Prompt: If I sample social media posts based on my priors instead of evidence how will the prior probability grow?
Summary
Having a weak prior is very useful when building beliefs based on data (Bayesian Inference - Importance of weak priors). However, because we are information gathering machines, we automatically update our beliefs every time we gather new information.
One of the problems with social media is that it tries to show you content that matches your beliefs. This increases engagement with the content and increases advertisement revenue for the social media companies.
However, if you are showed posts/videos based on your beliefs, your internal bayesian inference updating algorithm will assume that your beliefs are consistent with data and it will get stronger and stronger.
Hence, just being on social media will make someone polarized if even they start of with a weak prior.
Experiment
To do: use the globe tossing experiment and sample data from the population based on prior instead. See how the beliefs converge and test if your assumptions are consistent with the simulation.