Unbuzzing AI and Machine Learning

I really dislike buzz words and the way they are leveraged by some people as a tool for self promotion. 

Lets face it, most companies are not truly working in Artificial Intelligence, at best they are implementing an interface on top of existing platform that has been developed by other companies (there is absolutely nothing wrong with this). At worse they are working on AI creating a platform from scratch, and as exciting as this sounds, unless significant fire power is thrown at the initiative, its not likely to end well. 

Innovation does not always come from reinventing the wheel, sometimes small iterations with existing frameworks and ideas can create a significant impact in whichever ecosystem you operate in.

At iCarAsia we decided to use TensorFlow as a machine learning platform interfaced by another tool developed by Google to be used as the front facing interface. Is there a risk of us putting all eggs in one basket and risking looking control over our tech? Maybe. In the grand scheme of things does it matter? No!

The reality is that the benefits of executing this and pushing the benefits to our customers in a shorter time span, coupled with the massive learning gathered from exposing the product to our customers completely outweighs the risks. 

An AI project is only cool if it actually does something useful, if it helps people by enhancing an experience. We did not look at AI as a technological problem, it was seen as a  potential way to enrich and improve the way car buyers and sellers connect.