[tejohnso]

No, it's

  1) Write down the problem
  2) Frame it in such a way as to pique Feynman's curiosity
  3) Wait for Feynman to tell you how he solved it

[MaxBarraclough]

These days we call this approach 'lazyweb'