News for hackers. Stuff that matters.
Re:UCLA Computer Grad Constructs “Crown Jewel of Cryptography” by tptacek on Wednesday May 24, 2023 @07:36PM (#36065307)
Isn't that where we're already at with ML classifiers?
Re:UCLA Computer Grad Constructs “Crown Jewel of Cryptography” by astrange on Wednesday May 24, 2023 @07:57PM (#36065495)
Well, we know they're ML classifiers and they're not making syscalls. So there's obviously a limit to "we don't know what they do".
Re:UCLA Computer Grad Constructs “Crown Jewel of Cryptography” by tptacek on Wednesday May 24, 2023 @09:39PM (#36066178)
You'd still know what syscalls a cryptographically obfuscated userland program was making.
Re:UCLA Computer Grad Constructs “Crown Jewel of Cryptography” by sbierwagen on Wednesday May 24, 2023 @10:53PM (#36066577)
If it's a total black box, wouldn't the NOBUS thing to do would be to have some large key that it's watching input for that flips it into a malicious mode?

If BB(6) took years to execute, how long would you have to spend feeding random input to a suspected-hostile 10000 symbol Turing machine (whose source code and state you can't examine) in a sandbox before you decided it was safe?

Re:UCLA Computer Grad Constructs “Crown Jewel of Cryptography” by rini17 on Thursday May 25, 2023 @02:37AM (#36067707)
Easy, have all executable segments read-only with none (or fixed) syscall instructions.
Re:UCLA Computer Grad Constructs “Crown Jewel of Cryptography” by astrange on Wednesday May 24, 2023 @11:50PM (#36066881)
Think it'd be a bit dangerous to let them make any though, unless you were single stepping.