Surprise! My book arrived two weeks before the official release date. 🎉 "Spec-Driven Development: From Specs to Code with AI Agents" (Apress Pocket Guides) is now available, both as softcover and eBook. It is a compact guide on how to make specs the contract between you and your AI agents, based on the AI Unified Process (AIUP). Requirements, use cases, entity models, and a complete case study, all in about 150 pages. I hope it helps you bring more structure into AI-assisted development. 👉 link.springer.com/book/10.10… #SpecDrivenDevelopment #AIUnifiedProcess #GenerativeAI #KeepITSimple

On my way to Antwerp

I’m overwhelmed by the number of people planning to attend my deep dive tomorrow at @Devoxx 😀 See you all in Antwerp 🇧🇪

After many years I cancelled my JRebel subscription. I don't see any use case for it in times of AI.

I try to keep my setup for editing markdown as simple as possible. The screenshot shows on the left @CotEditor and on the right @macmdviewer for the preview. This follows the UNIX philosophy: one tool for one thing.

On Monday afternoon I will have a deep dive on Spec-Driven Development at Devoxx Belgium. So far over 200 people have favorited the session. That's an amazing number and I really look forward to it. I hope to see you all on Monday afternoon in Antwerp.

That's super cool. I love @flydotio Sprites

Safe travels to all Devoxxians. See you soon! #Devoxx

Spec Kit or AI Unified Process? Both start from specifications, but they solve different problems. People often ask me how the AI Unified Process (AIUP) compares to GitHub Spec Kit. Short answer: Spec Kit helps a developer ship a feature. The AI Unified Process helps a team maintain a system for years. GitHub Spec Kit is developer-centric → The developer writes the spec, mostly as a structured prompt for the coding agent → specify → plan → tasks → implement, one folder per feature → Once the feature ships, the spec has done its job and the code becomes the source of truth again That's a lightweight way for a developer to bring structure into their next AI coding session. The AI Unified Process is built to last → Use cases describe system behavior and stay the source of truth → Business analysts and requirements engineers own the specs, and AI drafts them → Specs live as long as the system does, versioned in Git next to the code → Every requirement traces to a use case, the code and its tests → For existing systems, you derive the specs from the code first The real difference is who owns the specification and how long it stays valid. AI can regenerate code in minutes. It can't regenerate the knowledge of what the system is supposed to do. That knowledge has to be maintained. Is your spec still accurate after the merge? 🔗 unifiedprocess.ai #SpecDrivenDevelopment #AIUnifiedProcess #RequirementsEngineering #SoftwareArchitecture #GitHubSpecKit

In the face of generative AI coding agents, the role of human-led architecture becomes even more important in the crafting of software-intensive systems. All architecture is design, but not all design is architecture. Architecture represents the set of significant design decisions that shape the form and function of a system, where significant as measured by cost of change. As such, remember that coding agents are reasonably good at transforming prompts into code, but they are not particularly good at making decisions that balance the static and dynamic forces on a system in order to yield a reasonably optimal solution: this is where human judgment, human understanding, human knowledge, and human experience shine.

One thing that I hate when flying is that. A few days before the flight, I get those emails sometimes multiple times.

Vaadin 25.3 is out, one of our largest feature releases yet. Dev Loop for AI agents, a self-hosted MCP server, Observability Kit 5 without a Java agent, auditable AI form filling, a new Switch component and more. Read the full release post: vaadin.com/blog/vaadin-25-3-… #java #ai

My advice has been don't use AI to scaffold a new Spring Boot project because you get outdated version. Today I did that in a demo, and guess what, Opus 5.5 uses the Spring Initializer 🤯 That model is awesome! Luckily switching to Sonnet 4.6 helped and it created a Spring Boot 3.3.4 project 😅

It’s artificial intelligence, Sundar. And in the history of computing, this puerile attempt to rename it simply because of the simpleminded whims of an incompetent and corrupt narcissist will be seen as a failure of leadership, an abandonment of intellectual integrity, and an embarrassment for any who join with him.

Gute Fehlermeldungen waren schon immer schwierig. Aber bitte @Outlook, soll ich einen neue Version vom Browser installieren oder die Seite neu laden?

In just over an hour, @simas_ch, @jbaruch, and yours truly are talking about their processes of AI (assisted) development in the 8th episode of #JRush. Set a reminder on YouTube! piped.video/live/nw6C6WNZO54…

Sonnet 5.5 is out. It gets remarkably close to Opus 5.5 on most coding and agentic work, at a fraction of the cost and noticeably faster. Opus is still the go-to for the gnarliest, long-horizon problems. For everything else, Sonnet 5.5 is going to be hard to beat.

Agent code quality may look worse to a human observer, but in my experience agent build quality is better. Features have less bugs on launch and tend to "just work", whereas human-driven builds always tended to be myopic and fail for embarrassing reasons.

A lot of tool companies are trying to mimic the current development process with AI. Why? We must rethink the development process.

⚡️Tomorrow at JRush: When AI meets real software development @simas_ch , @jbaruch , and @asm0di0 will talk about spec-driven development, coding-agent factories, context, reviews, and engineering decisions belonging to humans. Sign up and see you there! jrush.bell-sw.com/episode8

Amazing! This was in my inbox today. Congratulations! Your session From Developer to Builder: A Spec-Driven Development Deep Dive has been accepted for Jfokus 2027! Thank you @Jfokus 🙏