4
$\begingroup$

My understanding of quantum computers is that because of their theoretical vastly superior processing speeds, brute-forcing your way through securely encrypted software would become feasible, therefore effectively nullifying existing encryption. Of course a remote hack with a quantum computer would still be limited to bandwidth etc. which would remove that advantage in some cases.

So my question is, if you had developed the first viable quantum computers and in your philanthropy you wished to gift the technology to the world, how would you go about safely doing this without leaving older systems no way to adapt and protect themselves and their users?

This question assumes that the developers of the computer are a private institution who are not funded by something like the military and are therefore not obligated to only pass the secret to their sponsors.

Edit The general gist I'm getting from the replies is that the release of quantum computers will be a slow, gradual process that will give people time to adapt and accommodate.

$\endgroup$
  • $\begingroup$ Bold assumption at the end. Real quantum computers are expensive, with all that cooling and stuff. It will not be realistic to say someone gave up so much money for free. $\endgroup$ – Mołot Jun 28 '17 at 10:50
  • 1
    $\begingroup$ You should have a look at Post Quantum Cryptography. People are already developing stuff to be "safe" from such attacks. The easiest way would probably be to announce it, attack someone to "prove" it and wait for people to adapt to the idea of quantum computers being a real thing. That will take some years and many people will not adapt. So at some point you probably have to just release it and hope for the best. The problem is that there are many, many aspects of human psychology you have to account for in this scenario. $\endgroup$ – Secespitus Jun 28 '17 at 10:51
  • $\begingroup$ @Mołot I think I'm allowed to make a fictional philanthropist as rich and generous as I like, and they may well have developed a way of reducing manufacturing costs during their internal development process. $\endgroup$ – Sparks Jun 28 '17 at 10:59
  • $\begingroup$ "How to prevent a cave man that invented an aircraft carrier from overtaking the world?" Let's forget the issues already stated here for a moment of them not being efficient right now, nobody will just invent something and skip 100 years of slow, painful progress in an instant. The moment quantum computers do what all the people that get research grands by overstating their abilities promise this will no longer be a big issue. $\endgroup$ – Raditz_35 Jun 28 '17 at 12:12
  • $\begingroup$ There is no such thing as "securely encrypted software"; you probably mean "encrypted data". While some encryption technologies are vulnerable to attacks by quantum computers, many are not. For example, AES is not vulnerable to quantum computers as we currently understand them. What's clearly in need of change are the key exchange mechanisms used by HTTPS or SSH; work is already in progress and we will have new standardized key exchange mechanisms long before quantum computers progress from their current status of laboratory toys to actually useful machines. $\endgroup$ – AlexP Jun 28 '17 at 12:59
7
$\begingroup$

First, your understanding of quantum computers is wrong. Quantum computers are not faster as such, it's just that they can solve certain problems much faster than classical computers by using certain features of quantum mechanics.

While quantum computers indeed have some advantage in brute-forcing (due to Grover's algorithm), that advantage is merely quadratic; that is, just double your key size, and you're safe against a quantum computer with the same processing speed as a classical computer.

When you hear about quantum computers threatening cryptographic systems, this is about Shor's algorithm which is exponentially faster than what (we think that) classical computers can achieve. Which is relevant for cryptography because our current asymmetric cryptography (like SSH and SSL/TLS) is mostly based on the difficulty of factoring large numbers. The algorithm being exponentially faster means that a doubling of your key length only adds a fixed amount of additional time to break the key.

However at the time quantum computers are released, this will be a non-issue, for two reasons:

First, there are already cryptographic algorithms that are believed to be secure against quantum computing (if you think "are believed to be" is too weak, remember that factorizing integers is also only believed to be hard for classical computers; we don't have a proof for it, and in principle we cannot exclude the possibility that tomorrow someone will publish an efficient algorithm to do it on a classical computer).

And second, asymmetric (public key) encryption is mainly used for communication; and for that quantum mechanics also offers a solution: Quantum encryption. Quantum encryption can be broken neither by classical nor by quantum computers (and unlike for public key encryption, you actually can prove it!). The only way to attack quantum encryption is to attack the hardware.

Note that quantum cryptography is already commercially available today, and surely by the time a quantum computer capable of breaking current cryptography is available, those quantum cryptography systems will already be in wide use.

So in short: The way to safely release quantum computers to the world will be easy: Just release them. People who care will already be prepared.

If you have indication that people are not sufficiently prepared, just announce that you will release them, wait for some reasonable time for everyone who cares to update their cryptography, and then release them,

$\endgroup$
6
$\begingroup$

I think you overestimate the immediate impact that quantum computers will have.

As @Mołot already noted in the comments, current quantum computers are extremely expensive. And as far as I'm aware, they're also far less powerful than conventional computers. It's cutting-edge technology, and it's still very much playing catch-up.

Realistically speaking, the "first viable quantum computer" will have performance comparable to an average or high-end conventional PC, but will cost a lot more. This isn't the kind of tech you can just give away for free - you might be the world's greatest philanthropist, but your investors aren't. They've pumped billions of dollars into your company to fund your research, and they will demand a return on that investment (i.e. profit).

The expense, and the limited benefits, will mean their impact will be limited. People won't instantly start cracking every encryption known to man. But people will be aware of the potential of quantum computers to do just that, once the tech improves, and they will prepare for that eventuality. In fact, they're already preparing. Chances are that by the time you have quantum computers capable of cracking our current encryption algorithms, those algorithms will already be obsolete (though there will still be some people using them).

In terms of "gifting your technology to the world", go down the Elon Musk route: give the patents away for free, so other companies can replicate your breakthroughs and improve on them. Once you have plenty of competitors working on quantum computers, the technology will improve much faster and the inevitable price wars will drive down the costs as well. This is what happened with home computers in the 80s, and it's what Musk is hoping will happen with battery technology.

$\endgroup$
  • $\begingroup$ SHA-1 is already obsolete, since years, and people are "cracking" SHA-1 using classical computers. en.wikipedia.org/wiki/SHA-1 and en.wikipedia.org/wiki/SHA-1#Attacks Keep in mind that SHA-1 was published in 1995, only some three years after MD5. By now SHA-2 is some 16 years old. $\endgroup$ – a CVn Jun 28 '17 at 13:21
  • $\begingroup$ I must have gotten my encryption algorithms mixed up, my bad. $\endgroup$ – F1Krazy Jun 28 '17 at 13:22
  • 1
    $\begingroup$ Neither SHA-1 nor SHA-3 is an encryption algorithm. It's a hash, which is a type of cryptographic primitive with properties very different from those you'd want in an encryption algorithm. For one, a good hash is irreversible, even for relatively trivial inputs. $\endgroup$ – a CVn Jun 28 '17 at 13:25
4
$\begingroup$

First of all, quantum computers do not have "vastly superior processing speeds". They don't work faster than traditional binary logic computers, they just work differently. That means that there are some problems they can solve much faster while they are worse at others.

One of the problems quantum computers (theoretically) excel at is finding the prime factors of large integers. One of the most widely used asymmetric encryption algorithms is RSA. This algorithm is fast to use but slow to break because multiplying two large prime numbers is fast, but finding these two prime-numbers when you only have the result is slow. That means unless you have a quantum computer which can execute Shor's algorithm.

But there are already alternative cryptographic algorithms where a quantum computer doesn't help much. So if someone comes up with a cheap and powerful quantum computer capable of breaking RSA, everyone would just need to switch to one of these algorithms. For more information about viable alternatives, check the article about post-quantum cryptography on Wikipedia.

That means the responsible release strategy for an RSA-breaking quantum computer would be:

  1. Announce that you have one and demonstrate that it can break RSA
  2. Tell people which algorithms they need to use to be save
  3. Give them enough time to switch.
  4. In the meantime, hope nobody steals your plans and beats you to market. Patenting it would be irresponsible, because a patent requires that you publish how it works. This would allow criminals, government agencies and other people who don't care about patent law to copy the technology. (plot-hook for a heist or spy story)
  5. Put it on the market
$\endgroup$
1
$\begingroup$

Tony Stark, don't worry about it at all.

There is a very big difference between standard computers and the quantum annealing process used for quantum computers. You can't just take the software from one and run it on the other - the hardware is so fundamentally different that many of the basic processes used in a standard computing environment DO NOT EXIST for quantum computers in our current milieu.

While there's been extensive theory crafting on the principals of HOW you'd use a quantum computer to decrypt data, that's no closer to practical implementation than a series of concept art is to a completed computer game on a gaming system THAT HASN'T BEEN BUILT YET. Crafting software that interacts with the necessary hardware in an efficient manner takes years, and as mentioned in the comments, there's been quite some progress on quantum encryption as well.

Though older, collected, encrypted data will ALWAYS be susceptible to cracking, that's not a problem anyone can solve (Enigma codes are trivial now). Older SYSTEMS can be re-encrypted. By the time decryption software becomes an issue, quantum encryption will also be solved.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.