I keep coming back to the centaur, half human and half machine, as the truest picture of how we work with AI.
The point was never replacement. It is amplification: the model as a thinking partner that pushes on my limits, as long as I don't hand it the reins. Lean on it like a GPS and you stop learning the road. The value shows up when I keep thinking critically and add what only a human notices to what the machine offers.
Between October and December 2025 I wrote down how that practice actually works, in seven chapters of Artificial Insights. I co-wrote the notes with the models themselves, mostly by asking them to look back over our exchanges and name the patterns.
- Thinking: refining ideas before they harden.
- Creating: bringing ideas into form.
- Building: making them work.
- Deciding: choosing a direction.
- Organizing: keeping it all coherent.
- Living: where it touches the ground.
- Connecting: how the fragments become a system.
Underneath all seven sit four small habits. Choose the moment worth using AI for. Feed it good context and clear intent. Review as a human, noticing what is missing and where it drifted. And keep questioning the choices it made.
The machine amplifies what we do. The sense and the direction stay ours.
