Solving problems II

My favourite way of solving problems; and I still find it hard though it is probably the most satisfying, is to deal with them from first principles. It involves interrogating the problem, and boiling it down to the most core assumptions and principles you can. And that allows new ideas and solutions to be revealed. The most important thing is that first principles relies more on ground-up reasoning than to try and ‘copy-paste’ solution (which by the way, will be the next blog post).

We can ask why is the problem presented a problem; and seek out the more fundamental bottleneck present. A lot of times, we can be solving a derived problem rather than the actual problem itself when we don’t think from first principles.

For example, when you have a job vacancy where there’s a particular role that you envision involving a lot of different tasks and from your understanding of the tasks the prospective staff is expected to perform, you derive various requirements such as 2-3 experience in this, and versed in particular software, having had particular paper qualifications and so on. Now you created for yourself a new problem: finding a candidate who will be a match for those requirements you just created. But are they truly the requirements for the job?

Thinking from first principles would probably involve working out a proper way of assessing candidates for the job rather than rely too much on those screening requirements. It will involve what the job really is about and what you want this role for. The ultimate solution may be surprising. It might not involve hiring at all! In some instances, it might involve parceling out the work to existing team members, automating them or using other tools to finish the work.

When thinking from first principles, you’re better at problem solving than others not because of your experience or having dealt with the same problem before. Rather, it is precisely your ability to understand and probe a problem more deeply than others ever did which helps.

2 Comments

Comments are closed.