To truly solve a problem means to create a solution that can be perfectly reproduced multiple times by anyone anywhere.
This paper proposes three different ways of categorizing problems:
- Problems of simple circumstances - those that involve few variables
- Problems of unorganized complexity - those that involve multiple individually predictable moving components
- Problems of organized complexity - those that involve multiple individually unpredictable moving components
Warren says that by creating tools to solve the third type of problems we would find great improvement in humanity as most health, social and political problems are of the third type.
He also foresees that computers and multi-disciplinary methods of working are the right direction into trying to create tools to solve this third type of problem.
After 70 years, it's clear that computer simulations are essential for understanding complex systems. The development of artificial minds would bring new tools and ideas to the table so that we can aim at solving biological, social and political problems. Although, with great power comes great responsibility. Being able to predict or prototype complex systems would become a terrible tool in bad hands. I guess that this will always be the case as it has been ever since language was created.
In terms of my own writing, I'm realizing that my goal is not to create tools to solve abstract theoretical complex problems. My goal is to create a way of creating artificial minds so that we can augment our use of personal computers. So my publications are not intended to propose truths or reproducible solutions to established problems. My intention with them is to share instructions on how to build artificial minds. Any other relevant finding that may be worth putting into a formal scientific paper would also be published, but that's not my goal.