# Innovation is Nonlinear

I am getting obsessed with the nonlinear way of discovering things.

For example, let's say you want to use machine learning to classify many strings, but you don't know anything about this topic. The linear approach would be to start learning about ML from the basics, and once you have enough knowledge, you will be able to solve your specific problem. By the time you reach this point, you find out that your problem domain is not a good fit for what you've learned (your knowledge is not practical for the problem domain), and thus the experience is disappointing.

The nonlinear way would go like this. You want to classify many strings using machine learning, but you don't know anything about this topic. You download brain.js and start doing some random problems not related to your problem domain which ultimately leads to a discovery or an insight in a different area. Now you innovate.

The illusion everyone in tech has is that it is possible to develop innovative solutions by following a linear path: A, B, C, D. But, innovation is never linear. If innovation was linear, it would have been repeatable as it will be possible to work forward and backwards with relative ease. Instead, we know that innovation is random, and it comes from unexpected places.