I'm amazed any HN readers could believe Google would want to know what is stored if they could actually read it. That is completely useless to them. What is interesting and useful is who is connecting from where, when, with what. Yet again it feels like Snowden is forgotten. Metadata is what matters in high volume data, not content.
So why does Google engage in automated reading of Gmail email contents?
they haven't done that (allegedly) in years right
Right…?
As far as I know they don't but for arguments sake let's say that they do. A lot of smart people have worked on Gmail so any scanning of mails would be for useful metadata or if they scan all the text it would be to provide services from context (find something with a date, ask if it should put it in your calendar, etc.). Almost all emails are worthless information and I doubt even the NSA with 100 times the budget it has today could learn anything worthwhile from scanning everything even if everything is only "everything at gmail". At best they would end up with data worse than they get from metadata and more likely the need for thousands of people looking at mails that match wordlists (hate USA, bombmaking, presidents route...) and not learn anything useful at all.