Presumably, such a system would escalate to a more heavy weight authentication. We're already seeing something similar with sites trying to figure out if you're a bot or not. For example, if you make edits on stackoverflow, the site might decide to challenge you with a captcha from time to time.
The article doesn't make it sound like this would be a first line of defense, in fact the author seems quite adamant about its infallibility, but even if such algorithms would be used conservatively it would still be a huge hassle. Imagine every major website nudging you with a popup every once in a while: "we don't like the way you're typing, please use our code generator mobile app to prove it's still you".
I allege that this would have a huge number of false positives for no discernible reason to begin with, but on top of that: I might be on a different keyboard, using a different mouse, using a track pad, might be in bed lazily trying to log in with my laptop, might be injured, might be distracted, I might be in another country using an unfamiliar keyboard layout, a different screen size, the list goes on...
