The best problem solvers I know don't actually solve problems.
They redefine them until they're easy.
Spent the weekend reading "Are Your Lights On?" by Weinberg & Gause.
It's a problem-solving book from the 80s.
Everything in it applies to building AI products today.
// FROM THE BOOK
"We can never be sure we have a correct definition. Even after the problem is solved."
This broke my brain a bit.
Then I realized it's liberating.
// REAL EXAMPLE
Client comes to me: "Our brand is inconsistent."
Old me would:
• Audit everything
• Build massive design system
• Spend months "fixing" it
Now I ask:
• WHO perceives inconsistency?
• WHAT does inconsistent mean to them?
• WHEN does it matter?
Often discover:
• Sales team needs deck templates (solved in 1 day)
• Product team needs component library (solved in 1 week)
• "Inconsistency" = different people solving different needs
// THE TRAP
Book calls it "taking their solution method for a problem definition."
Client: "We need a brand guidelines document."
Me (old): "Okay, here's 50 pages."
Me (now): "What problem does the document solve?"
Often discover they don't actually know.
// THE SHIFT
Stop being a problem solver.
Become a problem definer.
// THE 3X RULE
"If you can't think of at least 3 things that might be wrong with your understanding of the problem, you don't understand the problem."
I use this constantly.
Can't think of 3 reasons a user would click that button?
You don't understand the user yet.
Can't think of 3 ways your "solution" creates new problems?
You don't understand the system yet.
// PRACTICAL
Next time you're stuck:
List 3 ways you're wrong about:
• The problem
• The solution
• The user
• The constraints
• The value
Not 1. 3.
First 2 are obvious. Third breaks the pattern.
That's where insight lives.