Recent Musings
What does it take to make data science, machine learning, and AI truly impactful in practice? Reflections, ideas, and insights on real-world challenges and current developments in a data-driven world.
Billions of Tokens for Nothing - The Cost of Context
September 29, 2026
Agentic setups expand the types of problems AI can help with. But it often also results in an exploding token usage...
Modern Programming Will Specify Behavior, Not Code.
July 21, 2026
Describing behavior will replace writing code when it comes to using AI to develop future applications...
Can we ever trust AI?
April 15, 2026
By now, AI models are used broadly. However, when it comes to trusting their outputs, we sometimes still run into a wall...
Bias in AI is bad – but is it really?
March 04, 2026
There is a lot of talk about bias in AI and how it results in bad and sometimes even harmful answers. But if you peel back a layer...
Agentic AI Orchestration – can Generalists Outperform Specialists?
May 21, 2025
AI assistants are evolving into cross-system agents. But can generalists outperform specialized tools in the long run?
Bootstrapping Agents - let them figure it out!
March 18, 2024
I have heard this a few too many times recently: “We can’t build an AI agent yet. Our knowledge base / data catalog / ... isn’t ready. These and other variations of this sentence stop many AI projects before they even begin...
Open Sourcing LLMs – is it really all about ethics?
March 08, 2024
Should GenAI be open source? That debate typically focuses on the risks of giving everybody with sufficient compute resources the ability to...