INNOQ Technology Night in Munich: Cloud-Migration & KI im Arbeitsalltag: Die produktive Einsamkeit 👉 11 Aug 18:00

Blog

The New Economics of Software Development

AI has dramatically reduced the time it takes to build software, but that doesn’t mean we should simply deliver more features. The productivity gains from coding agents give us an opportunity to rethink how we invest our development time. Instead of accelerating technical debt, we can use AI to build cleaner abstractions, improve maintainability, and leave our codebases better than we found them. This article explores how AI has changed the classic project management triangle, and why quality should become our new competitive advantage.

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.

Static Sites Without Big Tech Hosting

The Data Dilemma, Part 1

When AI agents fail in a company, it is rarely the model’s fault. The real problem is the data: semantics, quality, and context are written down nowhere, they live in the heads of the domain experts. Data Mesh addresses exactly this problem, but it organizationally overwhelms many mid-sized companies. This first part describes why that is and derives a Minimum Viable Data Mesh from it. The second part shows the implementation in four layers.

The Data Dilemma, Part 2

The first part showed why agents fail on undocumented data semantics and why Data Mesh is not a realistic answer for many mid-sized companies. This second part gets concrete: a Minimum Viable Data Mesh in four layers, from an honest inventory through curated views with ODCS contracts to agent access via MCP. A purchasing agent that compares suppliers and evaluates price histories serves as a running example.

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

Looking back on a year of Agentic Software Engineering training

Indecent Exposure

I noticed a trend in recent years where teams first create APIs for their Single Page Application (SPA) and then offer the same endpoints to other teams for systems integration. In my opinion that is a bad practice and you should think twice before you do that. In this blog post I will explain why.

What ECST Events, Domain Events, and ECST Really Are

“ECST Event”, “Domain Event”, “ECST”: these terms appear in almost every discussion about Event-Driven Architecture and Domain-Driven Design. Surprisingly often in practice, different people have a different understanding of them.

Trust but Sandbox

Running Coding Agents Safely with Docker Sandboxes

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.

Stop Using Bearer Tokens Like House Keys

You’ve built an API. You protected it with OAuth 2.0. You’re using JWTs. You feel secure. You’re not.

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.

OpenProject: A Real Alternative to Jira?

In his article “Using Jira Data Center with AI Features”, my colleague Nicolas Inden showed how to connect a self-hosted Jira instance with an AI model. One point stands out: Atlassian will only support Jira Data Center until early 2029. After that, support ends. Sales of Data Center licenses to new customers were already discontinued at the end of March 2026. So what do we do if we don’t want to return to Jira Cloud after 2029 - or want to move away from Jira right now?

The Agentic Trio

RAG and Fine-Tuning

Large Language Models (LLMs) don’t always produce reliable results. In practice, one thing becomes clear: combining Retrieval Augmented Generation (RAG) and fine-tuning can make them more robust. This post explains how the two approaches complement each other – and why a production feedback loop is critical.

Using the space, not optimizing the treadmill

Features per unit of time is not a meaningful metric. What counts is whether they satisfy a real user need and contribute to a desired outcome. Agentic development opens up the space to finally take that seriously.

The Missing Half of Your Data Strategy

Why Data Literacy Makes or Breaks Data Product Oriented Architectures

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.

Accidental Complexity in the Heart of Software

Domain-Driven Design went mainstream over the last ten years – in parallel with a wave of digitalization and modernization projects. That is not a coincidence. And it explains a lot about why so many of these projects fail to deliver what they promise.

Hail Mary: Why domain knowledge cannot be extracted from experts

Developing with AI through the cognitive Lens

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.

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.

AI Features for Jira Data Center – No Atlassian Cloud Required

Our Digital Independence Day recipe for going off-cloud

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.