|
This paper called Fast Factoring Integers by SVP Algorithms is making the rounds that was given this statement in the precis, spelt "destroyes". Paper is fulla maths so it could be legit or just bullshit, I don't know. What I do know is what RSA is -- that's one of the most common cryptography algorithms that protects a bunch of things from being hosed with by bad guys, most notably all your secure internet connections. The 's' in https, that padlock, that's more often than not RSA. The forums aren't, they're elliptic curve, which is different, but tons of sites use RSA, including Google. The way it works is that you generate two massive prime numbers, which is pretty easy for computers to do, and then you multiply them together, which is also very easy for computers to do. You keep those primes private, that's the private key. Then you publish the product of those primes, that's the public key. When someone wants to send you a message, they turn the message into a big number (data is already numbers in computers so that's free), and just calculate the that number to the power of the secret key. Now the only way to get it is to know the private key, or figure out the private key by trying to factorise a massive huge number. This works because factorising is much much harder than multiplying, and these are huge numbers so that difference really matters. So this paper seems to claim they have a way for that not to be a thing, that would be uh, pretty bad, security wise. RSA is so old, it's basically everywhere, so there's a million places your identity, passwords, credit card, even your straight up money could get stolen. If someone intercepts your bank login, they can log in to your bank and transfer themselves all your money. This has been floated as a kind of Y2K style techno-apocalypse, quite a lot, but it was always like, "nah mathematicians are almost positive that it's impossible to factorise primes fast, at least without quantum computers, and they're years away, we have plenty of time." It would be pretty crazy if they were wrong.
|
# ? Mar 3, 2021 05:59 |
|
|
# ? Mar 29, 2024 13:34 |
I've read this post top to bottom like five times and have absolutely no idea what the hell it's about Like there's some mathematical formula that will kill cryptology or something? Am I close? The first sentence alone is
|
|
# ? Mar 3, 2021 06:01 |
|
low key sex master posted:I've read this post top to bottom like five times and have absolutely no idea what the hell it's about Yeah, maybe, or it could be like a timecube guy.
|
# ? Mar 3, 2021 06:03 |
|
Just imagine believing that RSA hasn't been severely compromised by the NSA for at least two decades when they went from gently caress RSA to RSA FOR EVERYONE, IT'S ALL COOL NOW! Everyone that has paid any attention to RSA at all knows it was obviously severely compromised and just a matter of time before the shoe dropped. edit: https://www.reuters.com/article/us-...EA2U0TY20140331
|
# ? Mar 3, 2021 06:11 |
|
I would simply use bigger numbers op
|
# ? Mar 3, 2021 06:15 |
|
I now understand how chuds feel when anyone who knows anything is talking about anything: Extreme inferiority complex and resentment. You are just believing in Big ... uh... Numbers!
|
# ? Mar 3, 2021 06:17 |
|
Comparing it to the y2k bug ain't doing this any favors
|
# ? Mar 3, 2021 06:19 |
|
Three Olives posted:Just imagine believing that RSA hasn't been severely compromised by the NSA for at least two decades when they went from gently caress RSA to RSA FOR EVERYONE, IT'S ALL COOL NOW! This is pretty terrifying poo poo. I heard about the NSA trying to trick or legislate dumb dumbs into using their terrible schemes with back doors but it always seemed comically inept and obvious. That they’d plant poo poo into security companies is pretty gross.
|
# ? Mar 3, 2021 06:19 |
|
This is why I only drink filtered rainwater and grain alcohol.
|
# ? Mar 3, 2021 06:27 |
|
this is iterative advancement on general number field sieve and i dont think the polytime argument goes thru because the poo poo hes claiming is iid is not iid
|
# ? Mar 3, 2021 06:56 |
|
First they'll have to destroy my rear end
|
# ? Mar 3, 2021 06:57 |
|
also the empirical results are comparing w straight face a result from 1993
|
# ? Mar 3, 2021 07:04 |
|
I'll believe it when this is corroborated by an expert writing in English
|
# ? Mar 3, 2021 07:04 |
|
But can I crunch numbers into a speculative asset
|
# ? Mar 3, 2021 07:19 |
|
so what, are they claiming they turned the np problem of rsa into a p problem with One Weird Trick, Cryptographers Hate It! or something
|
# ? Mar 3, 2021 07:36 |
|
Strawberry Pyramid posted:so what, are they claiming they turned the np problem of rsa into a p problem with One Weird Trick, Cryptographers Hate It! or something tl;dr just take what bob dobbs is dead is saying as true. That nerd has dug deeper into the subject than anyone sane.
|
# ? Mar 3, 2021 07:39 |
|
Strawberry Pyramid posted:so what, are they claiming they turned the np problem of rsa into a p problem with One Weird Trick, Cryptographers Hate It! or something From a skim, it looks like they found a way to efficiently solve the np problem in an efficient amount of time using matrices for most cases. Kind of like A* pathing.
|
# ? Mar 3, 2021 07:40 |
|
wake me up when someone breaks the algorithm that bitcoin uses and all the musky techbros end up penniless
|
# ? Mar 3, 2021 07:48 |
|
I added an exclamation mark to the end of all my passwords. I'm good.
|
# ? Mar 3, 2021 07:50 |
|
like, if you have a real beefy for an academic thing you can factor semiprimes on the order of 2^1000something straightforwardly. the nsa w its datacenter in utah can probably squeeze out a hundred or so extra bits of size. you need something like 2^3000ish to start breaking poo poo that peeps use for money. theyre demoing on 10^14
|
# ? Mar 3, 2021 07:55 |
|
Isn’t this the plot of the movie “Sneakers”?
|
# ? Mar 3, 2021 07:57 |
|
why is this dated 31.10.2019?
|
# ? Mar 3, 2021 07:59 |
|
we dunno the relation of np completeness w factoring. factoring would be whacked by quantum computers whereas we dunno if np complete would be whacked quantum computers will realistically take 10 figures and 3 decades of investment before a decent one becomes available to intelligence agencies. its not gonna be a general compute dealio, its gonna be shoved into application specific poo poo
|
# ? Mar 3, 2021 08:00 |
|
|
# ? Mar 3, 2021 08:00 |
|
If they broke RSA then they could easily demonstrate it by posting the private key of a random public key found online. Since they didn't it's at most an improvement over the existing best polynomial time algorithm. The funny thing about P=NP is that proving P=NP is as simple as providing just one solution to a big enough problem, but a disproof would probably span an entire book and give the discoverer a harem of math groupies in every university around the world.
|
# ? Mar 3, 2021 08:15 |
|
human garbage bag posted:If they broke RSA then they could easily demonstrate it by posting the private key of a random public key found online. Since they didn't it's at most an improvement over the existing best polynomial time algorithm. i like this reality you have constructed, where whoever published the article had to do it online out pure logistics. you know, because during live math conferences they were drowned out over the sound of being on sucked for hours on end by calcusluts
|
# ? Mar 3, 2021 08:25 |
|
That sounds super op but couldn’t anyone just hold down the 1 key on the keyboard and write to the same memory address until it breaks the prime number cryptography?
|
# ? Mar 3, 2021 08:32 |
|
ClamdestineBoyster posted:That sounds super op but couldn’t anyone just hold down the 1 key on the keyboard and write to the same memory address until it breaks the prime number cryptography? you dont even need to press down 1. i spilled my special widecap gatorade bottle on my keyboard and im pretty sure i broke all the prime cryptojiggies
|
# ? Mar 3, 2021 08:35 |
|
ClamdestineBoyster posted:That sounds super op but couldn’t anyone just hold down the 1 key on the keyboard and write to the same memory address until it breaks the prime number cryptography? hell yeah i'm in
|
# ? Mar 3, 2021 08:40 |
|
ScRoTo TuRbOtUrD posted:you dont even need to press down 1. i spilled my special widecap gatorade bottle on my keyboard and im pretty sure i broke all the prime cryptojiggies Yeah why do all the work? You could just as easily tape a pencil to an exercise weight and prop that up against an upside down coffee mug to hold the key down. Then you could do whatever you want.
|
# ? Mar 3, 2021 08:44 |
|
Destroyed OP's moms RSA cryptosystem (rear end)
|
# ? Mar 3, 2021 08:48 |
|
ClamdestineBoyster posted:Yeah why do all the work? You could just as easily tape a pencil to an exercise weight and prop that up against an upside down coffee mug to hold the key down. Then you could do whatever you want. can you please diagram this i'm not a cryptography expert
|
# ? Mar 3, 2021 08:56 |
|
r u ready to WALK posted:hell yeah i'm in I saw a video about how that is a valid Win 95 key. Really wild, crazy stuff.
|
# ? Mar 3, 2021 09:04 |
|
I work with cryptography and I had a hard time to parse that paper But thanks for it, it was an interesting read nonetheless. Agreed that asymmetric cryptography as it is used today is not good enough. I am trying to implement a “post-quantum cryptography” project at my job but it mostly falls to deaf ears since people still live with the assumption that asymmetric cryptography is in no way running the risk of getting cracked, and that’s infuriating as hell.
|
# ? Mar 3, 2021 09:06 |
|
the only thing that anybody sane will guarantee you will never be cracked is one time pad. good luck actually doin one time pad lol
|
# ? Mar 3, 2021 09:17 |
|
There isn't a government on earth that wouldn't kill us all for that thing.
|
# ? Mar 3, 2021 09:37 |
|
it used to be that the govt crypto agencies were far far ahead of commercial and academic everything but hey guess what the nsa pays the equation group 75k usd and google pays its interns, annualized, 90k usd just mysteriously, the commercial places got good peeps too
|
# ? Mar 3, 2021 09:42 |
|
Chef Boyardeez Nuts posted:There isn't a government on earth that wouldn't kill us all for that thing. drat, I watched Sneakers jus a week ago and it’s amazing how such an old film still holds up regarding the techbabble/social engineering aspect.
|
# ? Mar 3, 2021 09:43 |
|
Lol just lol if you don't use elliptic curves
|
# ? Mar 3, 2021 10:51 |
|
|
# ? Mar 29, 2024 13:34 |
|
bob dobbs is dead posted:it used to be that the govt crypto agencies were far far ahead of commercial and academic everything but hey guess what the nsa pays the equation group 75k usd and google pays its interns, annualized, 90k usd I gave my resume to the NSA when I was about to graduate. They were recruiting from a supercomputing convention for people knowledgeable in, you guessed it, the supercomputing field. We happened to be there as students running a very expensive server rack full of high grade CPUs and GPUs in a competition with other student teams from universities across the US, Germany, China, and other countries, where one of the challenges was cracking hashed and sometimes salted passwords. We did great but one of the Chinese teams had more GPUs than CPUs and took the competition. Anyway, the NSA's job listings are full of super boring Turbodork Analyst Level II and Computer Toucher Level I in Virginia or Utah (where the big spooky data center is) or some other boring city paying, as you say, peanuts compared to private tech companies. I ghosted the recruiter pretty hard because it was all very unappealing, and I didn't want to have to tell the government about every joint I'd ever smoked and list everybody I'd ever had a conversation with in the last 20 years for a background check.
|
# ? Mar 3, 2021 10:55 |