Start with a toy

When designing something new, start with a toy.

Toys are fun. Starting with one forces you to create something desirable from the start (see Desirability, Feasibility, Viability).
Toys invite people to play. Play is how we learn new things.
Toys don't have directed goals. When something is truly new, you don't know what it is good for. Use-cases don't exist yet. A toy lets you find your use-case through exploration.
Toys don't threaten incumbents. Instead of squashing you, an incumbent will ignore you. (First they ignore you, then they laugh at you, then they fight you, then you win.)

Power users find it difficult to accept toys. They are used to the way the axe handle fits in the palm. The tool has become an extension of self. GUIs will never be as powerful as command lines. Laptops will never be as powerful as desktops. Phones will never be as powerful as laptops...

Large organizations find it difficult to justify funding toys. A large organization is hungry for large profit, but a toy can't justify itself with a large bottom line. This pushes incumbents toward well-trodden paths with measurable bottom lines and away from exploration.

If it works, it's obsolete.

— Marshall Mcluhan

"Measurable" means someone else got there first and measured it. "User segment" means someone else got there first and segmented it.

If something is new, it has yet to be measured. It has no users. It also has no competition.

Start with a toy.