Syllabus Lesson 103 of 141 · Past the tenth idea
Past the tenth idea

The first ten ideas

Why it matters Ask anyone, or any model, for feature ideas and the first few are the ones everyone has. That is not a failure of imagination. It is where the search starts.

Ask for ideas for a team to-do app and the list writes itself: reminders, tags, priorities, due dates, a calendar view, sharing, dark mode, search, notifications, AI suggestions. Every competitor has them, because every brainstorm produces them first. A language model produces them first for the same reason: they are the likeliest ideas.

Seeing those ideas early makes it harder to get past them. In a study of designers, those who were shown AI-generated example images while generating ideas fixated more on the examples and produced less varied work (Wadinambiarachchi et al., CHI 2024). The first answer you see becomes the shape of the next ones.

Two moves that get you past them

  • Ban the obvious. Write the likely ideas down on purpose, then forbid them. What is left has to come from somewhere else.
  • Change the world the product lives in. Constraint cards: no screen, no network, one user becomes a team, the user comes back after a year. Each card makes the familiar ideas stop fitting.

Novel, not odd: a novel idea is unusual and solves a real problem better for someone. An odd idea is only unusual. The test is always the user, never the surprise.

Recap The first ten ideas are the common ones, for people and models. Write them down, ban them, then change the constraints. Keep only what helps a real user.

Spotted a problem in this lesson? Report it · Ask about this lesson

This exercise opens soon

Finish the lesson before it and this one unlocks. The whole course is free, no account needed, and you can read this lesson right now.

Start at lesson 1 Back to syllabus