Having been on the receiving end of over 50k compromised accounts attacking my server (with verified phone enabled); largely stolen oauth/access tokens from real users, I can also see why they would do that as a platform owner.
On the other hand, as a rms-thoughtpattern, this is awful.