Discovery and Familiarity

Algorithms only look for averages.That is due to the absolute and never discussed fact that machines operate in averages.

When Spotify recommends you music is because it averages your taste and deems some songs to be  proximity ‘x’ to that average taste. Same with Netflix, Amazon and Google ads.

The brain (and subsequently humans) operate very differently.

We don’t calculate what we don’t need do. We simple do things, cutting through an array of signals, numbers and symbols (consider the mathematical complexity of catching a baseball).

We’re not rational, we’re behavioral.

We take longer to make spreadsheet calculations: analyzing insurance policy’s true risk score, or 1000 steps of permutation into a Go game.

However we normally know what to do, in an almost innate way. That is partly why trying to universalize the human expense into a general purpose machine is impossible.

For example consider talking to people from a different culture. 99.9% of intercultural conversations end up in a meaningful exchange of diverse points of view.

We don’t sit there and calculate the cultural distance of the other person, the anomaly from a singular median culture.

We’re just there, unaware of our inner making, and still able to pay attention to the right communication signals (body language) and ignoring the wrong ones (accent).

There is a deep fallacy in trying to universalize the human experience, condition and preferences (in its totality).

In design of tools we ought to encourage, not curb, the innate human experience.

Published by Nitzan

I am a designer, writer and strategist with interest in machine learning, liminal thinking and complexity science. In my commercial work I help companies build innovative tools, design better qualitative processes, and lead that human machine collaboration with complexity in mind.

Leave a comment

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.