>having users authenticate with a sample of their DNA.
I know it's not your point and I agree with everything you're saying in this thread 100%, but DNA based login would be a username, not a password. Like all biometrics. And it'd have a really bad false positive rate if you're just doing restriction enzyme cut location tests like lots of "police work". A full exome scan ~$300-500/per might do it.
Does that mean it's a great idea?
What about losing the yubikey?
In any case it's a different matter when it comes to high stake scenarios, in which case I don't mind something slightly more sophisticated.
But not as the default for everything.