OK. It's time to rethink this. I've spend the last several weeks working on a harness that tightly constrains the agents to work the way that I want them to work. I set up all kinds of gates, and tests, and tools, and protocols, and ... And while I was heads-down getting that to work, the agents got a LOT better. So much so that when I came up for air, the need for my harness was obviated. Indeed, the need for _any_ but the most liberal of harnesses may be obviated. Just how good these things have gotten blows me away. I have had long debates with grok and codex about the structure of systems -- as if they were senior engineers. They often disagree with me and have their own perspectives. I have, more than once, found myself agreeing with their views. I have not given up on constraints and tooling. Unit testing is still important. So is CRAP and Mutation testing. These tools still find bugs and offer useful constraints, though they can leave scars. However, the agents have gotten so good that I can now give one a very significant task with a few guidelines and it will faithfully implement it. I can walk away for 40 minutes and when I return it will be done. CRAP will be satisfied, Coverage will be high, and Mutation testing complete. The architecture will be clean, and the code will be very good. The end result may not behave perfectly, but it's so close that a couple of tweaks usually puts it into place. What does this mean going forward? I'm not sure. But I'm beginning to think that harnesses should not treat agents as components within a software design.

Databases certainly aren’t dead. I will say, however, that they are often used by default when other solutions write be more appropriate. I also think that SQL will likely see a gradual asymptotic decline as it gets shoved further out of the minds of programmers.

Code review is becoming the bottleneck in AI-assisted development. Uncle Bob says he doesn't read the code his AI agents write. The idea is replacing line-by-line review with tests, constraints, QA, and verification. My take: youtu.be/sClTAvkQDOU

Microsoft quietly published 99 ready-made .NET skills for AI Agents. Most .NET developers have no idea they exist 👇 These are not toy examples They are official and open-sourced. And they work across Claude Code, Codex, GitHub Copilot, and Cursor. Quick reminder first. A skill is a folder of instructions that your AI loads only when a task matches. You write the recipe once, and it follows your steps every time. Microsoft ships these in the dotnet/skills repo: 14 plugins, around 95 skills. A plugin is just a group of related skills. Here are 5 worth installing today: 𝟭. 𝗱𝗼𝘁𝗻𝗲𝘁-𝗱𝗮𝘁𝗮 The optimizing-ef-core-queries skill finds N+1 problems, fixes tracking modes, and removes common EF Core performance traps. 𝟮. 𝗱𝗼𝘁𝗻𝗲𝘁-𝘁𝗲𝘀𝘁 The largest plugin: 22 skills for running, filtering, and migrating tests between frameworks. 𝟯. 𝗱𝗼𝘁𝗻𝗲𝘁-𝘂𝗽𝗴𝗿𝗮𝗱𝗲 Migrates a project from .NET 8 to 10, and turns on nullable reference types. 𝟰. 𝗱𝗼𝘁𝗻𝗲𝘁-𝗮𝘀𝗽𝗻𝗲𝘁 Creates Web API endpoints and file-upload endpoints in Minimal APIs the right way. 𝟱. 𝗱𝗼𝘁𝗻𝗲𝘁-𝗮𝗶 Builds, debugs, and tests MCP servers with the C# SDK. Installing them in Claude Code takes 4 steps: antondevtips.com/blog/creati… But here is the catch. These skills know .NET. But they do not know YOUR project. They cannot tell where you put your entities, or that you don't use repositories, or that your handlers return Result. That is the gap your own skills fill. ✅ Microsoft's skills → general .NET tasks ✅ Your own skills → your house rules and conventions Pair them, and Claude (or other AI agent) has both the framework and your style. In my newest article, I break down the whole system: → What a skill is and the building blocks inside one → How to write the description so a skill triggers at the right moment → How to encode your conventions: rich entities, no repositories, full features → Using the Skill Creator to bundle your real code files → The complete Microsoft dotnet/skills setup Read the full guide here: 👉 antondevtips.com/blog/creati… Have you tried any of Microsoft's .NET skills yet, or is this the first you're hearing of them? —— 𝗕𝗲𝘀𝘁 𝘁𝗼𝗼𝗹𝘀 𝗼𝗳 𝘁𝗵𝗲 𝘄𝗲𝗲𝗸: Redis Iris is a real-time context engine that transforms fragmented data into live, agent-ready context and memory. Try for free today: fandf.co/4xLtOo8 exe[.]dev runs persistent VMs for the agent era. SSH and root, plus HTTPS and auth out of the box. Secrets injected at the network edge stay out of the LLM's hands. Run agents, internal tools, side projects, anything: fandf.co/4wegPKf —— ♻️ Repost to help other .NET developers discover these skills ➕ Follow me ( @AntonMartyniuk ) to improve your .NET and Architecture Skills

"I think that typing code manually is dying." - @shanselman on The Agentic Review podcast We got into what that means and much more on the last episode. If you're thinking about where the craft is actually heading, this one's worth your time.

JSON in a relational database is not only acceptable, but often encouraged

Picture credit: @AussieStard .

How do SQL Joins Work? The diagram below shows how 4 types of SQL joins work in detail. 🔹 INNER JOIN Returns matching rows in both tables. 🔹 LEFT JOIN Returns all records from the left table, and the matching records from the right table. 🔹 RIGHT JOIN Returns all records from the right table, and the matching records from the left table. 🔹 FULL OUTER JOIN Returns all records where there is a match in either left or right table. –- Subscribe to our weekly newsletter to get a Free System Design PDF (158 pages): bit.ly/3KCnWXq

At the University of Amsterdam, Jihadists beat students with sticks. Europe is committing suicide.

Dutton's now saying that he will ban foreign buyers and temporary residents from buying existing homes. He must have missed the fact that we're already doing that... It comes into effect next week. #Budget2025 ato.gov.au/about-ato/new-leg…

Implants can help deaf kids hear—but many still struggle with spoken language | Science | AAAS science.org/content/article/…