VITERBI
Scope an idea you can demo in 48 hours
You do not need a company. You need one person, one bad moment, and one thing your build lets them do.
Pick a user you can describe in a sentence. A dispatcher at the port. A neighbor during a blackout. A student trying to reach family when the cell network drops. If you cannot picture them, the idea is still too big.
Then pick the moment it fails for them. Not the whole crisis. The minute where they do not know what to do next. Your demo is that minute, fixed.
Write three lines before you open a laptop:
1. Who they are. 2. What just went wrong. 3. The one action your build lets them take.
Everything else is a later version. A map of the whole city, a second user type, an account system, a landing page: cut them. If it does not appear in the three-minute demo, it does not get built this weekend.
Build the demo path first. A button, a result, a sentence a stranger understands. Pretty screens come after that path works. If you are stuck at hour six, delete a feature and make the remaining one obvious.
Before Saturday night, read the three lines out loud to someone who is not on your team. If they ask "but what does it do?", you are still describing a topic, not a build.
A useful test: could you demo it with the wifi off, in under a minute, to a friend who skipped the explanation? If yes, the scope is small enough. If the demo needs a paragraph of setup, cut the setup. Judges have four questions, and none of them is "how big was the idea." They ask whether it works, whether it could ship, who is better off, and whether it helps Los Angeles on a normal day and in a crisis.
Write the out-of-scope list on paper and leave it on the table. When a new idea shows up at 1 AM, point at the list. You can build that one next month.