How I think
Make the next thing obvious.
A product should not ask a person to learn its entire internal logic before it helps them.
My product thesis is simple: people should not have to absorb the whole system at once. The job is to decide what they need first, what can wait, what should stay out of the way, and when a person, not software, should make the call.
What this looks like in practice.
Start with the uncomfortable decision.
Most of the work is hidden in the moment someone hesitates. That is where language, sequence, and a small amount of empathy do more than another feature.
Earn trust in stages.
I do not believe people should hand over control to an automated system all at once. The system should begin small, show its work, and earn more responsibility through useful behavior.
Use an editorial hierarchy.
Put the important idea first. Make the next action clear. Let the curious person go deeper without punishing everyone else with a wall of information.
Leave a human-sized exit.
When a situation is unusual, emotional, or consequential, the right answer is often to make it easy to reach a person. Good automation knows its edge.
What I am not interested in.
Busy interfaces that hide the decision. Intelligence that performs certainty. Growth that depends on someone misunderstanding what they agreed to. I would rather make one honest path that works than five clever paths that make people feel behind.
Outside of work, I am usually writing, rebuilding a workflow, or trying to make a desk more comfortable for long days of thinking and typing.