back

by matheusmoreira·14d ago·view on hn ↗
> Most analysts expect

Total bullshit.

There is no "effective" method without hardware remote attestation. If I control the system, I can just spoof whatever "verification" it is you're asking.

The whole point of hardware attestation is to put a cryptographic key in the computer that the users can't ever get at, then use that key to prove the computer booted a corporate owned operating system that's 100% aligned with government and capitalist surveillance and other cyberpunk dystopia nonsense.

Install a custom system that you control and they will say you have "tampered" with your device, and that transgression will get you ostracized from digital society.

This is what will happen, and if we let it happen might as well close down this site because everything the word hacker ever stood for will have been destroyed.

1 comments
You can of course create an independent attestation database at any time and mandate its use - verifying that the custom OS you use fits minimum security requirements for digital ID use.

We use that approach in several other industries.

But.... that requires work beyond just complaining.

>But.... that requires work beyond just complaining.

So you have to build an entire parallel internet just because you want to use Linux? That's what your argument boils down to.

The people who are complaining on HN are not platform operators, the platform operators don't care at all. To them it's not even about whether it requires work, they literally don't care.

For the people who care, it's not a matter of work, because they don't operate the platform.

> You can of course create an independent attestation database at any time

Ah yes. They're totally going to trust my self-signed certificates. They're totally not going to restrict their trust set to the corporate owned and surveillance friendly Google and Apple devices.

Come on now.

> minimum security requirements for digital ID use

Also known as "the user has no control over the device".

Because users who have control can simply spoof this silly "digital ID" and there's nothing anyone can do about it.

> We use that approach in several other industries.

Your industries include the user of the device in their threat models. They want the device secured against the user. Absolutely unacceptable.

> Ah yes. They're totally going to trust my self-signed certificates. They're totally not going to restrict their trust set to the corporate owned and surveillance friendly Google and Apple devices.

That sounds mostly like copium just to motivate your complete inaction.

Again - independent, EU based, attestation database is completely possible to make and we're using similar approval processes across multiple industries to certify hardware - locally, here in EU.

But yea, if you think you'll be able to print passport at home and then go travel and demand that government recognizes that as an ID document, you're a bit optimistic.

> we're using similar approval processes across multiple industries to certify hardware

Why not tell us more about the requirements for hardware certification?

Seriously doubt it's anything but the usual war on general purpose computing.

Requirement #1, the computer runs the mandated surveillance software.

Requirement #2, the computer does not allow the user to run any software not approved by the government.

Requirement #3, the computer resists tampering so as to preserve the previous requirements.

> Why not tell us more about the requirements for hardware certification?

Err, it's actually pretty simple: the token/certificate representing your ID (or credit card, or anything really) cannot be exfiltrated by userspace or installed kernel space apps or intercepted on the way to TPM when issued. And it cannot be duplicated.

It's the same set of requirements that are put on credit card smart chips and biometric chips in EU IDs and Passports (which are essentially also TPMs).

But sure, it's a all an evil conspiracy against general purpose computing. And they're all out to get ya. Now smash that downvote for a vote against the evil establishment.

>Err, it's actually pretty simple: the token/certificate representing your ID (or credit card, or anything really) cannot be exfiltrated by userspace or installed kernel space apps or intercepted on the way to TPM when issued. And it cannot be duplicated.

So you need a proprietary browser running on a proprietary OS (both userspace and the kernel) with proprietary TPM hardware. You just proved the point. No more Linux.

> cannot be exfiltrated by userspace or installed kernel space apps or intercepted on the way to TPM

So it must be secure against the user, as expected.

Preventing the user from "tampering" with the token means carving out a section of the machine and putting it out of his reach. You just created a government embassy on the user's machine. There's no telling what it will be abused for, and there's no escape.

> But sure, it's a all an evil conspiracy against general purpose computing.

You just advocated for putting an inescapable persisent cryptographic government ID on everybody's computers. This is the literal implementation of the surveillance state. Everything you do online, this token gets sent. It's the end of anonymity. Not even Tor gets around this.

Yes, that's what a TPM or smart card chip in your ID/Passport does.

Prevents tampering even by the "user".

I have advocated nothing of the sort you're accusing me of. Please leave your strawman at home.

Having a physical card fallback here is a necessity and nothing in these proposals shows that the physical card ID is going away.