Comment by ThePhysicist | Hacker News Reader

Comment by ThePhysicist | Hacker News Reader

I don't know, following that logic it would be impossible to claim copyright over an encrypted file as it's basically also just a very large random number, even if I can reassemble the original file if I have access to the cryptographic key. I think you could probably share the encrypted file on the web if you don't share the key, as no one can know what the file contains (assuming the cryptographic method is secure). Together with the key this data will become copyrightable though as there is an easy way for any user to re-assemble it on his/her computer.

So I'd say downloading individual randomized blocks is probably not problematic and akin to downloading an encrypted file without having access to the key. Downloading the URL that points to the descriptor list might be problematic though, as this will allow you to "decrypt" the other blocks.

I think the distinction this protocol makes is that by using XOR as the "encryption" method, given any input block you can choose a "key" to decrypt with to produce any other output block. A block in isolation provides zero information to the downloader. I think it could be argued that it is the knowledge of which blocks to combine is where the actual data is being stored, and maybe that's where the copyright owners could stake a claim.

So, the copyright owners will have a more difficulty job going after the users of the network. They have to prove that there is copyright infringement by monitoring the blocks downloaded as opposed to BitTorrent file hash identification.

So does that mean if I store child porn on AWS I don’t technically own it since it’s somewhere in the virtual cloud?

The only way I could access it is through that weird SSH key that doesn’t contain any videos in it.

Oh no - wait. Let’s take this idea even further.

I don’t actually store child porn - the videos I have are all decryption keys to files filled with zeros that I have encrypted with a one-time-pad cypher.

I don't think we are in any disagreement that whoever is uploading the data "owns" the data. The interesting idea is that the entity storing the bytes has 0 information about the data they represent, in the information theoretic sense, since they can decrypt the data to any value by choosing a sufficient key. This is not true for most other encryption schemes where the encrypted data has enough structure to it that theoretically it could be retrieved without the key, although the whole point of the encryption is that this isn't a practical undertaking.

Reply on news.ycombinator.com