Yeah, I’m pretty skeptical of the premise… it’s looking for browser “abnormalities”? I mean… there wasn’t a strong motivation to correct those abnormalities for bots when it didn’t matter. Now that it does, I just suspect they’ll correct those abnormalities.
Just because the abnormalities were present in the past doesn’t imply that it’s intrinsically more difficult to emulate browser behaviour than it is to defeat captchas. There just hasn’t been a reason to do so up until now.