As someone who builds Machine Learning algorithm for a living in a large corporation, I concur with what she says. People have this idea that FB employs thousands of humans that sit there and decide what to show in your feed and how to bias it. There aren’t enough people in the world you can hire to do that manually. A machine learning algorithm is at the core of such technologies. The core objective is to increase engagement: likes, shares, reactions, comments, etc. So a ML algorithm can help you do just that to increase engagement. Also AB testing is a critical part to prove if engagement increased with a statistical significance given a new algorithm. Over time FB got excellent at collecting high quality data, building reliable data pipelines, and improving their ML algorithms. I love building and optimizing ML models but FB is a case where the upper management willingly ignored the known terrible side effects.
Great interview, once again!
It’s way past time we act surprised that it’s “all about money”
I gave up Facebook 6 years ago and don’t miss it at all. Awful company and even worse leadership.
Peter Thiel is out of FB board.
Hope to see you on Louder with American crowder cupcake!
Your email address will not be published.
© Late Night TV website by Super Blog Me