I don't see the outrage here. Using heuristics to determine the likelihood of a URL being fake, sounds like a good idea as long as it's weighted against false positives.

That said, I've never understood why browsers do not highlight the hostname separately from the path. Many phishing domains are of the form: google.com.auth.something.else.realistic.looking.tk/fake-path-stuff and are so long that the user just sees google.com and moves on. Something as simple as underlining the hostname or making the path a slightly lighter hue would be a huge usability improvement in being able to stop phished hosts.