Username or EmailPassword
When a QC is actually assembled and working, let me know. Then we'll see when someone actually manages to program something on it. Until then, these kind of fluffy hype articles are silly.
Exactly. Weren't we meant to have all sorts of wonderful science fiction inventions by the year 2000?
Flying cars, anyone?
Even if one is built and actually works, I don't see *Personal* Computers as we know going away so soon... for 'super computers', it may change the game completely, but for personal computing it means much more the power, it needs to be affordable, common and also it needs personal computing software... looking at the past 20 years, PCs are just a natural and gradual evolution of the mainframe hardware and software, but QC would require a true revolution in software development too.
I honestly do not think that QC will come to be any time soon, leave alone applications that would run on it... And just out of curiosity - how fast do you want your word processor or e-mail client to be? Because these are the applications that current home PC users use the most...
Hopefully at the least it lets you write a letter.
I don't want to have to wade through three levels of photo realistic graphics just to get to the in-game typewriter so that I can write a letter and print it.
Well, the Quantum Computer will be so powerful that you won't need to type anything. It will guess automatically.
Actually, the Quantum Computer will be too advanced to 'accept' doing so silly things as typing a letter.
It may try sometimes to explain to mere humans the fate of the universe, loop quantum gravity, superstring theory, how many gods, what's inside girls' brain, how to win FIFA world cup. Things that matter.
Well, by that time perhaps we will have telepathy (or a near universal system of brain augmentation, with a personal cybernetic interface connection to a pervasive global network) as well as teleportation, so we won't need to write letters...;-)
QC has proven somewhat good at math but the technology is so young and untested, it will be a long, long time before it's on the desktop.
I hear the QC's will bring us the paperless office.
Definite *soon*. At least, I didn't see any quantum software for *personal*.
forget replacing the pc. That's the last thing on my mind if teleporting is possible
sorry, i didn't read the article. But the headline was just hilarious.
It's nice to say that a computer using a quantum computing core doesn't have to search for keywords, it can search the images themselves, but ummm, wouldn't we still have to enter the information into the images? I don't see any mention of a telepathic interface in the article.
It's nice to theorise but the truth is that, whatever is powering the computer, it won't be fast enough by then.
Yes, the article was disappointingly fluffy and vacuous. The Telegraph used to be regarded as one of the UK's `heavy' newspapers, before Conrad Black bought it.
As we're still struggling to develop ways of exploiting multiple cores properly*, developing algorithms for quantum machines is still some way off.
* --- well, I'll rephrase that. `as we're still addicted to pathetic 3GL high-level assembly languages rather than genuine programming languages that allow clear expression of algorithms'
...but it won't kill the Mac!
No it will not kill the Mac as Jobs will make a QC first!
Hides under something!
but that whole thing should have waited till the April 1st.
..and here is why. The size of quantum memory (number of qubits) doesn't scale so well as the size of normal memory.
In quantum computing, adding just a single quibit into existing memory is an engineering challenge. You have to find a new system with enough states that can be quantum entangled. Compare that with normal computers, where adding memory is not engineering challenge at all until you are going to double it.
So in quantum computers, each added bit of memory poses a problem. In normal computers, each doubling of memory poses a problem. So the memory in normal computers can scale exponentially while the quantum memory can scale only linearly.
This means if we were able to factor 1024-bit integers for cracking RSA on quantum computer, then everybody would start using 2048-bit integers, and they would be only very slightly more expensive. About the same order of more expensive as is factoring of 1025-bit integers on quantum computer. So you can see, this will not scale and is going to nowhere.
but the point is that adding a qbit to a QC doubles its memory. It can represent 2^n states simultaneously, so, increasing n by one, doubles the memory.
The point of QC is that classically there is a minimum power associated with the processing of 1bit. A quantum computation is reversible, meaning that it does not dissipate power.
In fact, you can classify codes by how much energy you would need to crack that code. You might not know how powerful a supercomputer your enemy can build, but you can state how much it would cost in energy, and put a price on that information.
I am completing my PhD in experimental QC, and I tell you, we are far far from building anything useful. We can simulate about 25qbits on a mac pro, and experimentally only do about 7. With the exponential difficulty in adding them, it will take some time before we catch up. But when power really starts to be an issue, QC will be there.
Sounds like he says something spectacular because he ran out of funds & wants someone to give him more money.
There are some swedish biologists that once every couple of years shouts "omgwtfbbq! Gullmarsfjorden is dying!" and they get funds for a few more years to do whatever it is they do.
quantum computers are practical today but this article and associated comments are full of hypes. Let me explain. I will tackle the problem (in bold), give a more accurate but far too concise statement (at the end, in italic) and I will try to explain it a bit more in the text that is found in between the bold and italic parts.
Typical personal computers calculate 64 bits of data at a time. A 64- qubit quantum computer would be about 18 billion billion times faster.
Untrue. This immediately follows from the idea of a superposition over all bits. But let me explain this a bit more clearly. It is true that a single quantum bit can be both true and false. However, another far more accessible way to look at this is that each option has a certain probability assigned to it. So, instead of saying that a bit is 0 with 100% certainty, you can say that a quantum bit is 79,42% a 0 and 21.58% certain a 1. So yes, a quantum bit can be both 0 and 1, but the problem is that you need to read the value. This is nothing like an ordinary computer. Reading a quantum bit actually destroys the superposition. This means that the qubit collapses to either 0 or 1 with a probability that was described above. It would be more accurate to state that a quantum qubit can somehow hold billions of billions times more information. This, however, is under the assumption that the quantum system is not disturbed and not measured.
Great, so I can hold a billion billion times more information, but reading it will collapse it to a single value, making it no more useful than an ordinary computer.
A working quantum computer could be so mind-bogglingly powerful that it would solve in seconds certain problems that would take the fastest supercomputer millions of years to complete.
True. But, well, only if you would clearly emphasize the could. In certain things, a quantum computer can be mind boggling fast. For example, if we search a deck of 2 cards for the card we want, we need to look at 1 card to be 100% certain which card it is. (It either is the card we look at, or it isn't. In the latter case, we know for sure that it is the other card we did not look at). A quantum computer can do better and need only look at a single card in a deck of 4 cards to be certain which card is the correct one.
Given a very concrete problem, a quantum computer can indeed be a lot faster. Overall, the speedup is not so mind boggling as this conclusion would suggest. But yes, a quantum computer can do things in seconds that would take an ordinary computer millions of years. Then again, a GPU can calculate pi a lot faster than an ordinary CPU. There is no more magic in it than that. It is better suited for a number of jobs, but grandly fails when it comes to some other jobs.
... But a QC could break the most complicated encryption in hours.
RSA is dead indeed. But we predicted this! RSA has an inherent weakness. Both you and the public know n, which is the product of two primes. RSA simply relies on the fact that, given n, it is really very hard to find out what the original primes were. However, there are oodles of special properties in this problem, making it only a matter of time before someone found a way to beat RSA. A quantum computer can indeed break RSA. Instead of thousands of years, it needs hours. But for this to work, we first need to make a quantum computer with a couple of thousands of qubits. The biggest we built that had full entanglement consisted out of 7 qubits. Ough. When we move to other encryption algorithms that do not rely on public and private keys, such as AES, a quantum computer fails to deliver to the promise. Agreed, a substantial speedup can be obtained. For example, a key of 256 bit can be cracked by a quantum computer about just as fast as a 128 bit key on an ordinary computer. So instead of 2^256 is is reduced to 2^128. The exponent is actually halved! However, the problem is still in NP. And making a bigger key is still P. Thus simply doubling the key size will stop a quantum computer from reading your most precious information.
RSA is dead when a big enough quantum computer can be built. This is an extreme challenge and not to be underestimated. However, alternatives exist and all other types of encryption remain quite safe. So, keep on using AES and the likes.
Quantum computers could also take advantage of another quantum property, teleportation. Teleportation allows information about one particle to be transmitted to another particle some distance away. A quantum computer could use teleportation instead of wires to move bits around inside itself.
Well, for this trick to work, you need to create an entangled qubit. This means that you actually have two qubits, but when you read one qubit, you immediately influence the other qubit. Some tricks, and some classical communication, make it possible to exactly reproduce the original qubit in some other place. This is quite cool, since normally a random qubit cannot be cloned. But starting from a basic entangled qubit and transmitting some simple information over a classical communication line makes it possible to destroy the original and obtain the new, exact, copy in the new location. But please, this is not Star Trek "beam me up" style teleportation. Researcher need their funds, and teleportation just sounds soooo cool that they immediately found the funds for fundamental, but far less spectacular research.
You can teleport things, given that you can communicate over a normal communication line and given that you already have an entangled qubit. This itself suggests that prior quantum communication needs to be possible before you can even consider teleportation. During the process, you need classical communication. So let's have some wires, okay?
I honestly do not think that QC will come to be any time soon, leave alone applications that would run on it.
A number of people stated this or similar ideas. But QC are here, today. Let us think again of a qubit as being both 0 and 1, but with some associated chance. Making a qubit that is 50% a 0 and 50% a 1 is actually very simple. Now, what would happen if we measure this? We get a truly random number generator! This and applications in quantum cryptography are hot in todays quantum world. Quantum cryptography makes it possible to detect when someone is eavesdropping and allows you to have an arbitrarily secure channel of communication. This makes it possible for example to share a common key that no one else knows, eliminating the need for RSA. So quantum computing will actually protect us from quantum computing power ;-).
Quantum computers exist today. They are great for generating truly random numbers and for some kick-ass cryptography. Go visit http://www.idquantique.com/ and order your random number generator USB stick today. It will outperform each and every random number generator that utilizes current day computers. Or pick us some quantum cryptography, but mind you that these are quite pricey!
Even if one is built and actually works, I don't see *Personal* Computers as we know going away so soon.
Current day classical computers are very speedy and they are very good at basic number crushing. A quantum computer is nothing magical and it cannot solve some random problem. It only consists out a few bits and can only outperform a classical computer given some very specific problems (mostly search problems). But classical computers are cheap, have tons of bits to spare and can do all the rough preliminary work!
Classical computers are here to stay, indeed.
I wrote a somewhat accessible paper this year on this subject for those that have sufficient computational and some mathematical background. This may clear up more things.
So basically, we can hope for a QPU with a RNG and Q(uantum)Sort for our next 128 cores silicon-based CPU but we shouldn't dream of running programs off it.
I'll put my bets on biochemical computers then. After all, we have a working model in our heads.
What is in your head can't crack a 32bit RSA key in hours, in days or in years.
Studying the brain is quite interesting for curing deseases, but I very sceptical when it comes to computer applications. If what was needed was a brain, we could as well use ours. If we have computer, it is to overcome our brain limitations. No human brain can break a 64bit RSA key in its life time. That is what computers are for.
Actually, I think it is feasible for a human to factor a 32-bit number, so you could crack 32-bit RSA key in your head.
Human brains might not be great at factoring huge numbers or doing rapid computations, but they are the *BOMB* at pattern recognition, and whether you realize it or not, it actually takes a great deal of "operations", so to speak, to accomplish this.
You seem to underestimate what your brain is doing at all times.
Our abstract thought abilities are a new feature implemented over many layers of cruft.
That you are able to detect movement, identify objects and calculate trajectories in realtime while keeping your body in the right position is already an amazing feat. But this is only the task on focus. At the same time, the brain is polling your tastebuds, your skin, your nose; everything for relevant input. A current supercomputing cluster cannot cope with all the information processing in a fly's brain.
In addition to that, the amount of useless information you store analogically without even realizing would set the Google servers in flames.
Silicon could possibly be faster, we don't have the benchmarks, but it surely doesn't scale well even if it beats QC at that. A 3D CPU with a few million cores would melt its way to the Earth's nucleum well before processing anything.
Great post! And your paper will be very useful to me. Thanks!
How do editors decide what to put into the news these days?
I'd see the point with a printed paper like:
Hey Bud we got to fill this page somehow - lets see Nessie was last week,bigfoot is on vacation and we can't have aliens abduct the president twice a week.
Heck - just write something about stuff that might happen in 20 years or so and give it a cool headline....
PC's going to be a very different in 20 years.Who would ever been able to think of that one?
Also just think about the possibilities - teleportation!
You can have those p2p criminals teleported right from their PC's to a RIAA torture camp in cyber space.
Of course microsoft's software is going to give them a couple of additional feet and legs on the head in the process.
Somebody post a bug on the LKML - we need a Linux kernel for this thing!
If you had read the article, you would know that it's not that sort of teleportation they mean.
The worlds oil supply will be expended by 1978, 1988, 1998, 2008 and 2016 respectively... We'll all have flying cars by 1976, 1986, 1996, 2006 and 2016... Bubble memory will replace static ram by 1986, by 1996 nobody will be using keyboards anymore, by 2006 the mouse will have been replaced, head per track hard drives will replace normal hard drives by 2006... By 2002 all internet commerce will be done in 'internet currancy' like beanz or flooz... Everyone will have smart appliances (powered by BeOS?) in their home by 2005... We'll all have nuclear powered vacuum cleaners by 1965... By 1930 all homes will have their power transmitted to them by DC current...
Yup, that subscription to popular mechanics is really paying off guys... File this with UFOs, astral projections, mental telepathy, ESP, clairvoyance, spirit photography, telekinetic movement, full trance mediums, the Loch Ness monster and the theory of Atlantis.
Man, I wish I hadnt already commented so I could mod you up. Stupid rules.
someone i met on http://www.fitnessloving.com told me that we should choose a good CPU of a good computer, is that true? by the way, they said people should go out regularly to avoid the harm of radiation from computer, right? besides, i find the members on that really good, maybe you'd like to find your lover and compose a happy family.