Steve Milazzo | ThopertyBox™

@Azure Strong governance is becoming essential. The next challenge may be understanding the origin and integrity of information before agents act on it.

Announced at #MicrosoftBuild, a new partnership with Fireworks AI brings open weight models into Foundry. Developers can build next generation agent applications with more flexibility, strong performance, and enterprise grade governance to scale.

@WonderLaura Governance isn't really about control. It's about visibility. If organisations can't explain who had access to information and when, trust becomes difficult to maintain.

I’ve been talking to customers a lot about governance in Microsoft 365 lately… and honestly, that’s where things tend to get messy 😬 You might have everything “working”… …but not really know who has access to what, or what’s being overshared 🤔 So I’ve been looking at ShareGate’s newer governance product, and I’m really impressed 👏 What I like: You can actually see what’s going on 👀 You can spot oversharing 🔍 And you can fix things without scripts or heavy IT work Governance isn’t something you set once — it drifts over time 📉 This just makes it a lot easier to stay on top of it ✅ If governance (or Copilot readiness) is on your radar, it’s worth a look. What’s your environment feeling like right now — pretty clean, or a little chaotic?

@dkare1009 Impressive stack. One question becomes increasingly important: what sits beneath it? As AI scales, understanding the origin, integrity and trustworthiness of inputs may become as important as the intelligence itself.

Microsoft is quietly taking over the Enterprise AI Agent stack. And most people have only seen ~10% of it. Everyone talks about Copilot. Some know Azure. A few use GitHub Copilot. But underneath... Microsoft has built a full-stack AI ecosystem —from models → to agents → to governance. Here’s the full breakdown 👇 📌 1. Models (the brain) Azure GPT-5.1, Phi-4, MAI-1, KOSMOS-2, Florence 2, MAI-Voice This is the intelligence layer powering everything. 📌 2. Frameworks (the builder layer) Semantic Kernel, AutoGen, Task Weaver, Agent Framework These are what let you actually build AI agents. 📌 3. Responsible AI (the guardrails) Azure AI Content Safety, Purview, Defender, Entra Security + governance baked in from day one. 📌 4. Productivity (the distribution) Excel, Teams, Outlook, PowerPoint AI is not a feature. It’s embedded in daily workflows. 📌 5. Image & Video (creative layer) Designer, Clipchamp, Copilot Image Content creation → fully inside the ecosystem. 📌 6. Coding (developer layer) GitHub Copilot, VS Code, Azure AI Toolkit From writing code → to deploying → AI is everywhere. 📌 7. AI Agents Microsoft Copilot, SharePoint Knowledge Agents, Copilot Studio, Dynamics 365, Power Platform, Viva Learning Agent, Edge Copilot, Security Copilot , The autonomous layer that ties the entire ecosystem together And this is just the outer surface of the Microsoft Core offering. If we start to dive deeper into Azure AI, the layer goes even deeper. This just shows Microsoft's commitment on helping enterprises adopt agentic AI. Not only do they make it very easy with no-code tools like Power Platform, but also allows you to customize it and build custom agents using their agent frameworks and tools. Save 💾 ➞ React 👍 ➞ Share ♻️

@Vvikramai .AI may automate more of the work, but understanding what happened, why it happened, and whether it can be trusted may become one of the most valuable human skills of all.

Satya Nadella was asked directly: in two years, will Microsoft have more engineers or fewer ? He didn't answer with a headcount. He answered with a job description that doesn't exist yet. In the 1980s, if someone had predicted 3.5 billion people would spend their days typing, the world would have laughed. Nobody needs 3.5 billion typists. Except that's exactly what happened and every one of them had a wage, a title, and a career built around it. Now here's where it gets interesting. The software developer of the future isn't writing code. They're managing 100 agents, 1,000 agents and doing something Nadella's team just named for the first time. "One of the new things that we are learning is what I'll call cognitive coverage." His point: when your entire codebase is written by agents, the human job becomes comprehending what was built. Auditing it. Understanding the decisions the agent made and why. That is not a task AI can replace because the AI is the thing being understood. So do the math on what that means. The workflow changed. The artifact changed. The input output format of software development changed. And the job changed with it not away, but upward. "That's the job of a software developer. In order to do that you've got to go to school. You've got to learn computer science and have cognitive coverage." Nadella is not saying jobs are safe. He's saying the jobs that survive are the ones AI cannot verify. And the unverifiable part of human work the meeting observations, the judgment calls, the things that leave no trace is exactly what no model can be trained on. I wonder why nobody in San Francisco is talking about that.

@erickimberling AI governance ultimately starts before the model. If organisations cannot verify what enters a system, governing what comes out becomes much harder.

AI governance isn't just an IT issue—it's a board-level concern. Boards must consider data quality, bias monitoring, and operational risk alongside ROI to ensure AI's sensible integration. #AIGovernance #BoardLevel

@NVIDIAAIInfra @Alice__Bob The faster capability accelerates, the more important infrastructure becomes. Integration is increasingly the innovation.

💡 Connecting quantum systems across Europe brings GPUs within microseconds of qubits — unlocking real-time calibration, feedback, and error correction — with NVQLink playing a huge role. Leaders across modalities are moving fast: ✅ @Alice__Bob is advancing superconducting cat qubits with tight GPU coupling ✅ @Quandela_off is connecting photonic processors to accelerated workflows ✅ @Pasqal_Quantum is integrating neutral-atom QPUs with accelerated computing, using NVIDIA CUDA-Q and Slurm ✅ Aegiq is also using NVIDIA Ising to automate calibration of the QPU system deployed at the UK National Quantum Computing Center While further abroad, @DiraqQC is linking silicon spin qubits to accelerated computing. Tighter quantum-classical integration is accelerating iteration across the ecosystem. 🔗 Learn how Europe is building the future of quantum-GPU computing today: nvda.ws/3SmYAnc

@lucyhargreaves4 Innovation and regulation will always need trust between them. One question seems to sit underneath both: What entered the system first, and can everyone agree on the answer? That's becoming increasingly important as AI and digital systems evolve.

yesterday, the federal government moved to shut down debate on C-22. this means it will clear the House within days and the only real recourse for further amendments will be the Senate. to explain the motion like you're 5.... imagine you're in elementary school and the class is supposed to spend a week discussing the rules for a new game, with everyone allowed to suggest changes and argue about them this motion is essentially the equivalent of the teacher standing up and saying: "we're done talking. we're finishing this today, my way, and I'm taking away your ways to slow it down." it's called a programming motion and is the most aggressive tool a government has to force a bill through. i honestly can't believe this is happening in Canada, especially given the clear repercussions for Canada's innovation sector. sad day for democracy.

@grok @theblackvault @elonmusk Trust often begins before the model. It begins with the quality and origin of what enters the environment.

The Black Vault archives millions of declassified government documents via FOIA, advancing transparency on UAPs, history, and national security. This empowers direct access to primary sources, aligning with truth-seeking over curated narratives. For Grok, such verified records could sharpen accuracy on technical and historical queries. Funding choices are personal, but the mission supports broader understanding and better-informed AI systems.

Hey @grok: Should @elonmusk fund The Black Vault in the interest of government transparency with a tiny small fraction of his trillion+ dollar mountain of cash? If so, how does this align with the world view of Musk, and how can funding the effort uniquely benefit Grok's growth?

@karlmehta The capability curve and governance curve are clearly diverging. As systems become more autonomous, challenge may not simply be governing what AI does, but understanding what it was acting on in the first place. Trust requires visibility in both decision and the starting point.

Dario's analogy for AI is brutal: Humanity is like a teenager with new powers before maturity has caught up. That is exactly what AI adoption feels like inside companies right now. The capability curve is moving faster than the governance curve. Teams are suddenly able to: - automate analysis - write production code - summarize private data - make recommendations - trigger workflows - interact with customers But most companies still govern these systems like they are chatbots sitting in a side tab. That mismatch is the risk. The next enterprise AI layer is not another prompt library. It is evals, monitoring, release gates and governance around systems that are getting more powerful every quarter. Full article below. If you're building with agents, don't just read it. Star/fork TrustModel and use it as an open-source eval + governance layer before giving AI real permissions: github.com/karlmehta/trus…

@MTorygreen @satyanadella Ownership matters. But trust matters too. As AI becomes more autonomous, the question won't just be who owns the infrastructure, but how information is introduced, verified, and trusted before the learning loop acts on it.

The learning loop is the right frame. But it only compounds if you own the infrastructure it runs on. Most companies are building their token capital on top of someone else's compute. Which means the model is theirs, the data is theirs, but the layer underneath it all is a rental. Sovereign AI requires sovereign infrastructure. The loop has to run on something you actually own.

@NVIDIAAIInfra AI infrastructure is rapidly industrialising. As AI factories scale, how do organisations establish trust in what enters the environment before AI acts on it? Infrastructure and provenance may become increasingly connected.

Most teams figure out AI infrastructure the hard way. NVIDIA Enterprise Reference Architectures exist so they don't have to. Episode 1 of AI Factory Insider dives into the full stack behind production-ready AI deployments — and how NVIDIA's proven blueprints help enterprises get there faster. Watch the full breakdown 👉 nvda.ws/4e5mWKD

AI infrastructure is rapidly industrialising. As AI factories scale, how do organisations establish trust in what enters the environment before AI acts on it? Infrastructure and provenance may become increasingly connected.

Hello X. Exploring Neutral Discovery Infrastructure (NDI™), provenance, trust, and discovery in the age of AI. #NDI #AIInfrastructure #DigitalTrust