AI
Building on Vercel’s eve + Azure Cosmos DB: An Agent That Remembers
Most "AI agent" demos forget everything the moment the process exits. That's fine for a toy project, but useless for anything real. An agent that helps you write, triage, or support needs two things a language model alone can't give it: durable state and the ability to recall the right context by meaning. This post shows how to build exactly that by integrating two pieces that fit together surprisingly well: Eve — Vercel's filesystem-first agent platform. Drop a file in agent/tools/, and it becomes a tool the model can call. Azure Cosmos DB JavaScript SDK — the official, promise-based client for Cosmos DB N...
The hidden variables in your agent eval
This is the seventh article in a series about Agent Experience (AX): the practice of making AI coding agents work correctly with your technology. The series covers what you can and can't control in the agent stack, how to measure whether your extensions are helping or hurting, and how to iterate toward better outcomes. You build an eval. You run it on your machine. You get a score. Your colleague runs the same eval on their machine and gets a different score. Same scenario, same setup. What changed? In the previous article, we covered why public benchmarks can't tell you which model works best for your stack. T...
Don’t rewrite your CLI for agents
There's advice making the rounds: replace your CLI args with a single payload so agents can use your tool more effectively. The thinking being, that agents already think in structured formats, and nested data maps cleanly to JSON. Flat args on the other hand, force awkward conventions like repeating to delimit multi-value groups, which is inherently ambiguous. Not to mention, that the agent needs to get the types of all values right. It's a reasonable hypothesis, and we wanted to know if it holds up under measurement. The data we collected, showed something interesting. What we tested We built a synthetic CL...
Not all model upgrades are upgrades
A new model drops with lower per-token pricing and better benchmarks. You switch. A week later someone asks why the agent is burning 12x more tokens on the same task while producing worse output. We ran 150 agent tasks across 15 scenarios on two models, Claude Sonnet 4.6 and Claude Sonnet 5, using GitHub Copilot Chat in VS Code on Windows. The scenarios covered two types of work: architecture and design tasks grounded in Microsoft Learn documentation, and SharePoint Framework project upgrades. Sonnet 5 is the newer model with 33% lower per-token pricing across every token category. The assumption we wanted to te...
Enabling MLflow OpenAI Autolog on PySpark Workers
When distributing LLM calls across PySpark workers via mapInPandas, MLflow autolog silently fails. Here is how to fix it.
What AI benchmarks are not telling you
This is the sixth article in a series about Agent Experience (AX): the practice of making AI coding agents work correctly with your technology. The series covers what you can and can't control in the agent stack, how to measure whether your extensions are helping or hurting, and how to iterate toward better outcomes. We love benchmarks. A new model drops, the leaderboard says 92% on SWE-bench, and your timeline declares it "the best coding model." You switch to it, run your agent on your codebase, and outcomes are... the same. Maybe worse. The leaderboard said 92%, so what happened? In the previous article, we ...
What’s new across Microsoft SQL in 2026 so far (SQL Server, Azure SQL, and SQL database in Fabric)
We’re halfway through 2026, and Microsoft SQL has not slowed down. Since SQLCon/FabCon in March (where we released a ton of things, and those updates can be found in this udpates video), we shipped a wave of updates across SQL Server, Azure SQL, and SQL database in Fabric, with Microsoft Build 2026 as the centerpiece. If you want the details on Build, start with my recap blog, The Era of the Agentic Database Developer. This post collects everything new from mid-March through today, organized by service so you can find what matters to you. Every item is tagged Preview or Generally Available with a link. You can...
MCP Beyond the Chat Window: Build Diagnostics in CI
A practical tour of the Model Context Protocol tools for .NET build diagnostics - the full Binlog MCP toolset, how those tools run inside a GitHub Actions workflow, and what the evaluation data says about the efficiency gains.
Spring AI 2.0 is GA: Vector Search, Memory, and Agents on Azure Cosmos DB
The wait is over. Spring AI 2.0 is generally available, and Azure Cosmos DB is right there with it. With this release, Spring AI graduates into a mature, production-ready framework for building AI applications in Java, and Azure Cosmos DB ships dedicated, vendor-maintained integrations that plug straight into the Spring AI ecosystem. The Spring AI 2.0 GA announcement names Azure Cosmos DB among its vendor-maintained modules, maintained directly by Microsoft rather than the core Spring AI team. This means the integration is built and supported by the engineers who work on Cosmos DB itself, bringing deep, first-ha...
Your agent already has a plan
If an agent isn't doing the right thing, the obvious move is to make the docs clearer. Add a tip, spell out the correct command, describe the right approach more prominently. You do all of that, and the agent still ignores it. It does what it had already decided to do. The tip wasn't ignored because it was unclear, it was ignored because the agent had already made its plan before it read the page. The agent plans before it reads An AI coding agent doesn't arrive at your documentation as a blank slate. The moment you give it a task, it forms a plan based on what it learned during training, before it goes and...