM Quantum Booth #400 Juan Cruz-Benito will showcase our new AI-powered quantum computing development environment—designed to visually bridge the gap between quantum theory and practical implementation. - 📅 Thursday, September 4 – 10:00AM 🎤 Invited Talk at the 3rd Workshop on Quantum Algorithm Design Automation Nicolas Dupuis  presents: "Large Language Models for Generating Quantum Computing Code" - Learn how we’re using QPUs for post-training LLMs to power the next generation of Quantum and Qiskit Code Assistants. We’re looking forward to connecting with the community, exchanging ideas, and pushing the boundaries of what’s possible with Quantum + AI. If you're attending, come say hi! #QuantumComputing #AI #MachineLearning #Qiskit #IEEEQuantumWeek #QuantumAI #ReinforcementLearning #IBMQuantum

110

1 Comment

I was recently quoted in this IBM article exploring the rise of DeepSeek-R1—a new open-source AI model that’s quickly gaining attention on Hugging Face and beyond. What’s exciting about DeepSeek-R1 isn’t just its performance, but what it represents: the growing impact of open-source innovation and the speed at which new players are shaping the future of generative AI. It’s a great reminder that we’re still in the early days of this transformation, with room for many contributors. Appreciate the opportunity to share a few thoughts alongside such an interesting story. Big thanks to Aili McConnon for the thoughtful piece and for including my perspective. Read the article here: https://lnkd.in/ea4q8z-F You can now explore DeepSeek on #IBM #watsonx.ai, where distilled versions of the model can be deployed for secure and scalable inference. A step-by-step tutorial is available to help you get started: https://lnkd.in/ergmsGUh #AI #DeepSeek #Wastonx #IBM #OpenSource #GenerativeAI #Innovation #HuggingFace #IBMResearch #IBMAI #IBMThink #AIInfrastructure #OpenSourceAI

54

1 Comment

💡 Rethinking Knowledge-Intensive AI: The Case for Cache-Augmented Generation (CAG) Have you ever thought: "What if we could eliminate retrieval bottlenecks in AI systems?" (This question is at the heart of Cache-Augmented Generation.) To date, Retrieval-Augmented Generation (RAG) has been the gold standard for improving large language models. But it’s not without challenges: ⏳ Retrieval delays. ❌ Error-prone results. 🔧 Complex architectures. Enter Cache-Augmented Generation (CAG): A simpler, faster alternative. 💾 Preloaded knowledge. ⚡ Cached runtime parameters. 📚 Static context windows. What does this mean for knowledge tasks? 1️⃣ No retrieval delays → Answers are immediate. 2️⃣ Simplified architecture → Just the LLM and its extended context. 3️⃣ Reduced errors → Preselected, accurate knowledge at your fingertips. The results? 👉 In benchmarks like TriviaQA and WebQ, CAG performed as well or better than RAG—while cutting latency and complexity. The takeaway: For static or semi-static knowledge bases, CAG is a game-changer. ✨ "Innovative, efficient, disruptive." Well if you don't even know what is RAG you can go there : https://lnkd.in/e8jj6U59

124

5 Comments

Unlocking the Power of RAG: Check out these 8 innovative Retrieval-Augmented Generation (RAG) architectures: from classic Naive and Multimodal RAG to advanced Agentic and Adaptive approaches. Each architecture brings unique strengths for combining LLMs with information retrieval, knowledge graphs, and agentic reasoning. Whether you’re building smarter chatbots, tackling complex queries, or integrating multimodal data, understanding these RAG variations is crucial for next-level AI solutions. Which RAG architecture are you most excited to try? #AI #MachineLearning #RAG #LLM #Innovation #DataScience

38

2 Comments

Meet ZebraLogic: A Comprehensive AI Evaluation Framework for Assessing LLM Reasoning Performance on Logic Grid Puzzles Derived from Constraint Satisfaction Problems (CSPs) A research team from the University of Washington, Allen Institute for AI, and Stanford University introduced ZebraLogic, a benchmarking framework developed to rigorously test LLMs’ logical reasoning performance. ZebraLogic generates logic puzzles with quantifiable complexity, ensuring a controlled environment for systematic evaluation. The framework prevents data leakage and enables a detailed analysis of an LLM’s ability to handle increasingly complex reasoning tasks. ZebraLogic serves as a crucial step toward understanding the fundamental constraints of LLMs in structured reasoning and scaling limitations. The ZebraLogic framework constructs logic puzzles with varying difficulty levels based on two primary complexity measures: search space size and Z3 conflict count, a metric derived from an SMT solver. The study tested leading LLMs, including Meta’s Llama, OpenAI’s o1 models, and DeepSeekR1, and revealed significant accuracy declines as puzzle complexity increased. The framework allowed for a precise assessment of reasoning capabilities across different levels of problem difficulty, making it one of the most structured evaluations of LLMs to date. By systematically varying the constraints, researchers could determine the impact of problem size on logical reasoning performance..... Read the full article: https://lnkd.in/gRn4XSFC Paper: https://lnkd.in/gaziAss3 Project Page: https://lnkd.in/gkxYiwQc Bill Yuchen Lin Ronan Le Bras Kyle R. Ashish Sabharwal Radha Poovendran Peter Clark Yejin Choi

30

AI is creating intelligence on a path to AGI and ASI. This is one of the most important things happening in 2025. Jensen Huang on AI scaling: 1. Pre-training is like college 2. Fine-tuning is like a PhD 3. Compute at test-time is like deep thinking and reasoning

1,802

53 Comments