Latest Update:
Everyone Talks About Agents. Nobody Talks About State.
Over the past year, the discussion in AI has gradually shifted away from models as isolated reasoning engines and toward agents as autonomous operational systems. Large language models are no…
Most recent updates
Why Chasing the Next Big Thing Is a Career Trap
Some days ago in the evening, after I finished speaking at a meetup in Dubai, two young guys waited until most people had left the room. They were not interested in debating Kafka internals or LLM benchmarks. They asked something much more personal, and much more relevant. Where should we put our effort? What will be the next big thing?Should we scale vertical or horizontal? It is a question almost everyone asks at some point. It sounds strategic. It sounds ambitious. It sounds like you are planning ahead. But the…
OpenClaw Is Not the Autonomy Revolution You Think It Is
When you scroll through social media today, you might come away believing that OpenClaw has ushered in a new era of autonomous AI assistants that you can drop straight into production and have them “just work.” That impression is misleading. OpenClaw, formerly known as Clawdbot and Moltbot, is a clever and technically interesting side project created by a single developer for other developers. Its rise to GitHub trending status says as much about narrative dynamics as it does about engineering merit. OpenClaw is not a turn-key automation platform for enterprise…
Running Code AI Locally: An Engineering Reality Check
Over the last couple of days, my LinkedIn feed has been flooded with euphoric posts about “Code AI” and “local coding assistants”. Screenshots of terminals, bold claims about productivity exploding, and the familiar undertone that if you are not running an LLM locally via Ollama, OpenCode, or Copilot, you are already falling behind. I know that not only engineers read my blog. A fair number of managers and tech leads do as well, often just to get a polite rant about trends and what is actually happening behind the screenshots….
Teaching a Machine to Recognize Traveling Bears
This project did not start as an attempt to build a generic image recognition system or to benchmark computer vision frameworks. It started with three teddy bears that have been traveling with me since 2017. Over the years, they have accompanied me on flights, through airports, into hotel rooms, conference venues, cafés, and occasionally onto beaches. Airline by airline, the family slowly grew. Each bear developed its own character, its own role, and eventually its own name. They are not famous on Social Media, and they were never meant to…
Vibe Coding: Why It Feels Productive and Why It Fails Engineering
There is a growing belief that software engineering has become an optional skill and a 20-dollar subscription with the right prompts can build complex systems without understanding architecture, versioning, security, or operational reality. Engineers, according to this narrative, are a bottleneck that can be removed. I am skeptical of claims like these, but I do not consider dismissal without evidence a serious position. Instead of arguing in the abstract, I tested the approach myself under real constraints, using a realistic technology stack and enough complexity to move beyond toy examples….
Teaching a Machine to Clean Up My Document Chaos
This „project“ did not start with the ambition to build a generic document classifier or to compete with existing document management systems. It started with a much more personal and probably familiar situation. I wanted to explore whether machine learning could help me to organize my PDFs better. Not reminding me of deadlines or summarizing documents, but assisting with a very concrete task: taking a new PDF and proposing the folder where it belongs, based on how similar documents were filed in the past. And for safety reasons, the system…