Hacker Newsnew | past | comments | ask | show | jobs | submit | Retr0id's commentslogin

Ahhhhhh! Thank you for mentioning Precordial catch syndrome. I read about it some time ago, and experienced it several times, and then forgot what it was called. Even with the help of LLMs I was unable to re-discover the name of it.

This also lets them clamp down more on youtube-downloader tools. When they rely on the TV's own browser engine, the client fingerprints can vary a lot between devices and firmware updates. This makes "pretend to be a random smart TV" an attractive way of implementing a download tool.

If they control the engine, they control the fingerprint, making it easier for them to detect and harder for 3rd party clients to spoof.


The tobacco, alcohol, and gambling industries want a lot of things, but it doesn't make those things bad by default.

I'm more interested by the fact that it apparently didn't work this way before Tahoe.

IIRC there are options for exportable and un-exportable private keys when you make one in the secure enclave. Going to guess Tahoe made them un-exportable by default.

There is no way to export data out of secure enclave. You cannot write arbitrary data to secure enclave, as well as read it out.

What happen though is the ability to encrypt the data with secure enclave, store it on our own. When migrating, we decrypt with secure enclave, get back original data and re-encryp on the new device.


As far as I know, you have never been able to import or export keys from the Secure Enclave. It’s more likely that previously keys were stored in the keychain and now they are generated in the SE by default.

Yes you can make private keys that are encrypted by the secure enclave, rather than stored inside it. The "sc_auth" tool on macos has a "ssh-exportable" variant

> including the ~16 hypotheses I had to retract along the way

If you're going to paste raw claude output, please at least replace "I" with "Claude".


Hi, I wrote that note by hand; I didn't use Claude. The hypotheses were based on the fact that other emulations used different paths to reach the playback device. For example, I analyzed Rockbox to see how they managed to play music on this chip. I did use Claude and Gemini to investigate and create the emulator code

But that was the single hardest claim in the model register!

I, Claudius

Really sounds like: If you’re going to paste raw executable binaries, please at least replace “I” with “the compiler”

More like, don't commit compiled executables to git.

Right, because no one does that.

Hence the equivalence.

I think one of us missed the sarcasm. Lots of people post compiled binaries to git, so you can't really use that as a "you shouldn't forget to mention LLM use because you shouldn't submit compiled binaries" because lots of people do it in order to use it as a distribution channel rather than requiring users to compile it themselves or requiring the author(s) to host the binaries elsewhere.

Lots of people paste LLM output into human discussion forums.

Or "if you're going to submit a comment, credit Tim Berners-Lee".

> we sign a perceptual hash (pHash) of the image rather than an exact pixel checksum

Perceptual hashes are non-cryptographic. There are certainly collision attacks, but what about preimages? A preimage would completely break this scheme.

This paper demonstrates second-preimage attacks against PhotoDNA and PDQ: https://eprint.iacr.org/2021/1531.pdf


Yep, that breaks this scheme, making it useless. But it's far from the only thing making it useless, and the github page even explicitly lists those:

- "Small content edits slip under the threshold. [...] A localised edit covering ~15%x20% of the frame [...] passes as authentic" - this is the worst part. 15%x20% is huge, for example enough to change the face of the person or the book/text on the image.

- "Cropping is not survivable, at any amount" - given the purported reason for perceptual hashing is surviving light editing, it's pretty disappointing that one of the most common light editing operation is not supported.

Oh, and the whole "cryptographic chip" angle is absolutely bogus from the security perspective. OK, attacker can't extract the private key from chip. But they can simply connect the chip to a different device and have it sign anything! Given that the attacker in this model is device owner, this is absolutely trivial.


Considering child comment, I agree it's not great. That being said, would a bit-hash have worked? How many users are capable of sending bit-perfect images to someone else? I don't expect more than 5% of people to know...

Yes, it is a hard problem, but that doesn't mean that we should accept non-solutions.

Hi, I'm the author of the above. I also got a Sony A7 IV camera a few days ago. I already have a root shell on it. Although that wasn't even required to break Sony's C2PA implementation...

I'll write more about this in the future.


Simpler than that, you can just talk to the cryptography IC yourself and ask it to sign stuff. No need for an FPGA, just an arduino. Given the datasheet I imagine any LLM from the last year should be able to oneshot it.

if

I imagine on apple silicon this is buried deep in silicon / ISP IP block, and isn't a discrete IC.


But it is, the discrete IC is pictured in the article.

I think you're talking past each other.

You're talking about the device that's in the blog post; the person you're replying to is talking about the thing that Apple is shipping soon.


I could tell from the HN submission title + domain name.

I think this is the first time an "I made an agent swarm do something" blog post was actually written by a human, what a breath of fresh air.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: