Writing
A weekly, no-nonsense take on building with AI. Real examples from real products. It goes out on LinkedIn first and lives here for good.
→ Follow on LinkedInAI without bullshit #3: What is a prompt, really?
Role, format, constraints and examples: what actually changes when you improve a prompt.
→ ReadAI without bullshit #2: Why AI lies about numbers
LLMs don't calculate. They predict what a calculation looks like. Use code for numbers, use the LLM for language.
→ ReadAI without bullshit #1: What is an LLM, really?
Training happened once. Everything you type triggers prediction. Why a confident answer is not a verified answer.
→ ReadAI without bullshit
Why I write this every week, and what you can expect: practical, honest, no hype.
→ Read