Quantum computing.

[Asterix, the former Gaul.]

Back in the day when I was being taught about computers we were informed by a learned professor that our lovely new Personal Computers were actually inelegant and clunky.  The reason being that binary is agin nature and has to be forced.  Electric current doesn't actually go on and off neatly enough to be really efficient and neat for one thing.  All those years ago he had no idea how it would be managed but he though the next development would be more analogue, possibly through an organic medium.

Now it seems that he was dead right and we are heading for quantum computers that are like Schrödinger's cat and can be in at least two states at once.  This apparently enables quantum computers to run hugely (bigly?) faster than binary ones just like that and represents a paradigm shift just as Moore's Law was beginning to expire.

This could be good news for vinyl fans as analogue could be one day be back.  There may be other benefits.

[Kim]

Back in the day when I was being taught about computers we were informed by a learned professor that our lovely new Personal Computers were actually inelegant and clunky.  The reason being that binary is agin nature and has to be forced.  Electric current doesn't actually go on and off neatly enough to be really efficient and neat for one thing.  All those years ago he had no idea how it would be managed but he though the next development would be more analogue, possibly through an organic medium.

Analogue computing (think simulating stuff with the flow of buckets of water or electrons) is deeply unfashionable these days, but was used for all sorts of clever things before digital computing became ubiquitous.

 

Now it seems that he was dead right and we are heading for quantum computers that are like Schrödinger's cat and can be in at least two states at once.  This apparently enables quantum computers to run hugely (bigly?) faster than binary ones just like that and represents a paradigm shift just as Moore's Law was beginning to expire.

Faster or not, the real paradigm shift is that a quantum computer can (at least in theory) exploit *handwaves* spooky quantum weirdness to run certain more efficient algorithms that simply don't work on a classical computer (whether it's built from transistors, optics, DNA, paper tape, sprockets and steam, rooms full of people who are good at sums, specially trained ants, or whatever). 

You can also potentially exploit quantum entanglement for secure communication, which is related but not quite the same thing.

This could be good news for vinyl fans as analogue could be one day be back.  There may be other benefits.

Analogue never went away.  That's a false dichotomy perpetuated by lazy journalists and people selling record players.  The real world (including our senses) remains as analogue as it always was, and any non-trivial interface with a digital system (eg. the screen you're probably reading this on) involves some pretty nifty analogue engineering.

[Asterix, the former Gaul.]

Yeah, I think it was a different view of analogue though.

The real world (including our senses) remains as analogue as it always was, and any non-trivial interface with a digital system (eg. the screen you're probably reading this on) involves some pretty nifty analogue engineering.

Which is what I meant by this:

The reason being that binary is agin nature and has to be forced.

Faster or not, the real paradigm shift is that a quantum computer can (at least in theory) exploit *handwaves* spooky quantum weirdness to run certain more efficient algorithms that simply don't work on a classical computer

Yes:

This apparently enables quantum computers to run hugely (bigly?) faster than binary ones just like that and represents a paradigm shift just as Moore's Law was beginning to expire.

So you are telling me I am wrong but I am not wrong.  That's very Schrödinger's cat; I like it!

[Morat]

Quantum Computing has been the brave new future for quite some time now, rather like Nuclear Fusion. I hope to live to see both but I'm not holding my breath (because that would be counter productive).

[ian]

Interesting analogy, we can do quantum computing and we can do fusion, we just struggle to do them on a useful scale. That said, quantum computer comes with a smaller price tag than fusion (at least with the current mega-projects). There's still the problem of decoherence and noise, but error correction is improving. Unfortunately noise scales with the number of qubits.

There's also the problem of programming quantum computers – these are entirely different categories of algorithms that aren't easy to create in the first place.

Quantum computing isn't faster per se, it's just potentially better at solving a range of problems that are difficult or intractable to classical computing. Classical binary computers are pretty much like one of those games where you try to guess something through yes/no questions, it just speeds up how quickly those questions can be posed and answered (and of course you can run those questions in parallel). Quantum computing introduces a probabilistic maybe to the mix and given not every question has a yes/no answer, that's extremely powerful.

I think it'll happen, but I also think quantum and classical computing are complementary technologies, not a replacement.

[Morat]

Of course the NSA have already applied Quantum computing to breaking crypto, they're just not telling us!

[mrcharly-YHT]

Of course the NSA have already applied Quantum computing to breaking crypto, they're just not telling us!

Silly person, they use quantum computing for the CGI to generate the faked photos of the 'spherical' world. The 'globe' is flat! Don't be a sheep!

[ian]

Indeed, factoring large semiprimes is one of those things that's difficult to do on a Casio calculator. Even you've got a lot of secret agents and you give them all their own calculator. You wanted a gun, a wad of currency, and six different identities? Come on, look, it does scientific notation. Radians!

I was talking to a quantum computing bod a while back and he pointed out that some of the problems potentially soluble by quantum computing require algorithms that we have no idea how to create, and indeed our teeny little squishy human brains may not be capable. So they'll need to be written by other quantum computers.

Humans will be the acoustic couplers of this brave new world.

[Morat]

Of course the NSA have already applied Quantum computing to breaking crypto, they're just not telling us!

Silly person, they use quantum computing for the CGI to generate the faked photos of the 'spherical' world. The 'globe' is flat! Don't be a sheep!

So that's how they did it!! mind:blown

[David Martin]

I had an enjoyable read of John Gribbin's book on the development of the quantum computer. Quite fascinating - I now have a far more sophisticated level of ignorance than I had before.

[Asterix, the former Gaul.]

A friend bought over "Quantum: A Guide For The Perplexed" (he has a law degree).  The cat example proved unpopular so we replaced it with a lemonade example, i.e. you have a bottle of lemonade.  It might be flat or it might not. Opening the bottle is the only way to find out.  We might then have mixed it with something..

[ian]

Not quite, with the lemonade, its fizzical (sorry) state is just hidden. That's a vanilla (or rather lemon) mystery. There's no probabilistic indeterminacy. In the cat thought experiment, there's radioactive decay, a random process. You can't say precisely whether or not it has happened in a given period, so you don't know if a decay has cracked the vial of hydrocyanic acid or not, so you don't know if the cat is alive or dead. It's the indeterminacy that's key to what Schrödinger was saying. Since you can't know if the decay event has happened you can't know about the fate of the cat. The cat is both alive and dead. It's only through looking in the box that you resolve the superposition of feline fates.

Which is the mind bending thing about the quantum world: that something had only happened only because we look. For double weird, the cat obviously knows if it's alive, even if you don't. So by opening the box, you've created it's history. Alternatively the cat may be pondering your macroscopic indeterminacy, so you might only exist when it looks up and sees you looking back. Especially if you don't have tuna fish.

That's the Copenhagen interpretation, of course. Other interpretations are available.

[freeflow]

There's no uncertainty in Heisenberg's experiment. After a certain time the cat is always dead because it suffocated. Another way of breaking the experiment is to put a light in the box and make the box out of a perfect one way mirror. You can always see what is happening to the cat even though it can't see you.

[vorsprung]

Quantum Computing has been the brave new future for quite some time now, rather like Nuclear Fusion. I hope to live to see both but I'm not holding my breath (because that would be counter productive).

It is reckoned (by people that know) that by 2024 quantum computers will have enough grunt to make much existing cryptography completely useless.  It's a race against time to make new algorithms, test them and add them to open source crypto libraries like OpenSSL

[Morat]

Quantum Computing has been the brave new future for quite some time now, rather like Nuclear Fusion. I hope to live to see both but I'm not holding my breath (because that would be counter productive).

It is reckoned (by people that know) that by 2024 quantum computers will have enough grunt to make much existing cryptography completely useless.  It's a race against time to make new algorithms, test them and add them to open source crypto libraries like OpenSSL

Far be it for me to claim any authority in this field, it's all being researched by people with brains so much bigger then mine I couldn't even fathom the basics but my scepticism tends to ramp up in direct proportion to the proposed schedule.
<1yr - probably going to happen
1-2 years - getting a bit ahead of themselves when announcing The Next Big Thing
3-5 years - OK, you've got your grant now come back with some results
>5years - Science Fiction

[Mr Larrington]

See also sustained and controllable nuclear fusion, which has been twenty years away as long as I can remember.

[vorsprung]

Quantum Computing has been the brave new future for quite some time now, rather like Nuclear Fusion. I hope to live to see both but I'm not holding my breath (because that would be counter productive).

It is reckoned (by people that know) that by 2024 quantum computers will have enough grunt to make much existing cryptography completely useless.  It's a race against time to make new algorithms, test them and add them to open source crypto libraries like OpenSSL

Far be it for me to claim any authority in this field, it's all being researched by people with brains so much bigger then mine I couldn't even fathom the basics but my scepticism tends to ramp up in direct proportion to the proposed schedule.
<1yr - probably going to happen
1-2 years - getting a bit ahead of themselves when announcing The Next Big Thing
3-5 years - OK, you've got your grant now come back with some results
>5years - Science Fiction

generally I agree that 5years is sci fi

My source in this case is a world class researcher

As you may know there is already a quantum algorithm that will crack certain types of currently unbreakable encryption called Shor's Algorithm

I asked the researcher - how many qubits would it need to use Shor's algorithm vs a large (1k - 4k bits) key. I'd previously heard that it would be a square of the bit length (ie 1M - 16M qubits) but she said that even a smaller quantum device coupled with state of the art classical methods would be enough to crack the encryption.  Her estimate is that such a device will be available in the next few years

[ScumOfTheRoad]

Ian puts things very well regarding the nature of observation, and the Schrodingers Cat experiment.
When the cat is in the box, it is in a state of superimposed alive and dead wavefunctions.

The thing though that I cannot believe is the 'many worlds' interpretation. Sorry, but to me its a load of bollock(TM).

[ScumOfTheRoad]

ps. Sorry to move away from quantum computing, but regarding binary computing did not the Russians produce working ternary arithmetic devices?

[Kim]

ps. Sorry to move away from quantum computing, but regarding binary computing did not the Russians produce working ternary arithmetic devices?

I believe so.  AIUI the main advantage was they could represent negative numbers without resorting to sign bits or two's complement.

Many early computers used decimal arithmetic, albeit represented by groups of on/off logic states.  Binary-coded decimal (where you count in binary but stop after 9 rather than going all the way to F) is alive and well in mundane systems that deal with numeric displays.  Historically all sorts of wacky encoding schemes have been used - IIRC UNIVAC used an abacus-style representation, for example, and Babbage's mechanical computers used a true decimal representation.

[David Martin]

There's no uncertainty in Heisenberg's experiment. After a certain time the cat is always dead because it suffocated. Another way of breaking the experiment is to put a light in the box and make the box out of a perfect one way mirror. You can always see what is happening to the cat even though it can't see you.

The non-hypothetical experiment is the electron phasing through multiple polarising mirrors. That does my head in.

[ian]

Schrödinger deliberately chose the cat-in-a-box to make an example of how stupidly weird the concept was.

When you realise you can do the experiment with in actual things (admittedly, by and large a lot smaller and colder than a cat, but also less scratchy when you try and put them in a box) indeed does your head in. Even the simple double-slit experiment is fundamentally mind-boggling.

Of course, the universe doesn't have to intuitive to humans, so it might just be how things are. Equally, maybe there's something fundamental that we just haven't yet understood and one day someone will have a Eureka moment and everything will make sense again.

I have a hard time with the many-worlds too, I don't know where the universe(s) would conjure the infinities of everything required. But then I just have a layperson's understanding of all this.

[mrcharly-YHT]

There are multiple uses of encryption. When people talk of 'breaking' encryption, really they are only talking of decrypting a message.
'so what' I say.
Someone being able to read your messages is a bit of a bugger, but being able to read a message is not the same as being able to modify the message. That still requires the keys, which you don't get by decrypting the message.

[Greenbank]

As you say, there's more than one meaning for "breaking encryption" and it depends on context.

A government being able to take a public key and derive the private key would allow them to read all messages and modify at will.

What most governments want is not the ability to modify messages, but read all.

What most messaging apps that do end-to-end encryption (like WhatsApp) do is take your message on the device and encrypt it with the public keys of all the intended recipients and then transmit those encrypted messages. Only the people with the corresponding private keys can decrypt their message.

What governments want to do to "break" this type of encryption is add an extra 'government' recipient public key to all messages, so the Government gets an encrypted copy of every message that only it can decode. Of course you can trust the Government, etc. Assuming the people you correspond with are careful checking the public-keys of the people they correspond with there's no possibility of a man-in-the-middle attack to allow a third party to modify the messages being sent.

Of course, the public key you think belongs to Joe Bloggs might actually belong to someone else, and they decode your message and then re-encode it (possibly after modifying it) and send that on to Joe Bloggs. He then replies to what he thinks is you (but is actually a MITM) who then modifies/re-encodes and sends on to you. This is why you, of course, verify the keys you use when you meet people in real life. Don't you?

What Quantum Computing suggests it will be able to do (with sufficiently powerful quantum computers) is allow the derivation of the private key from a public key, which truly renders the encryption "broken" as it allows them to not just read all messages but modify too.

5 years for that? No chance.

[Kim]

They'll just put a backdoor in the devices, instead.