Focus

Agentic Software Engineering

AI is changing the way we build software. Agentic Software Engineering takes it further. Instead of applying AI to isolated steps, specialized agents handle distinct parts of the software lifecycle, such as planning, implementation, testing or documentation. These agents follow defined roles and are coordinated within structured workflows. Engineers set goals and constraints, manage agent interactions and review results. This creates a new kind of collaboration between humans and machines, shifting the focus toward quality, speed and scalability.

Why 75 Dollars in Tokens Can Create So Much Value

75 dollars in token costs sometimes buy a week of work or more. What that means economically, and why companies often draw the wrong conclusion from it.

How do you train people on a subject that changes every week?

For the past year, we’ve been training development teams to use AI agents across the entire software development lifecycle. In that time, we’ve delivered 47 trainings to more than 500 participants. Here’s what goals companies bring to us, what sets our training apart, which moments have stuck with our trainers, and why the real work starts after the training ends.

The New Economics of Software Development

Invest AI productivity gains in software quality instead of increasing throughput

Beyond Claude & GPT

Sure, we enjoy trying out the latest frontier model as much as anyone. But in day-to-day project work, what matters most is control over cost, data, and availability. That’s why we increasingly rely on open-weights models paired with open source tooling. Here’s what our setup looks like.

Trust but Sandbox

Coding agents can now do (almost) everything. They read and write every file you have access to, they run arbitrary commands, talk over the network, read your environment variables, and drive your browser with all your cookies. In the spirit of an agentic workflow, where we let agents work as autonomously as possible, that is actually the point, because it is exactly what makes them useful. Even so, we should find a way to gain more control over what happens in there.

Alternatives to U.S. Artificial Intelligence

Why rising AI costs and digital sovereignty concerns are pushing me to look beyond Claude and GPT

Faster Is Not Better: On Experimentation Culture and What It Costs

Faster experimentation sounds like an obvious advantage. But experimentation culture is not a universal paradigm. It is a product of a very specific context. And the people who bear the heaviest cost of poor product development never appear in any experiment.

The Agentic Trio

The Product Trio separated discovery from delivery because delivery required a larger team. Agentic development removes this constraint. That changes what a product team can look like.

Using the space, not optimizing the treadmill

Why agentic development needs product discovery

Nebu: Self-made sovereignty

There are plenty of open-source alternatives to Slack. None of them deliver real sovereignty. Community editions are deliberately pared down so no company can run them in production, while enterprise features sit behind commercial licenses. This article shows why “building it yourself” is no longer a utopian idea today, but a realistic option for anyone with a product vision and a bit of DIY courage.

From Vibe Coder to Agentic Engineer

I no longer read every line of code that’s produced in my projects. I think most people who work intensively with AI agents feel the same way. When an agent produces a thousand lines in ten minutes, line-by-line review is no longer realistic. At the same time, that code ends up in production, and someone has to take responsibility for it.

Hail Mary: Why domain knowledge cannot be extracted from experts

Developing with AI through the cognitive Lens

The Right Kind of Hard

More structure, more output, more exhaustion. Spec-driven development has helped me get better results from AI agents – and drained my energy in the process. About dead weight, false clarity, and hidden costs I didn’t plan for.

Where AI Helps (and Hurts) Across Different Coding Scenarios

Where does LLM-assisted software development improve developer productivity—and where does it fall short? Instead of treating AI in software development as a one-dimensional productivity booster, we explore the question across several dimensions suggested by a Stanford-adjacent study: project maturity, task complexity, and the popularity of the programming language. The goal is to create a more realistic expectation baseline for both software developers and leaders—well away from today’s hype.

Spec-Driven Development is Domain-Driven Design’s Impatient Cousin

Why BMAD won’t save you

REST Beats MCP

Instead of using dedicated APIs, agents can operate existing web applications directly. Like humans, they rely on the most consistent implementation of hypermedia and benefit from existing context, validation, and access controls. Using an expense report as an example, I show how agents can automate complex, context-dependent tasks—without having to implement new APIs.

Agents good in the end?

Software development consists of a constant chain of trade-offs. As long as I’ve been building systems, there have always been things you don’t do despite wanting to do them. But now, with agents, we can fulfill all our dreams and finally build everything we always wanted and had the feeling we were missing. Spoiler alert: we shouldn’t do that.

Understanding AI Coding Patterns Through Cognitive Load Theory

Developing with AI through the cognitive Lens

Neuland reloaded

When Angela Merkel said in 2013 that the internet is ‘Neuland’ (uncharted territory) for all of us, the amusement online was considerable. Viewed from some temporal and substantive distance, one can now recognize a perspective from which she was indeed correct.

I sandboxed my coding agents. Now I control their network.

I want my AI coding agents to work independently, but I don’t want them to have unrestricted access to the internet. In this post, I describe how I routed all network traffic from my development sandbox through a strict proxy allowlist, allowing only a small set of explicitly approved domains. This setup finally gave me enough confidence to loosen the guardrails without constantly staying in the loop.

From Vibe Coder to Code Owner

Spec-Driven Architecture: When Agents Build, Architecture Must Speak

Spec-Driven Development gives agents a clear foundation for implementation. What it doesn’t solve is how a portfolio of systems stays coherent. Spec-Driven Architecture applies the same principle at the architecture level, using contracts as versioned boundaries and guarantees—enforceable in agentic workflows and in the CI/CD pipeline.

Let’s Not Normalize Insecure AI Assistants

AI assistants like OpenClaw promise convenience, autonomy, and increasingly personalized help. But beneath that promise lies an architecture that quietly combines private data, internet access, and exposure to untrusted content—the lethal trifecta of security risk. Sandboxing and physical isolation help, but they don’t address the core problem. As we add more capabilities, the potential blast radius only grows. This post is a case for slowing down, questioning defaults, and refusing to normalize insecure architectures.

What to Do While the AI Is Thinking

How deep work, multitasking, and code reviews change as we delegate tasks to AI

AI and Elaboration: Which Coding Patterns Build Understanding?

AI tools let you complete coding tasks without connecting new information to your existing mental models—a cognitive process known as elaboration that is crucial for building understanding. But some AI interaction patterns preserve this elaboration while others bypass it entirely. Let’s explore what elaboration is, why it helps with learning, and how we can use AI tools in a way that helps with this process rather than circumventing it.

I sandboxed my coding agents. You should too.

LLM coding agents are extremely powerful because they can run programs on our computers using our permissions. However, this same power also makes us very vulnerable. It only takes one mistake or one prompt injection to compromise the whole system.

Speed vs. Skill

Developing with AI through the cognitive Lens

Context Engineering: Managing AI-Generated Code Complexity

AI tools make developers more productive at writing code, but can overwhelm code reviewers with massive changes. Learn practical strategies for managing context in AI-assisted development to keep your code comprehensible, your reviews manageable, and your team’s productivity genuinely improved. Small scope was always good practice—with AI, it’s essential.

Swiss Army Knife for Salesforce: LLM with In-Memory Database

Large Language Models (LLMs) struggle with transforming large datasets, particularly when performing aggregations that require calculations - essentially anything where you’d normally use GROUP BY or ORDER BY in SQL. But what if we need to retrieve extensive data from a system and process it with an LLM, but only have an API with limited functionality available? In this blog post, I present a solution I’ve implemented and tested.

AI — Behind the Buzzword Garbage

First Agile, Then Agentic

Agentic AI is supposed to accelerate software development. But new technologies can only reach their full potential when organizations adapt their structure, processes, and culture. Most organizations today are not yet able to truly benefit from faster software development. The prerequisite for this are the capabilities shaped by the agile and DevOps movements.

Agentic Software Engineering

Designing software architectures for AI and ML systems

iSAQB® Training