Conversation
Notices
-
Alex Gleason (alex@gleasonator.com)'s status on Monday, 16-Oct-2023 20:55:54 JST Alex Gleason Solve these impossible problems:
1. Create an encryption that can be decrypted after a specific date, without relying on any third-party.
2. Prove that a website is running a specific commit without accessing the server.
Either of these would change the world. But unfortunately they're not possible.- Machismo repeated this.
-
Tadano ❄️🎅 (tadano@amala.schwartzwelt.xyz)'s status on Monday, 16-Oct-2023 20:56:13 JST Tadano ❄️🎅 @alex Regarding the second, when you say accessing the server I assume accessing the server excludes visiting the site in a browser? -
Alex Gleason (alex@gleasonator.com)'s status on Monday, 16-Oct-2023 20:56:18 JST Alex Gleason @realcaseyrollins Signing is easy, it's verifying that isn't possible. You can give me a commit hash (which is comprised of all source code files plus commit history including authors and commit messages), but I can't confirm that the commit hash you gave me matches the code on your website. -
realcaseyrollins (realcaseyrollins@social.teci.world)'s status on Monday, 16-Oct-2023 20:56:19 JST realcaseyrollins @alex That first one looks tough. But on the second one, couldn’t you use a similar method that programs use to create signatures for programs? I think it has something to do with using the original source code as a hash or something? I forget how it all works.
Machismo repeated this. -
marlin (marlin@poa.st)'s status on Monday, 16-Oct-2023 20:56:30 JST marlin @alex The first one is implicit, you have to assume Moore’s law and select a cypher that will be brute forced near a selected future date.
-
Machismo (zerglingman@freespeechextremist.com)'s status on Monday, 16-Oct-2023 20:58:42 JST Machismo @alex 2. Take every possible permutation of outputs, create signature, perform the same on the server.
You won't actually prove it's running that code at all, but you'll prove that there's no practical difference. -
silverpill (silverpill@mitra.social)'s status on Tuesday, 17-Oct-2023 07:02:12 JST silverpill 1. Time-lock encryption
2. IPFSNot impossible.
-
Niclas Hedhman (niclas@angrytoday.com)'s status on Tuesday, 17-Oct-2023 18:01:14 JST Niclas Hedhman Interesting problems and isn't the
3. prove that the client is unmodified
in the same bucket, that the gaming industry is trying to solve, but after 20+ years of multi-player cheating still can't manage.
@9989500413fb756d8437912cc32be0730dbe1bfc6b5d2eef759e1456c239f905 @marlin
@teknomunk @Zerglingman -
Machismo (zerglingman@freespeechextremist.com)'s status on Tuesday, 17-Oct-2023 18:01:24 JST Machismo @niclas @teknomunk @marlin @9989500413fb756d8437912cc32be0730dbe1bfc6b5d2eef759e1456c239f905 @alex oh yeah that's a funny one, just don't trust the client ezpz