LLM on Cyril's Cloud & AI Daily

How 3 Java Champions are using GitHub Copilot for increased productivity

Tue, 20 Oct 2026 00:00:00 +0000

There’s been an important development in the GitHub Copilot space that caught my attention, and I wanted to break it down — not just what was announced, but why it matters and how it fits into the bigger picture.

What’s New

AI coding tools can generate code quickly, but speed alone does not solve the hardest software engineering problems. Developers still need to provide context, validate results, protect maintainability, and remain accountable for what ships.

Azure MCP Server 3.0 stable release: More tools. More possibilities

Wed, 14 Oct 2026 00:00:00 +0000

What’s New

We’re excited to share the stable release of Azure MCP Server 3.0, released on October 12, 2026. From reviewing backup protection to evaluating infrastructure changes, this release brings more everyday Azure operations into your agent workflows. It also introduces clearer tool contracts and security-focused defaults for new resources.

From extraction to action: Using grounding and confidence to build reliable document workflows

Wed, 14 Oct 2026 00:00:00 +0000

There’s been an important development in the Microsoft Foundry space that caught my attention, and I wanted to break it down — not just what was announced, but why it matters and how it fits into the bigger picture.

What’s New

“Civilization advances by extending the number of important operations which we can perform without thinking about them.” — Alfred North Whitehead

“Civilization advances by extending the number of important operations which we can perform without thinking about them.” — Alfred North Whitehead

Azure Content Understanding Prebuilt Analyzers: packaging domain knowledge for document AI

Tue, 13 Oct 2026 00:00:00 +0000

There’s been an important development in the AI Agents space that caught my attention, and I wanted to break it down — not just what was announced, but why it matters and how it fits into the bigger picture.

What’s New

For many intelligent document processing and emerging agentic workflows, extracting text is only the starting point. Applications increasingly need to capture a document’s structure and implicit business meaning in a form that agents and downstream systems can interpret accurately and act on.

GitHub Copilot app for Beginners: How to teach Copilot about your codebase

Mon, 12 Oct 2026 00:00:00 +0000

What’s New

Context matters when you’re working with AI. Every codebase has its own house rules: how components are structured, which scripts to run, and which conventions apply to different frameworks and file types.

Dynamic Workflows in Azure Functions Hosted Skills: Durable, AI-Led Work for Event-Driven Apps

Thu, 08 Oct 2026 00:00:00 +0000

What’s New

Azure Functions is the event-driven application platform in Azure. Your code runs when a queue message, a blob upload, a database change, or an HTTP request arrives.Azure Functions Hosted Skillsadds AI reasoning to these apps. When an event arrives, an LLM reads it, applies your instructions, skills, and tools, and decides what to do.

Bring agentic reasoning to Python function apps with Azure Functions Agent bindings (preview)

Tue, 06 Oct 2026 00:00:00 +0000

There’s been an important development in the Microsoft Agent Framework space that caught my attention, and I wanted to break it down — not just what was announced, but why it matters and how it fits into the bigger picture.

What’s New

Azure Functions applications already have a strong model for event-driven work: a trigger starts the function, bindings connect it to other services, and application code validates input and controls the result. AI reasoning doesn’t need to replace that model.

ReviewBench: An open benchmark for AI code review

Mon, 05 Oct 2026 00:00:00 +0000

What’s New

Agentic code review is becoming an essential piece of how development happens. It helps you inspect pull requests, catch issues, and decide what deserves attention before code ships.

But the quality of existing AI reviewers can be hard to measure, and you need to know the strengths of a reviewer before you know if it will help you. Some reviewers surface more issues, some produce less noise, and some are stronger at catching critical problems while others surface smaller improvements, too. You may need code review to do different things within your workflow.

Why content extraction still matters in the GenAI era

Tue, 29 Sep 2026 00:00:00 +0000

What’s New

The future of enterprise AI is not determined by the model you choose. It isdetermined by whether your agents can trust the content they act on.

Five years into the era of large language models, that thesis still holds. If anything, it is truer today than when we started. The models are stronger. The content is messier. Most of the real engineering lies in turning a raw PDF, image, or audio file into grounded, verifiable information an agent can act on.

How we found 24 Android vulnerabilities using our open source AI security agent

Mon, 28 Sep 2026 00:00:00 +0000

What’s New

With the rise of AI in the security space, our team created theGitHub Security Lab Taskflow Agentas a way for security researchers to easily automate, package, and share the AI prompts and workflows that they find effective for their work. In this blog post, I’ll share how I created auditing taskflows to find vulnerabilities in Android applications.

AI-powered fuzzing with the GitHub Security Lab Taskflow Agent

Thu, 24 Sep 2026 00:00:00 +0000

There’s been an important development in the Agentic RAG & Foundry IQ space that caught my attention, and I wanted to break it down — not just what was announced, but why it matters and how it fits into the bigger picture.

What’s New

If you’re new to fuzzing and want to learn the fundamentals first, check out our Fuzzing 101 course atgh.io/fuzzing101.

Continuous fuzzing is not a magic solution that solves all your problems . Even projects that have been enrolled in OSS-Fuzz for years can still hide critical bugs, and the reason is almost always the same: someone needs to keep an eye on coverage, write new harnesses for the code that nobody is reaching, and triage the crashes that come out the other end. In other words, fuzzing still needs a human in the loop.

Using Azure Blob Storage as a durable filesystem for LangChain Deep Agents

Tue, 22 Sep 2026 00:00:00 +0000

There’s been an important development in the Azure OpenAI & GPT Models space that caught my attention, and I wanted to break it down — not just what was announced, but why it matters and how it fits into the bigger picture.

What’s New

LangChain Deep Agentsis an open-source agent harness with built-in capabilities for building LLM-powered agents and applications, including complex, multi-step workflows.

A model can reason and generate responses, but it needs a harness to do useful work over time. The harness provides the tools and runtime that let the model retrieve the right context, take actions, and manage work across multiple steps. Deep Agents supplies that structure through planning, context management, a virtual filesystem, memory and skills, specialized subagents, and human approval points.

Announcing Azure AI Speech LLM 2607: Better Multilingual Accuracy, Easier Customization

Thu, 10 Sep 2026 00:00:00 +0000

What’s New

Azure AI Speech continues to deliver regular improvements to LLM Speech throughout the year, helping developers build more accurate and capable voice experiences. We’re announcingLLM Speech 2607, our latest model update, which brings improvements in multilingual recognition and mixed-language audio inputs, along with an enhanced experience for the phrase list feature.

What’s new in Microsoft Foundry: July and August 2026

Wed, 09 Sep 2026 00:00:00 +0000

What’s New

Microsoft Foundry’s July and August 2026 updates make Hosted Agents, Voice Live integration, and Toolboxes generally available; expand Claude tools hosted on Azure; update Model Router’s regions and model pool; add Foundry Local capabilities; and advance the Python, JavaScript, Java, and .NET SDKs.

How to evaluate LLMs before production

Tue, 25 Aug 2026 00:00:00 +0000

There’s been an important development in the Cloud & AI space that caught my attention, and I wanted to break it down — not just what was announced, but why it matters and how it fits into the bigger picture.

What’s New

A language model can perform well on a clean benchmark and still struggle with the cases that matter in production.

Benchmarks and curated datasets are useful when prototyping an LLM-based system. They help teams compare models, test an initial prompt, and determine whether an idea is technically plausible.

From single call to agents: five new Claude capabilities now available in Microsoft Foundry

Mon, 17 Aug 2026 00:00:00 +0000

What’s New

From single call to agents: five new Claude capabilities now available in Microsoft Foundry

Structured outputs, Web search, Web fetch, MCP connector, and Tool search are now available for Claude models in Microsoft Foundry hosted on Azure, the building blocks that turn a model endpoint into a production agent platform.

Azure Content Understanding GPT-5 Series Guide: Model Selection, Grounding Improvements, and Confidence Enhancements

Wed, 12 Aug 2026 00:00:00 +0000

What’s New

Enterprise content is no longer just something people consume. As organizations increasingly rely on AI to extract and act on information from documents, images, audio, and video, Azure Content Understanding is expanding support for the GPT-5 series and improving grounding and confidence to deliver greater flexibility, efficiency, and quality.

From Sync APIs to support for the GPT-5 model series and agentic workflows: What’s new in Azure Content Understanding – August 2026

Wed, 12 Aug 2026 00:00:00 +0000

What’s New

Enterprise content is no longer just something people read. AI apps and agents are only as useful as the information they can understand, yet much of the world’s enterprise knowledge is locked in documents, forms, tables, images, audio, and video. The latest Azure Content Understanding updates help developers turn that content into structured, grounded data with less custom processing. This release brings broader support for the GPT-5 model series, lower token usage, improved confidence scoring, new synchronous APIs for Read and Layout, advanced contextualization for higher-quality extraction, new tax-focused prebuilt analyzers, semantic chunking for retrieval workflows, and agentic document reasoning for more complex extraction scenarios.

Automating cross-repo documentation with GitHub Agentic Workflows

Wed, 08 Jul 2026 00:00:00 +0000

What’s New

“Where are the docs?” It’s a question nobody on a product team enjoys answering. The honest reply is usually some variant of “behind.” A writer is staring at a closed pull request, trying to reverse-engineer what changed. The pull request’s author has already moved on. By the time the doc actually publishes, the feature has shipped, sometimes more than once.

What’s New in Microsoft Foundry | June 2026

Tue, 07 Jul 2026 00:00:00 +0000

What’s New

Here’s what shipped in Microsoft Foundry in June 2026 — Claude reached general availability, Foundry agents started publishing straight into Microsoft 365 Copilot and Teams, and Toolboxes, Routines, and Memory all got meaningfully more capable.

How we made GitHub Copilot CLI more selective about delegation

Fri, 12 Jun 2026 00:00:00 +0000

What’s New

In agentic systems, more delegation isn’t always better. Imagine asking Copilot CLI to make a simple change. Instead of handling it directly, it spins up a helper agent that searches the repository, waits on a result, and stalls. Work that should have taken one step now takes three. While some tasks genuinely benefit from a specialist subagent—like exploring an unfamiliar repository, checking an independent area of the code, or running a long command while the main agent keeps moving—delegation isn’t free. Every handoff adds coordination overhead, tool calls, and wait time. If an agent delegates too eagerly, the “help” can become friction.

Making secret scanning more trustworthy: Reducing false positives at scale

Thu, 11 Jun 2026 00:00:00 +0000

What’s New

Secret scanning plays a critical role in protecting developers and organizations. It helps catch exposed credentials early and prevents small mistakes from turning into real incidents.

At GitHub’s scale, even small inefficiencies create real friction. Too many false positives make alerts harder to trust.

Build 2026: From observability to ROI for AI agents on any framework

Wed, 03 Jun 2026 00:00:00 +0000

What’s New

9 min read · June 3, 2026 · Sebastian Kohlmeier

Shipping an AI agent is the easy part. Keeping it accurate, safe, and accountable in production is where teams get stuck. Agents are non-deterministic. Their behavior shifts as models update, tools change, and traffic patterns evolve and most of that drift happens silently, long after the demo.End-to-end observabilitycovering the full development lifecycle is how you close that gap: See every step an agent takes, evaluate quality and safety against criteria you define, optimize what isn’t working, and prove the business value of what is.

Build smarter document workflows: What’s new in Azure Content Understanding at Build 2026

Wed, 03 Jun 2026 00:00:00 +0000

What’s New

Azure Content Understanding in Foundry Tools is Microsoft’s comprehensive content AI service. It ingests diverse data types — documents, audio, images, and video — and extracts the most critical information to power well-grounded, reliable generative AI and agentic solutions. Azure Content Understanding brings together Azure Document Intelligence’s proven traditional AI with advanced LLM-based content reasoning, enabling both structured and unstructured content extraction, as well as multimodal understanding to address your full spectrum of processing needs.

Expanding the Reach of Document Translation – New Capabilities Announced at Microsoft Build

Wed, 03 Jun 2026 00:00:00 +0000

What’s New

Our commitment to advancing cross-language communication takes another major step forward at Microsoft Build. Azure Translator in Foundry Tools is expanding Document Translation with a set of new capabilities designed to meet enterprises where their content lives today — across formats, modalities, and languages. From large language model–powered translations to native image translation and expanded structured content support, these updates help organizations translate more of what they produce, with higher fidelity and less friction.

Azure Translator: Improving Translation Quality with Adaptive Datasets and Few‑Shot Learning

Tue, 02 Jun 2026 00:00:00 +0000

What’s New

Your healthcare app needs “La médica” not “El médico.” Your legal documents need precise terminology, not generic translations. When domain-specific language matters, generic LLM translation falls short.

Azure Translator’s adaptive translation lets you teach the model your terminology with just a handful of examples—no model training required. In this walkthrough, you’ll create an adaptive dataset, compare baseline vs. adapted translations side-by-side, and see exactly how much difference domain context makes.

What’s new in Microsoft Foundry | May 2026

Sun, 31 May 2026 00:00:00 +0000

What’s New

TL;DR

Trace-based evaluation for external and hosted agents:Grade real production traces from Foundry, GCP, AWS, or any framework — no hand-curated datasets required.
Grok 4.3:xAI’s latest model lands in Foundry for advanced agentic and domain-specific workloads.
DeepSeek V4:DeepSeek’s newest model family expands open-model choice in the catalog.
Fireworks AI — May update:DeepSeek V4 Pro and Kimi 2.6 arrive via Fireworks for high-performance open-model inference.

Eliminate LLM Cold starts: Load models up to 6x Faster with Azure Blob Storage and Run:AI Model Streamer

Tue, 19 May 2026 00:00:00 +0000

What’s New

Stop paying for idle GPUs while model weights copy to disk. Stream them straight into GPU memory instead with Run:AI Streamer from Azure Blob Storage.

The Problem: Every Cold Start Costs You More Than Money

GPU compute is among the most expensive cloud infrastructure, and every second a GPU is allocated but unavailable for serving is real money lost. The cost also goes beyond your Azure bill: slow cold starts can delay responses, stress SLAs, and degrade user experience during traffic spikes, when users need capacity most.

How to run evals for the model router

Tue, 19 May 2026 00:00:00 +0000

What’s New

One endpoint. Smarter spend. Model router in Foundry Models picks the optimal LLM for every prompt in real time based on signals like complexity, reasoning, and task type. Now with access to 28 frontier models, the model router makes model selection easier for developers and reduces manual overhead. This article walks you through how to run evaluations using a new open-source GitHub repo designed explicitly for the model router.Access the Eval Repo on GitHub

Foundry Local 1.1: Live Transcription, Embeddings, and Responses API

Tue, 12 May 2026 00:00:00 +0000

There’s been an important development in the RAG (Retrieval Augmented Generation) space that caught my attention, and I wanted to break it down — not just what was announced, but why it matters and how it fits into the bigger picture.

What’s New

Today we’re announcing the1.1.0 release of Foundry Local— Microsoft’s cross-platform local AI solution that lets developers bringAI directly into their applicationswith no cloud dependency, no network latency, and no per-token costs.