are you really arguing that opus 5.5 beating the best of openai's models handily is worse than you expected? What exactly were you basing your expectations on?
I mean, why would you? Chinese models through an API are more expensive than any subsidized subscription and the only one offering a transparent subsidy (opencode) only gives you $15 of credit on any relevant models, and still has a 5 hour limit.
Actuallllly these hypothetical people from a few hundred years ago hypothetically said you are an awful person for misrepresenting them, stop putting words in their mouth, thank you!
Wouldn't I have the same problem with passwords? I would lose my phone and be unable to two factor log in to my Gmail account or bank from someone else's device?
With 2FA? Yes. 2FA is almost just as stupid and almost just as bad for regular users as passkeys are, but it got adopted due to historical loophole - by far the most popular form was TOTP delivered via SMS, then "authenticator" apps, both of which keep the code transferable - you can receive it on one device, transfer to yourself or someone else (which is a feature, not a bug) via any channel, and get it to work. Passkeys were created specifically to close that loophole.
Plus, SMS didn't allow for vendor lock-in, which is arguably why they're so hated in security circles (SIM-jacking is real, but doesn't scale anywhere near enough to warrant deprecating it as mechanism for regular users).
I have over 2,000 accounts. The majority of them have unique passwords. I can count the services that have passwords that I know on one hand, with quite a few spare fingers.
Passwords just don't provide a broad recoverability benefit once you commit to doing strong, unique passwords.
> Passkeys were created specifically to close that loophole.
Credential sharing? I have a family share with passwords and passkeys in it. Passkeys most certainly didn't stop this.
They did add friction to having a user being duped to share their password to someone claiming to be tech support, since you can no longer request plaintext secrets be sent over arbitrary channels.
> Plus, SMS didn't allow for vendor lock-in, which is arguably why they're so hated in security circles (SIM-jacking is real, but doesn't scale anywhere near enough to warrant deprecating it as mechanism for regular users).
SMS is an ugly user experience and more importantly is expensive. Now a lot of services do emailed codes when they don't have a regulatory reason to require SMS - an even worse user experience, but less expensive.
We have authenticator apps which use a standard OATH setup, and quite a few platforms which have integrated support to try to sand over the worst part of the UX. Unfortunately they just didn't become popular, and OATH fails the same regulatory requirements that emailed codes fail.
Passkeys are not meant to have a vendor lock-in story. Credential exchange allows for credentials stored in consumer credential managers to be imported into another credential managers. The platforms have added infrastructure specifically to make this easier for users.
The exceptions are when enterprises run their own passkey stores, or when the user explicitly picks a solution that doesn't allow export (like a physical Yubikey).
No, it's fine. Most common example, I log in to my bank, they send an SMS, I copy the code and paste it into the web site. Works fine and I don't have to pick up my stupid phone, like I would if I used an authenticator app instead.
"and more importantly is expensive."
Not for me. The bank made several billion dollars last quarter, don't think it's a big problem for them, either.
Most common example, I log in to my bank, they say they sent an SMS, I sit there waiting to access my account, then 20 minutes later the SMS finally actually gets delivered after I've given up and left to do something else. Rinse/repeat.
Or another example, someone hijacks my SMS'es, and then they log in as me.
SMS sucks. SMS is insecure.
The amount of times I've experienced customers complaining about SMS 2FA not working well despite it being on their carrier failing to deliver the messages in a timely fashion really showed me how terrible it is.
Dunno what to tell you. Several possible points of failure there. Maybe your bank's system for sending SMS just sucks. Maybe you're in a place with poor mobile network coverage (which is a real problem for SMS 2FA, I agree, having lived in a place with nearly none for several years).
But for me, where I live now, with my bank, and my phone carrier, it works great. Never had a problem that I can think of. Typically, the SMS arrives within seconds. It's the least annoying and most reliable 2FA method I use.
I'm not that worried about potential security issues. Sure, it's a possible problem, but pretty low on the list of things to worry about.
It's different, but it is not at all clear whether or not that matters for the purposes of whether or not they can experience things, seeing as we don't know how to measure that even for other humans other than by relying on self-reporting to infer without other evidence that their experience matches ours.
It's not a "trap", it's the real argument which makes me hesitant to dismiss consciousness of machines. Yes, it seems silly to dismiss a sick patient as having a lesser consciousness, which is exactly the point
It's not a real argument. Dementia patients can have lucid periods that plainly show their long-term memory isn't gone. They're mental state is obviously degraded, but its disingenuous to say they have no long-term memory and that they are any way similar to LLMs.
Yeah I was gonna add I know there are case studies where sick humans don't consciously have access to long term memories but will still respond nonverbally to people they "recognize".
Yes it's seemingly the go-to rebuttal, I've lost count of how many times I've seen it with barely any variation.
Not sure if it's a product of LessWrong or Slate Star Codex or something. I kinda suspect it might be LW, given how it superficially resembles an unanswerable philosophical paradox but in reality is about as challenging as a question posed in a 200 level philosophy course like introduction to ethics.
I used [AI Studio.](https://aistudio.google.com/) It's possible the same issues are present there, but AI Studio is intended for serious work. I don't see why they would intentionally hobble the model's capabilities in AI Studio.
Couldn't I simply give a Chinese friend my key on Open router?
reply