Remix.run Logo
▲ xg15 4 hours ago

Then why do we need all the stuff about attestation and data bits?

▲danpalmer 4 hours ago | parent | next [-]

Because it prevents an attacker from stealing the password. It's both extremely hard to "steal" a passkey, and to phish a login from one. Neither of those are true for passwords, and TOTP only partially solves the stealing part.

Passkeys aren't solving for security nerds who like portability and have 30 character long passwords, they're solving for grandpa who will happily read their bank's 2FA SMS to the "bank" over the phone. The FIDO alliance has literally billions more of these users than they do users who know what hardware attestation is and how it impacts their auth workflows.

▲xg15 4 hours ago | parent [-]

This level of security is perfectly achievable with an unattested passkey.

And Grandpa will have fun if he has to juggle 20 passkeys with different security requirements on his devices.

▲danpalmer 4 hours ago | parent [-]

The security requirements are generally hidden from users. I have never noticed different behaviour between my passkeys.

But as I said, there are plenty of UX issues with passkeys, those UX issues however don't include getting your accounts hacked, phished, passwords lost in data breaches, etc, which I'd suggest are a lot worse.

▲WorldMaker 28 minutes ago | parent | prev | next [-]

We don't need attestation and most of the data bits. Apple consumer devices don't support attestation. Only Apple enterprise managed devices do. It's a paranoia tool for big enterprises with strict device management requirements. Apple has a lot of good reasons to believe attestation is a privacy violation for consumers and seems likely to keep up the good fight in making sure it remains an enterprise-only tool.

▲thadt 4 hours ago | parent | prev [-]

So companies can use passkeys and set policies around what hardware they're stored on. IMHO, not particularly useful or desirable for everyday Internet users, but really nice if you have specific requirements.