Recent content on Unladen swallow - Olivier Wulveryck
A manifesto is something you sign. A guiding principle is something you execute. The whole argument of this article fits in that distinction. Introduction When I first saw the AI-Driven Development Manifesto (AIDD), I thought it was an excellent idea. Four values, twelve commitments, written by developers against vibe coding, in the “X over Y” lineage of the Agile Manifesto. At first…
Introduction The previous article argued an architecture: move governance from the edges of the agentic loop (a wall of instructions at minute zero, a CI gate at the end) to the inside of the loop. It introduced a working proof of concept, the Platform Planning Gateway (PPG). This article does something simpler: it tells a story. One task, one agent, two runs. In the first run there is no gateway;…
Introduction The previous article closed the architecture of the Platform Planning Gateway: a proof of concept that governs how agents execute. Plans are linted against Rego policies derived from Architecture Decision Records, capability tickets constrain every tool call, and the smart tools verify the ticket before acting. The execution plane is governed. There is a second plane that is not: the…
Introduction This article closes a trajectory started three posts ago. In See, Act, Correct , I described the three levers that turn a code agent from a gadget into a production tool, and introduced a grid to sort durable platform investments from temporary crutches. In Who does what? , I mapped the agentic organization onto Team Topologies . In Codifying the Rules , the enabling team baked its…
Introduction This article is a follow-up to our deep dive on Team Topologies . Building reliable software at an organizational scale—across multiple products and teams—requires a fundamental shift in how we operate. In a modern, AI-driven Software Delivery Lifecycle (SDLC), the stream-aligned team should focus entirely on the solution , delegating the actual implementation to an AI system. This…
Exposing the problem Giving every developer a powerful, local AI agent feels like the ultimate productivity hack. But for organizations running at scale, it is a governance and cost trap waiting to spring. Currently, the AI revolution in the Software Development Lifecycle (SDLC) is happening almost entirely on developers’ laptops. We are building isolated, monolithic agent loops. I’ve…
The agentic platform defines what needs to be provided. Team Topologies defines who provides it, and how teams interact to make it happen. This is a second version more human written. The first version was roughly a translation made by AI. This version is human-edited. Hopefully it will be easier to read by a human. In the first article of this series , we asked the what : which systemic…
The agentic platform defines what needs to be provided. Team Topologies defines who provides it, and how teams interact to make it happen. ⚠️ This article is now obsolete and is replaced by a human-reviewed version here ⚠️ EDIT: Author’s Note & Hacker News Aftermath: This article was originally written for a French audience, and I used an AI ghostwriter to translate and adapt it for English…
Generative AI has transformed how code is produced. In just a few months, we went from autocomplete to agents capable of writing, testing, and deploying entire applications. The market is now flooded with methods for framing these agents and making them produce quality code. But this abundance raises a question few organizations are asking yet: what happens when you are not building one app, but…
An out-of-the-box code agent only sees a repo and a shell. For professional engineering, that is not enough. Here are the principles that make the difference between a gadget and a production tool. Foreword. This article grew out of the talk “Beyond the Basics with Claude Code” by Daisy Holman, an engineer on the Claude Code team (May 2026). The founding ideas come from that talk, then…
I am a consultant, and I regularly build presentations with Google Slides. My communication team has created dozens of pre-formatted templates (slides designed to convince , not just to present). The problem: choosing the right slides to illustrate the right narrative takes time, and filling them in mechanically adds no value. I built a multi-agent system to automate that part and focus on what…
Today, we see many initiatives around the agentic paradigm. Most revolve around systems built by AI giants (Anthropic, Google, OpenAI) and often boil down to pushing natural language directives to an off-the-shelf orchestrator. You specialize an integrated system like Claude Code through markdown files, skills, and tools. Frameworks like BMAD illustrate this approach well: they transform such a…
Note : this article has been translated by an AI; original is in French Introduction I have always been fascinated by AlphaGo , then AlphaZero , and this concept of reinforcement learning . A machine that plays against itself, learns on its own, and reaches superhuman levels. There is something magical there that goes beyond mere technical prowess. When in 2016, AlphaGo defeated Lee Sedol, one of…
Since its announcement a couple of months ago, the Model Context Protocol (MCP) has gained significant attention. Initially, MCP served as a straightforward mechanism, essentially a plugin system, to grant Large Language Models (LLMs) the ability to interact seamlessly with their environment, thus enhancing AI assistants and agents. Applying Wardley’s theory to trace the protocol’s…
This exploration of the Model Context Protocol (MCP) is presented in three distinct parts: Part I , the current article , introduces the foundational concepts of MCP. Part II will delve into a specific implementation , demonstrating a custom host leveraging Google’s VertexAI API and the Gemini model . Part III will showcase a practical custom server implementation tailored for a particular…
In the first part of this series, I explored some concepts and convictions regarding agentivity in AI and the potential of tooling to enhance agents. The final words were about the host (the application) running LLM-powered assistants ( Claude , ChatGPT , …). These applications will represent the true battleground. The companies that gain a monopoly on assistant systems will govern…
In the first two parts (see part I and part II ) of this series, we have: described the agentic system and how it could be extended via tool usage; explained how a standard like MCP was key to building a virtuous ecosystem; built a host “from scratch”: we implemented a private chatbot and gave it the ability to run a tool. In this final part, we will fulfill the initial promise: we…
In today’s hyper-competitive world, businesses no longer rely solely on gut decisions or intuition; they depend on data-driven insights to stay agile and make fast, smart decisions. However, data alone isn’t the answer; it’s the enabler to create momentum on a business & technology flywheel : a model where data drives decisions, decisions drive actions, and those actions drive…
Context In a recent meetup I organized in my hometown of Lille, I had the pleasure of hosting Jean-Georges Perrin, who provided a comprehensive introduction to data contracts. As a geek, I felt compelled to test this concept to fully grasp its practical implications. The goal of this article is to demonstrate how data contracts can be applied to and add value within a small ecosystem facing…
I am Olivier Wulveryck and this is my personal blog.