Earlier writing
New articles now appear in my blog on storz.net. This archive collects links to earlier pieces published on Medium and Substack. They remain available in their original language.
Selected writing
- Frontier vs. Local LLMs: I Tested 8 Models on a 340-Page Book
- CoB prompting had no effect, whereas CoT prompting sometimes helped with small OS LLMs
- The secret to a fast AI assistant is caching (and parallelism)
- Steps Towards an AI-Augmented Organization
More writing on Medium
- Why I Don’t Like the New Antigravity
- How a 5-Minute AI-driven Background Check Saved Me $200
- GenAI in SMEs: Real Benefit or just Hype/Fad?
- Mapping People to Company Size from Their Email — A Mini-Data-Engineering Journey
- The Next Big Thing in AI: The Model Context Protocol
- Is it just me, or is Google Cloud (Run) Function way easier to handle than AWS Lambda?
- O1, Gemini, and Perplexity Deep Research in Comparison
- From Speech to Task: Building a Lightweight, LLM-powered, Voice-driven TODO App
- Market Research with LLM support
- AI in 2025 — Developments and Opportunities
- Using a small local LLM (llama 3.1 8B) to anonymize data for a remote big LLM (ChatGPT/Claude Sonnet)
- Import ChatGPT generated mindmaps into Miro
- Let opposing LLMs discuss for better results — improving AI output through personas with contrasting perspectives
- Clean Language Coaching with ChatGPT — explore yourself and gain insights