Smart but not wise

ELC Prague 2026

The path to the AI-native engineering leader, in three layers. Wisdom is never found in multiplying matrices.

We spent a year getting faster at producing things and almost no time getting better at deciding which things were worth producing. That is the trap: AI makes everyone generically competent, and generic is not a position you can defend.

Three layers, because the problem shows up differently at each one. At the product layer, most pilots never scale, so the talk hands over the scorecard I use to decide whether something gets funded, gets two more weeks, or dies this afternoon. At the team layer, the question is whether anyone is still building judgment, which is what the impact plan, the team pulse and the judgment 1:1s are for. At the leadership layer it is simply that visions do not get written while you are busy.

It closes on a checklist you can start on Monday, including the one item everybody deletes first: four protected hours of being bored.