Tor traffic is monitored. Many high-bandwidth endpoints and relays are run by law enforcement, and Tor is configured to preferentially use high-bandwidth nodes. It is a given that, at minimum, all Tor metadata is logged; if not all content.
The fact that exit nodes are controlled by rogue third parties doesn't mean a thing if you use Tor to anonymise your connection. That's the purpose this is why it was built that way.
Tor anonymises extremely well both client and server. All other stuff in between should be used through SSL and with caution (after all SSL is kinda broken and the NSA can perform MiTM attack or steal and sign certificates from major authorities).
If you 'leak' metadata using JS or by logging to services e.g. to Facebook (the dumbest use of TOR I've ever seen), then obviously state-level authorities are not your concern. Otherwise you wouldn't real had public profiles on social networks anyway.
Comparing the neighbours Open WiFi to TOR is very naive, even if the NSA is after you. The Open WiFi will a give 1km radius of your CURRENT position on-the-fly. On the other hand to analyse metadata sniffed from an exit-node will probably take hours if not weeks and still would be a (mostly) wild guess over where you where at the time of the attack.
That's absolutely not true and if the author doesn't understand why, then he doesn't have a clue about what he is talking about.