What building 3 dead AI projects taught me about the 4th one

✦ Project 1: Too broad. Tried to help "everyone who uses AI."

✦ Project 2: Too clever. Solved a problem people didn't know they had.

✦ Project 3: Too early. The infrastructure wasn't ready.

⤹ Project 4 (the current one) started with one constraint: find one person with a specific painful problem and solve it completely for them.

I found them. Built for them.

Start smaller than feels comfortable.