I'm reading Walter Isaacson's latest book, The Innovators. In the second chapter, he introduces

*Entscheidungsproblem,*which translates to 'decision problem'.

David Hilbert, who posed the problem, asks whether there is an algorithm that can be used to tell if a given problem statement can be solved keeping in mind a set of axioms and logical rules.

Later, Alan Turing proved that some problems exist that cannot be solved with an algorithm. These are called undecidable problems as you will only know for sure that you can solve it once you have actually solved it. Those problems that can be solved using an algorithm are called decidable problems. For example, finding the average speed of an airplane flying from Bangalore to Mumbai is decidable while finding a new business idea that will earn profits is undecidable.

While you can spend hours and hours getting better at decidable tasks, you must know that it is all in vain because the moment it gets cheaper to get a machine to run an algorithm to do that task, you are replaced. Why, you are replaced even if another person who can do it just as fast and reliably for lesser money comes around.

Getting better at decidable tasks is like competing on price. The profit margins constantly drop and unlike the Flipkarts of the world, you can't scale out your work beyond 16 hours a day.

So, with this background, I must assume it would be a decidable problem if I asked you whether you want to get better at doing decidable tasks or undecidable tasks.

Short and crisp, nice read :)

ReplyDeleteIs this in any way similar to the P vs NP problem??Reminds me of the Minesweeper game. A good read, thanks! And the other write-ups too

ReplyDelete