Flywheel is a kind of wheel in machinery that keeps the whole system in a cycle or rotation even without the need of stimulator. Flywheel epitomizes continuity…so what’s great deal about it in AI? According to my personal views, flywheel effect in AI is nothing but a better cycle of ML algorithms and datasets. It is a combination of both. It means more data,which means better output and product, which is further used by mass of people, resulting in generating data sets after data sets.
Garbage of data is generated by this effect. In short, this concept is known as GIGO (Garbage-in-Garbage Out). You cannot compete against flywheel system, once they are set in motion – they just keep going like eternity. For better understanding take the example of Google Search Engine. How will you develop a set of algorithms that can race ahead than the Google Search Engine? Practically, that is near to impossible because Google generates a lot of data when users search its engine for seeking queries. On the other hand, bloggers and social media and myriads of websites are being continuously crawled by it, thus, the process of creating datasets is always on. In fact, the process is continuous.
It’s a dark secret that giants like FB, Google, and Amazon are working out of big data and precisely focused on AI/ML/DL algorithms. Their datasets never exhaust up, it keeps adding on. In a sense, they are running huge factories of flywheels. Had Yahoo focused on big data, they could have been at a place where we now see Google. Now you got the idea of another compatible search engine is almost impossible.