Originally posted by Wariat
any of you experience this sense of dread when you drink s lot or do drugs in a given period of time then cut it off completely for the day or so and during this day experience dread?
Yes, it's rebound anxiety. Booze calms you, and when you sober up the brain bounces back, but is used to the booze slowing it down, so your mind tends to go too quickly. It's very annoying.
Originally posted by Wariat
any of you experience this sense of dread when you drink s lot or do drugs in a given period of time then cut it off completely for the day or so and during this day experience dread?
Originally posted by Sweet
Factoring large semiprimes. How can it be so hard?
Tell the internet that it's literally impossible to factor the numbers you're trying to factor. It can't be done and you'd be a big dummy if you even tried.
Originally posted by mmQ
Tell the internet that it's literally impossible to factor the numbers you're trying to factor. It can't be done and you'd be a big dummy if you even tried.
They literally already did that, I'm part of that internet going "we'll see about that stoopid"
Currently the most advanced factoring algorithm is known as General Number Field Sieve and it's an extremely sophisticated algorithm that leverages several advanced concepts from number theory and linear algebra to factor large semiprimes.
The thing is. The second you have to start "guessing" or "searching" anything, shit gets horrible for large semiprimes. Lots of methods work for small semiprimes but if any step involves a search or guess, it's almost guaranteed it won't scale well to a large semiprime.
So I am really hoping there's some more direct way to just manipulate the integer. That would essentially "crack" factoring in full. But of course that's probably pretty hard and it's likely some simple algorithm or set of functions like I wanted doesn't exist.
it's pretty much the basis of asymmetrical (one-way) crypto, it's computationally easy to generate the number but much more difficult to take it apart
on the face of it the only way to find the primes is to loop backwards through the semiprime (or half of it anyway) and see if each combination of numbers works, but the goal is to devise an algorithm that is able to do it much quicker. from the look of the wiki they've only managed to improve speeds by using the latest hardware and optimising code for it; there's no special algorithm that'll skip a lot of failure.
you could save a lot of time if you precalculated prime numbers but these are 100-or so digits so looping through each one would take even longer than trying to bruteforce the semiprime as above
I am reading about something callef holographically reduced representations. They have some semi-miraculous mathematics behind them that allow you to solve some pretty "hard" properties of graphs in polynomial time, specifically Perfect Match.
I think you can use an advanced version of Dijkstra's algorithm on them to solve some very hard computational problems but it hasn't been well explored as an algorithmic technique yet. Holographic algorithms in general have not been well explored yet.
I'm not exactly sure what they could have to do with speeding up factoring yet but I feel like an "expansion" of a semiprime integer is required rather than a reduction of any search component of existing factoring algorithms.
But I'm not ruling out that possibility either.
I have a weird feeling like you can do the Area Method without needing to just search for Pythagorean triples. That's where the scaling goes to shit in Area Method, Fermat Method (search for difference of squares) etc at large integer lengths.
It's like, there is only one rectangle of integer side lengths that can have semiprime area. I feel like that's some kinda clue. There are a bunch of other geometric relations you can put the semiprime into the middle of. They all kinda evade giving you the unique length though. That might be a hint at the problem being fundamentally hard but that just seems so "accidental", no? So really there's not even so e relation in any number of dimensions where even by accident you can plug in the integer to 2 sets of manipulations and output 2 numbers that correspond to the 2 factors, for reasons that are otherwise totally unrelated to "solving factoring"? I mean you can write a function to make any integer output any other shit.
I dunno man, I got into cryptography for a while but was never good enough at math to get very far. I can understand the general concepts well enough but the actual mechanisms to get things done are a nightmare
if you're interested in that sort of thing, it'd be worth trying Project Euler
Tbh I honestly don't care about cryptography so much as I care about the basic math itself tbh. It just seems so arbitrary and stupid that factoring semiprimes should be so hard.
I think lots of amateurs dicking around with basic algebra and geometry and shit is useful. I think there is still a lot of juice left even in basic euclidian math because simply put we have never before din history had the sheer volume of math literate human beings walking around the Earth to the level we have now. E.g. Every guy who knows basic algebra would be one of the smartest people walking around ancient Greece.
There is the "infinity monkeys on infinity typewriters" principle but if you go from monkeys to humans and typewriters to the right basic tools, you can start putting more reasonable and useful bounds on what kind of realistic outcomes you can get and tbh maths is one of those things where the basic stuff we're taught is adjacent to an absolute monstrous amount of silent basic dependent logical structures that nobody ever really talks about or feels the need to explore.
Who knows what kind of interesting stuff is yet to be hit upon?
Eg if you make all the various rectangles of various integer areas and superimpose them in 3D in terms of their "extremeness" between being just a 1xN or Nx1, then do it for half I tegers, tenth integers etc you can start to make a weird 3D hypershape that seems to have some smooth and definite properties that change between each integer.
Their "waist" is always the square formed by the square root of the prime, but that's not always on the aforementioned "slice map". It seems useful to always map the square root anyway.
Semiprimes only have 3 rectangles with whole integer lengths, Nx1, 1xN and then the prime factors. Shit starts getting wild when you start mapping various non integer rectangles but I feel like there is some kind of hint there.
Originally posted by WellHung
Tara doesn't feel like being herself, yet, today… all the world's a stage, folks… all of us want to be actors and actresses…all of us want to escape from the monotonous, dreary day to day doldrums, folks. Even great white shark cage divers get sick of their job eventually!
