Holy cow, maybe it's not faster at home. (I'm sure the fact that the "production" db is used more and well cached had something to do with this: )
Query OK, 789195 rows affected (1 min 56.18 sec)
Rows matched: 789359 Changed: 789195 Warnings: 0
HOWEVER, I actually had to reboot the stupid server (oh well, 441 days of uptime later...) because apache died on me while the sql statement was running and wouldn't start up again because 'something' was already listening on port 80.
So, my advice is that if you want to avoid the risk of a reboot, stop apache while you run the sql replace command.
Now to put on the .htpasswd file and I'll call it a day.