And yeah, I personally use Bitwarden with 2FA on the Bitwarden account, but don't store any 2FA tokens in Bitwarden, handling them all separately. Don't want all the eggs in one basket.
Though given I have fingerprint access on phone and computer for bitwarden, i suppose that one basket is my finger. But if someone is taking my finger, I've got more immediate concerns than my passwords.
Assuming the victim does not use 2FA on their Bitwarden account
A pretty tall assumption given that we're already talking about someone who knows to turn on 2FA for other things. If someone knows about 2FA and password managers, they'd be insane not to have 2FA set up on the password manager itself.
It doesn't do very much, but the things it doesn't do aren't things I'm all that interested in anyway. What it does do, it's kinda hit and miss. But it is cheap as hell for a smartwatch and fully open source, so were I more inclined to code I could potentially improve some of the ways it does things.
Now I'm wondering if they trained the AI on the exact same machine that they used in their testing. Seems to me that tiny variances in the keyboards of even the exact same model could have significant effects on this sort of thing. And then there's different levels of wear from usage, crumbs in the keyboard, etc.
I'd be amazed if it is actually this accurate across all machines of the same model.