We rarely explain the algorithms when we talk to potential Viva users.
There are a few reasons for this, and it is not because the algorithms are not fascinating. I think the algorithms behind Viva are some of the most fascinating parts of the whole experience so far – how the algorithms identify the footprints of human behaviour and link them together into a chain of social preferences.
I spent half an hour tonight with the CEO of a UK blockchain company talking about how Viva’s algorithms learn a user’s preferences and use those to suggest events to the user (I wish I can tell you who the CEO is, but I am trying not to cross my day and night jobs – the photon streams should never cross).
As I speak with more people about these algorithms, what I am beginning to realise is that there are not many people who have developed anything similar for practical use. What I do not know at the moment is whether that is because my idea is so brilliant that no one else has thought of it, so crazy stupid that most people in the know instantly dismiss the idea, or simply because I have not read enough into the topic. If history is any indicator, then it would be the last of the alternatives.
The CEO suggested that I should look into Google’s machine learning software TensorFlow before heading into versions 2.0 and above of Viva.
This is all theoretical, until we see Viva working to give the best suggestions to its users.