AI
Practical applications of artificial intelligence in work, products, and business.
Digital solutions without illusions: what works and what remains a buzzword
Digitalization without illusions: what really works and what remains a buzzword. Facts, statistics and books.
Codex without magic: how I hand over a real project
Context, task boundaries, checkpoints, and the moment when it is better to take control back.
Where Codex fails — and why that is normal
Typical agent failures, the cost of weak context, and tasks I no longer delegate.
Claude Code or Codex: choosing a workflow, not a brand
One task, two agents, and concrete differences in workflow, control, and output.
A CLAUDE.md that actually helps
What to document for the agent — and what only bloats context.
When AI autocomplete starts getting in the way
Why writing code faster is not the same as shipping a product faster.
GitHub Copilot: where it actually helps
A separate AI product across the editor and GitHub: review, navigation, and routine edits without conflating products.
How to avoid generic AI design in v0
Context, references, and constraints that make the result feel like your product.
