Quantum computers

[KhillA]

https://news.vice.com/article/why-wall- ... vicenewsfb" onclick="window.open(this.href);return false;

Whose been keeping up with this? I work in the RF (mmWave) field and mostly concern myself in frequencies from 50-1700GHz. We are making great advancements in the analog world but the digital has to come with it. As communication speeds skyrockets (I've done a lot of work to facilitate 5G development; mostly in 55-95GHz) we need the computing power to match it. I believe this is it. I know very little about digital processing but my BS focus was in semiconductor fab so I believe in the core technology here.

Thoughts?

[KhillA]

The first bitcoin miners that get their hands on this are going to make a quick fortune and then crash the whole Bitcoin system, imo

[ChaMerZ]

I only know of the ramifications in the computational complexity field once these things come out in the wild. The standard Turing Machine will fail to be an adequate model for examining P = NP problems. P will effectively be equal to NP at least in practice. The P vs NP problem as originally stated will remain unsolved but nobody will care aside from idealistic grad students and post-docs trying to advance their academic careers. All the papers currently being pumped out in CS labs will be relegated to the nethermost reaches of the web.

I'm not a hardware guy. Didn't know you are an engineer btw ... cool.

The first bitcoin miners that get their hands on this are going to make a quick fortune and then crash the whole Bitcoin system, imo

Brute force will stop being expensive. No need for dynamic programming, you can write the most garbage code you can think of (or not think of haha), because well, it won't matter. Quantum computing will change the game for real.

[ric]

there's a past thread about this same subject but I'm on my phone so searching and linking will be huge hassle.

[KhillA]

Good post Chamerz. I'm not gonna pretend to know much about P=NP theory bc I didn't even know much about it until you posted it which sent me thru several hours of reading through wiki (it's a fucking rabbit hole). I didn't know about most of the unsolved computer science problems in fact, which is where I spent the most time on. Or that they have a million dollar bounty on the P=NP problem.

I do disagree that the P=NP problem will remained unsolved unless that by saying "as originally stated" you are referring to something that I'm not aware of.

@ric three questions:
1. Why do you never quote but instead use the @ symbol?
2. What phone are you using that makes searching and linking a huge hassle?
3. Could you provide a link when you get to a computer?

[ric]

@khilla
quoting takes a lot of space and effort. I browse with my phone mostly nowadays because my computer blows and I wasn't able to get the new one I wanted due to financial concerns. anyway it's just that clicking links etc is a bitch with this phone (htc)

past article

http://www.philaflava.com/forum/viewtop ... m#p2579478" onclick="window.open(this.href);return false;

[ChaMerZ]

Good post Chamerz. I'm not gonna pretend to know much about P=NP theory bc I didn't even know much about it until you posted it which sent me thru several hours of reading through wiki (it's a fucking rabbit hole). I didn't know about most of the unsolved computer science problems in fact, which is where I spent the most time on. Or that they have a million dollar bounty on the P=NP problem.

I do disagree that the P=NP problem will remained unsolved unless that by saying "as originally stated" you are referring to something that I'm not aware of.

@ric three questions:
1. Why do you never quote but instead use the @ symbol?
2. What phone are you using that makes searching and linking a huge hassle?
3. Could you provide a link when you get to a computer?

P = NP is pretty straightforward once you do away with the jargon that academics like to use to obscure what they are talking about.

P: class of problems that we know can be solved easily with computers (finding the greatest common divisor of two numbers)
NP: class of problems that we know can't be solved easily with computers (finding cliques in a graph, trying to brute force your password)

& if you show P = NP, not only will you be a rich man ($1 million prize), you will have effectively shattered the theoretical underpinnings of the public key security systems we use regularly. Nobody will be safe

[907]

I liked them better when they were called SoleSides computers.

[KhillA]

Good post Chamerz. I'm not gonna pretend to know much about P=NP theory bc I didn't even know much about it until you posted it which sent me thru several hours of reading through wiki (it's a fucking rabbit hole). I didn't know about most of the unsolved computer science problems in fact, which is where I spent the most time on. Or that they have a million dollar bounty on the P=NP problem.

I do disagree that the P=NP problem will remained unsolved unless that by saying "as originally stated" you are referring to something that I'm not aware of.

@ric three questions:
1. Why do you never quote but instead use the @ symbol?
2. What phone are you using that makes searching and linking a huge hassle?
3. Could you provide a link when you get to a computer?

P = NP is pretty straightforward once you do away with the jargon that academics like to use to obscure what they are talking about.

P: class of problems that we know can be solved easily with computers (finding the greatest common divisor of two numbers)
NP: class of problems that we know can't be solved easily with computers (finding cliques in a graph, trying to brute force your password)

& if you show P = NP, not only will you be a rich man ($1 million prize), you will have effectively shattered the theoretical underpinnings of the public key security systems we use regularly. Nobody will be safe

How do you define "easily" in this context? You gave examples which help but I would think it had to be explicitly quantified.

[ChaMerZ]

Good post Chamerz. I'm not gonna pretend to know much about P=NP theory bc I didn't even know much about it until you posted it which sent me thru several hours of reading through wiki (it's a fucking rabbit hole). I didn't know about most of the unsolved computer science problems in fact, which is where I spent the most time on. Or that they have a million dollar bounty on the P=NP problem.

I do disagree that the P=NP problem will remained unsolved unless that by saying "as originally stated" you are referring to something that I'm not aware of.

@ric three questions:
1. Why do you never quote but instead use the @ symbol?
2. What phone are you using that makes searching and linking a huge hassle?
3. Could you provide a link when you get to a computer?

P = NP is pretty straightforward once you do away with the jargon that academics like to use to obscure what they are talking about.

P: class of problems that we know can be solved easily with computers (finding the greatest common divisor of two numbers)
NP: class of problems that we know can't be solved easily with computers (finding cliques in a graph, trying to brute force your password)

& if you show P = NP, not only will you be a rich man ($1 million prize), you will have effectively shattered the theoretical underpinnings of the public key security systems we use regularly. Nobody will be safe

How do you define "easily" in this context? You gave examples which help but I would think it had to be explicitly quantified.

Sorry my dude just saw this.

Now this is where it gets complicated. Computer scientists have precise definitions of what "easy" is and what "hard" is but it is quite abstract and so you have to watch the terminology. In CS, "easy" generally refers to anything that takes "polynomial time" to run. An example would be a program that takes X amount of time to find the gcd of two numbers A and B, and 2 times X amount of time to find the gcd of 2 times A and 2 times B and 3 times X amount of time ... I hope you get the idea. It takes a long time but compared to the size of the input, the performance scales in a way that isn't remarkably bad.

"Hard" to computer scientists is anything that takes "exponential" time, so stuff like brute-forcing passwords falls into this category. Let's say you want to brute force Tommy Bunz password and you know it consists of the standard 26 letter alphabet. Brute forcing a one letter password will require 26 steps (because at worst case you will have to check every single letter). If it is a two letter password, it is going to take 26 squared steps (I hope that makes sense). If its three letters it will take 26 cubed steps. For a password of size n it will take 26^n steps. Now compare this performance to polynomial time. This is really, really shitty performance. Really bad. For a sufficiently long password the universe would end before you have a chance to crack it, assuming you have the worst possible luck and have to try every single combination.

Anyways, that's just a general idea. Its all hand-wavey too (& quite possibly false, since i might be butchering the exact definition of P and NP here), so if there's any other mathematicians, cs ppl here, don't crucify me.

If you want to have a firmer grasp on these things unfortunately you are going to have to furrow your brow, and remember your math fundamentals and work it all out from first principles. That's the only way you are really going to "get" it. But luckily the "high level" ideas are pretty easy to get across.

[360]

I haven't kept up with it for a bit. How many qbits are we up to now?

Last I checked it was 1? Maybe 2?
I'm hoping within 10 years we can have something we can actually call a quantum computer.

It's one of the glass ceilings.


Still, I can't wait to get into vr using a quantum computer. Then reality truly becomes subjective.