You're being overly pessimistic. Feynman's 'algorithm' isn't nothing, it's actually very rudimentary and important. The act of writing the problem down require

You're being overly pessimistic. Feynman's 'algorithm' isn't nothing, it's actually very rudimentary and important. The act of writing the problem down requires you to have a language for discussing it. You also have to be able to specify what you are looking for in a solution. (That is, how do you know if you've solved the problem?) Then you work to produce a set of transformations from one to the other. That is the best way to purposefully solve any problem. Just read between the lines a bit...