It's particularly bad for academic journal searches. Since those pages change so infrequently, this points to further shortcomings in the breadth of their search methods. Ostensibly, they are polling 'many' sources, but whatever their strategy, it's not working.
And then a few days later he off-handedly commented about the time I left for work that morning.... what a gross feeling.
You can't point to any single company and blame them for this mess, because the problem is in the interactions /between/ companies. There are hundreds of companies you've never heard of whose entire business is the gathering and merging of data, then trading it with others. This isn't a surveillance state, it's a surveillance market, and that's a much harder thing to track.
Sure, it would be good and valuable to make the Big Names less dangerous. But it's important to understand: if you just cut G, F, A & M out of your life, you won't be much safer. Most of the data about you is being gathered from sources with which you have no easily-manageable relationship.
Not that it excuses the behaviour of the others of course but Facebook is by far the worst offender in this list. They've been repeatedly and consistently caught misusing the data they collect. At least Google are only using it to sell better targeted advertising (for now).
Facebook is also the easiest to avoid.