Slashdot Mirror
IBM Gives Everyone Access To Its Five-Qubit Quantum Computer (fortune.com)
An anonymous reader writes: IBM said on Wednesday that it's giving everyone access to one of its quantum computing processors, which can be used to crunch large amounts of data. Anyone can apply through IBM Research's website to test the processor, however, IBM will determine how much access people will have to the processor depending on their technology background -- specifically how knowledgeable they are about quantum technology. With the project being "broadly accessible," IBM hopes more people will be interested in the technology, said Jerry Chow, manager of IBM's experimental quantum computing group. Users can interact with the quantum processor through the Internet, even though the chip is stored at IBM's research center in Yorktown Heights, New York, in a complex refrigeration system that keeps the chip cooled near absolute zero.
They still make hardware?
And IBM will also give no one access to its Five-Qubit Quantum Computer?
Five qubits? According to understanding of quantum computing, which I got by reading two seconds of various Wired articles before the paywall kicked in, that means that this amazing machine is capable of storing every value between 0 and 31 simultaneously!
Users can interact with thing through the Internet, even though the thing is stored at location.
Wow! What an amazing application of the Internet!
WATSON = 5 letters
IBM quantum computer = 5 qubits
Coincidence? I think NOT!
What's with the wildly inaccurate and misleading titles on slashdot lately?
You Must Be New Here.
Quantum computing is next to useless if you're just computing elementary algebra because the answer is definitively known. As a matter of fact, I thought they still offload these kinds of operations to co-processors because they are still either incapable or dramatically inferior at doing them. On the other hand, if you're doing something more involved like mapping a function on a chart, then having the entire range of answers "instantly" available is far more useful. Imagine not having to worry about stack heap exhaustion, or recursion depth limit BS; you would have code that actually did what you tell it to without having to resort to hackish ad-hoc douche-baggary with a full paragraph of comments begging your successor for their forgiveness. Of course we're knocking our selves back to 1975 in terms of programming languages and destroying every advancement we've ever made in encryption by doing all of this, but quantum entanglement should make encryption obsolete anyway.
The ones I've seen any detail on behave like a finite state machine, except that the state retains all possible positions from the full range of inputs as modified to that point of the algorithm and a 50% chance of entering an error flow that results in an incorrect answer. Fortunately, the incorrect answer is not consistent, so if the algorithm does have a consistent correct answer it will be the most common answer after multiple executions.
Once your sequence of parallel finite state analyses gives you a strong assertion of a correct answer, it is relatively trivial to test that answer with even a naive classical algorithm.
In theory, a quantum-inclusive machine can solve many problems in 2O(n) time (as in, two separate O(n) sub-algorithms) that would take a classical system O(n*n). How perfectly it scales to reducing O(n!), O(x^n) and similar is beyond my current familiarity.