So I canswitchbetweenheadsandtailsand I canswitchthezerosandonesinmycomputersothatitrepresentswhat I wantedtorepresentlikean A andwithquantumcomputerswehavenewrules, wegettouseto, wecanactuallyspinoneofourquarters.
A lotofthosediscoveryquestionsaremuchhardertosolveusingthecomputerswehavetoday.
很多這樣的發現問題,用我們今天的計算機來解決要難得多。
Sothereasonwe'rebuildingthesekindsofcomputersisbecausewethinkthatmaybeonedaythey'regoingtodo a lotofreallyimportantthingslikehelpusunderstandnaturebetter.
Actuallythisiscalled a dilutionrefrigeratorandit's a refrigerator.
實際上,這被稱為稀釋冰箱,是一個冰箱。
Itdoesyoulooklike a normalrefrigerator, right?
它使你看起來像一個正常的冰箱,對嗎?
Butit's somethingthatweuseactuallythere's usually a casearoundittocoolourquantumchipsdowncoldenoughthatwecancreatesuperpositionsandwecanentanglecubitsandtheinformationisn't losttotheenvironment.
Frommyunderstandingwith a quantumcomputerratherthanusingtransistorsisusingspins, youcanhavesuperpositionofspins, sodifferentstatesmorecombinationsmeansmorememory.
Okay, soit's thisideathatyouhavetwoobjectsandwhenyouentanglethemtogethertheybecomeconnectedandthentheresortofpermanentlyconnectedtoeachotherandtheybehaveinwaysthataresortof a systemnow.
Weneedtousedeviceswheretheycantheycansupport a quantumstateandwecanmanipulatethatquantumstate, atomsionsandinourcasesuperconductingqubits, wemakecubitsoutofsuperconductingmaterials.
Butaslike a programmer, howwouldquantumcomputingaffect a differentwayofwriting a program?
但就像一個程序員一樣,量子計算會如何影響不同的程序編寫方式?
It's a perfectquestion.
這是個完美的問題。
I meanit's veryearlyforquantumcomputingbutwe'rebuildingassemblylanguages.
我的意思是,對於量子計算來說,現在還很早,但我們正在建立彙編語言。
We'rebuildinglayersofabstractionthataregonnagetyouto a pointas a programmerwhereyoucaninterchangeablybeprogrammingsomethingthewaythatyoualreadydoandthenmakecallsto a quantumcomputersothatyoucanbringitinwhenitmakessense.
Insteadofnowwritingout a giantsimulationovermanyterms, youtryandactuallymimicthesystem, you'retryingtosimulatedirectlyon a quantumcomputerwhichwecandoforchemistry.
Onething I havecolleaguesherethatyoucantakeuptwoweekstotraincertainneuralnetworksandactuallymachinelearningisoneresearchdirectionwherewe'rereallyhopingthatwe'regoingtofindkeypartsofthemachinelearningcomputationthatcanbespedupusingquantumcomputing.
But I guessit's notimmediatelyobvioushowyoudothatwithquantumcomputers.
但我想,你如何用量子計算機做到這一點並不明顯。
Ifyouthinkaboutevenjustclassicalinformationlikethisrightattheendofthedaywhenyoustore a bitinyourharddrive, there's a magneticdomainandyouhave a magneticpolarization, right?
Quantumsystemswerestillmanipulating a deviceandchangingthequantumstateofthatdevice.
量子系統仍然在操縱一個設備並改變該設備的量子狀態。
Youcanimagineifit's a spinthatyoucanhavespinupandspindown.
你可以想象,如果是一個旋轉,你可以有旋轉上升和旋轉下降。
Butyoucanalso, ifyouisolateitenough, youcanhave a superpositionofupanddown.
但你也可以,如果你把它隔離得足夠好,你可以有一個向上和向下的疊加。
Sowhatwedowhenwetrytosolveproblemswith a quantumcomputerisweencodepartoftheproblemwe'retryingtosolveinto a complexquantumstateandthenwemanipulatethatstatetodriveittowardswhatwilleventuallyrepresentthesolution.
Thisactuallyis a modeloftheinsideofoneofourquantumcomputers.
這實際上是我們一個量子計算機內部的模型。
Soyouneed a chipwithcubits.
所以你需要一個帶立方體的芯片。
Eachcubitis a carrierofquantuminformationandthewaywecontrolthestateofthatcube, itisusingmicrowavepulses.
每個立方體都是量子資訊的載體,而我們控制該立方體狀態的方式,就是使用微波脈衝。
Yousendthemallthewaydownthesecablesandwe'vecalibratedthesemicrowavepulsessothatweknowexactlythiskindofpulsewiththisfreethisdurationwillputthecubeitintosuperpositionorwillflipthestateofthecubitfrom 0 to 1.
Andthesealgorithmsweredesignedassumingthatyouhad a perfectfaulttolerantquantumcomputerwhichismanydecadesaway.
而這些算法的設計是假設你有一個完美的容錯量子計算機,而這是幾十年後的事情。
Sowe'recurrentlyin a phasewherewe'reexploringwhatcanwedowiththeseneartermquantumcomputersAndtheanswerisgoingtobewhatweneeddifferentkindsofalgorithmstoreallyevenexplorethatquestion.
There's thingsthatweknoworwehavestrongreasontobelievearegoingtobefastertodoon a quantumcomputerandthenthere's thingsthatwe'lldiscoverjustbyvirtueofhavingone.
I wasaskingcomputerscientistsrecentlywheretoreadaboutfaulttoleranceinclassicalcomputing.
我最近在問計算機科學家,在哪裡可以讀到關於經典計算的容錯性。
Hesaid, ohtheydon't teachthatincomputerscienceclassesanymorebecausethehardwarehasbecomesoreliablein a quantumsystemwhenyoulookatitormakemeasurementsititcanchangein a waythat's beyondyourcontrol.
Theredoesexistanalgorithmshore's algorithmwhichhasbeenprovenmathematicallythatifyouhad a largeenoughquantumcomputeryoucouldfindtheprimefactorsoflargenumbers.