Placeholder Image

字幕表 動画を再生する

  • Quantum computers use the natural world to produce machines

  • with staggeringly powerful processing potential.

  • I think it's gonna be the most important computing

  • technology of this century, which we are really just about

  • one fifth into.

  • We could use quantum computers to simulate molecules, to

  • build new drugs and new materials and to solve problems

  • plaguing physicists for decades.

  • Wall Street could use them to optimize portfolios, simulate

  • economic forecasts and for complex risk analysis.

  • Quantum computing could also help scientists speed up

  • discoveries in adjacent fields like machine learning and

  • artificial intelligence.

  • Amazon, Google, IBM and Microsoft, plus a host of smaller

  • companies such as Rigetti and D-Wave, are all betting big

  • on Quantum. If you were a billionaire, how many of your

  • billion would you give over for an extra 10 years of life?

  • There are some simply astonishing financial opportunities

  • in quantum computing. This is why there's so much interest.

  • Even though it's so far down the road.

  • But nothing is ever a sure thing.

  • And dealing with the quirky nature of quantum physics

  • creates some big hurdles for this nascent technology.

  • From the very beginning, it was understood that building a

  • useful quantum computer was going to be a staggeringly hard

  • engineering problem if it was even possible at all.

  • And there were even distinguished physicists in the 90s who

  • said this will never work.

  • Is Quantum truly the next big thing in computing, or is it

  • destined to become something more like nuclear fusion?

  • Destined to always be the technology of the future, never

  • the present. In October 2019, Google made a big

  • announcement. Google said it had achieved quantum

  • supremacy. That's the moment when quantum computers can

  • beat out the world's most powerful supercomputers for

  • certain tasks.

  • They have demonstrated with a quantum computer that it can

  • perform a computation in seconds.

  • What would take the world's fastest supercomputer?

  • Years, thousands of years to do that same calculation.

  • And in the field, this is known as quantum supremacy and

  • it's a really important milestone.

  • Google used a 53 qubit processor named Sycamore to complete

  • the computation, a completely arbitrary mathematical

  • problem with no real world application.

  • The Google Quantum computer spit out an answer in about 200

  • seconds. It would have taken the world's fastest computer

  • around 10000 years to come up with a solution, according to

  • Google scientists.

  • With that, Google claimed it had won the race to quantum

  • supremacy. But IBM had an issue with the findings.

  • Yes, IBM, the storied tech company that helped usher in

  • giant mainframes and personal computing.

  • It's a major player in quantum computing.

  • IBM said one of its massive supercomputer networks, this

  • one at the Oak Ridge National Laboratories in Tennessee,

  • could simulate a quantum computer and theoretically solve

  • the same problem in a matter of days, not the 10000 years

  • that Google had claimed. Either way, it was a huge

  • milestone for quantum computers, and Silicon Valley is

  • taking notice. Venture capital investors are pouring

  • hundreds of millions of dollars into quantum computing

  • startups, even though practical applications are years or

  • even decades away by 2019.

  • Private investors have backed at least 52 quantum

  • technology companies around the world since 2012, according

  • to an analysis by nature.

  • Many of them were spun out of research teams at

  • universities in 2017 and 2018.

  • Companies received at least $450 million in private funding

  • more than four times the funding from the previous two

  • years. That's nowhere near the amount of funding going into

  • a field like artificial intelligence.

  • About $9.3

  • billion with a venture capital money poured into AI firms

  • in 2018. But the growth in quantum computing funding is

  • happening quickly for an industry without a real

  • application. Yet it is not easy to figure out how to

  • actually use a quantum computer to do something useful.

  • So nature gives you this very, very bizarre hammer in the

  • form of these this interference effect among all of these

  • amplitudes. Right.

  • And it's up to us as quantum computer scientists to figure

  • out what nails that hammer can hit.

  • That's leading to some backlash against the hype and

  • concern that quantum computing could soon become a bubble

  • and then dry up just as fast if progress stalls.

  • Quantum computers are also notoriously fickle.

  • They need tightly controlled environments to operate in.

  • Changes in nearby temperatures and electromagnetic waves

  • can cause them to mess up.

  • And then there's the temperature of the quantum chips

  • themselves. They need to be kept at temperatures colder

  • than interstellar space, close to absolute zero.

  • One of the central tenets of quantum physics is called

  • superposition. That means a subatomic particle like an

  • electron can exist in two different states at the same

  • time. It was and still is super hard for normal computers

  • to simulate quantum mechanics because of superposition.

  • No, it was only in the early eighties that a few

  • physicists, such as Richard Feynman had the amazing

  • suggestion that if nature is giving us that computational

  • lemon, well, why not make it into lemonade?

  • You've probably heard or read this explanation of how a

  • quantum computer works.

  • Regular or classical computers run on bits.

  • Bits can either be a 1 or a zero.

  • Quantum computers, on the other hand, run on quantum bits

  • or cubits. Cubits can be either 1 or zero or both or a

  • combination of the two at the same time.

  • That's not wrong per say, but it only scratches the

  • surface. According to Scott Aaronson, who teaches computer

  • science and quantum computing at the University of Texas in

  • Austin. We asked him to explain how quantum computing

  • actually works. Well, let me start with this.

  • You never hear your weather forecaster say we know there's

  • a negative 30 percent chance of rain tomorrow.

  • Right. That would just be non-sense, right?

  • Did the chance of something happening, as always, between 0

  • percent and 100 percent.

  • But now quantum mechanics is based on numbers called

  • amplitudes. Amplitudes can be positive or negative.

  • In fact, they can even be complex numbers involving the

  • square root of negative one.

  • So so a qubit is a bit that has an amplitude for being zero

  • and another amplitude for being one.

  • The goal for quantum computers is to make sure the

  • amplitudes leading to wrong answers cancel each other out.

  • And it scientists reading the output of the quantum

  • computers are left with amplitudes leading to the right

  • answer of whatever problem they're trying to solve.

  • So what does a quantum computer look like in the real

  • world? The quantum computers developed by companies such as

  • Google, IBM and Rigetti were all made using a process

  • called superconducting

  • And this is where you have a chip the size of an ordinary

  • computer chip and you have little coils of wire in the

  • chip, you know, which are actually quite enormous by the

  • standards of cubits.

  • There are, you know, nearly big enough to see with the

  • naked eye. But you can have two different quantum states of

  • current that are flowing through these coils that

  • correspond to a zero or a one.

  • And of course, you can also have super positions of the

  • two. Now the coil can interact with each other via

  • something called Josef's injunctions.

  • So they're laid out in roughly a rectangular array and the

  • nearby ones can talk to each other and thereby generate

  • these very complicated states, what we call entangled

  • states, which is one of the essentials of quantum computing

  • and the way that the cubists interact with each other is

  • fully programmable.

  • OK. So you can send electrical signals to the chip to say

  • which cube it should interact with each other ones at which

  • time. Now the order for this to work, the whole chip is

  • placed in that evolution refrigerator.

  • That's the size of a closet roughly.

  • And the calls it do about one hundredth of a degree above

  • absolute zero. That's where you get the superconductivity

  • that allows these bits to briefly behave as cubits.

  • And IBM's research lab in Yorktown Heights, New York, the

  • big tech company, houses several quantum computers already

  • hooked up to the cloud. Corporate clients such as Goldman

  • Sachs and JP Morgan are part of IBM's Q Network, where they

  • can experiment with the quantum machines and their

  • programming language.

  • So far, it's a way for companies to get used to quantum

  • computing rather than make money from it.

  • Quantum computers need exponentially more cubits before

  • they start doing anything useful.

  • IBM recently unveiled a fifty three cubic computer the same

  • size as Google's sycamore processor.

  • We think we're actually going to need tens of thousands,

  • hundreds of thousands of qubits to get to real business

  • problems. So you can see quite a lot of advances and

  • doubling every year or perhaps even a little faster is what

  • we need to get us there. That's why it's 10 years out, at

  • least.

  • Quantum computing would need to see some big advances

  • between then and now, bigger advances than what occurred

  • during the timeline of classical computing and Moore's Law.

  • Oh, we need better than Moore's Law.

  • Moore's Law is doubling every two years.

  • We're talking doubling every year.

  • And occasionally some really big jumps.

  • So what's quantum computers become useful?

  • What can they do? Scientists first came up with the idea

  • for quantum computers as a way to better simulate quantum

  • mechanics. That's still the main purpose for them.

  • And it also holds the most moneymaking potential.

  • So one example is the caffeine molecule.

  • Now, if you're like me, you've probably ingested billions

  • or trillions of. Caffeine molecules so far today.

  • Now, if computers are really that good, really that

  • powerful. We have these these tremendous supercomputers

  • that are out there. We should be able to really take a

  • molecule and represented exactly in a computer.

  • And this would be great for many fields, health care,

  • pharmaceuticals, creating new materials, creating new

  • flavorings anywhere where molecules are in play.

  • So if we just start with this basic idea of caffeine, it

  • turns out it's absolutely impossible to represent one

  • simple little caffeine molecule in a classical computer

  • because the amount of information you would need to

  • represent it, the number of zeros and ones you would need

  • is around ten to forty eight.

  • Now, that's a big number. That's one with forty eight zeros

  • following it. The number of atoms in the earth are about 10

  • to 100 times that number.

  • So in the worst case, one caffeine molecule could use 10

  • percent of all the atoms in the earth just for storage.

  • That's never going to happen.

  • However, if we have a quantum computer with one hundred and

  • sixty cubits and this is a model of a 50 kubert machine

  • behind me, you can kind of figure, well, if we make good

  • progress, eventually we'll get up to 160 good cubits.

  • It looks like we'll be able to do something with caffeine,

  • a quantum computer, and it's never going to be possible.

  • Classical computer and other potential use comes from Wall

  • Street. Complex risk analysis and economic forecasting.

  • Quantum computing also has big potential for portfolio

  • optimization. Perhaps the biggest business opportunity out

  • of quantum computing in the short term is simply preparing

  • for the widespread use of them.

  • Companies and governments are already attempting to quantum

  • proof their most sensitive data and secrets.

  • In 1994, a scientist at Bell Labs named Peter Shaw came up

  • with an algorithm that proved quantum computers could

  • factor huge numbers much more quickly than their classical

  • counterparts. That also means quantum computers is powerful

  • and efficient enough could theoretically break RSA

  • encryption. RSA is the type of encryption that underpins

  • the entire internet.

  • Quantum computers, the way they're built now, would need

  • millions of cubits to crack RSA cryptography.

  • But that milestone could be 20 or 30 years away and

  • governments and companies are beginning to get ready for

  • it. For a lot of people, that doesn't matter.

  • But for example, for health records, if health records to

  • be opened up that could compromise all kinds of things.

  • Government communications. Banking records.

  • Sometimes even banking records from decades ago contain

  • important information that you don't want exposed.

  • But the problem we've got is we don't really know when

  • we'll be able to do this or even if we'll ever build one

  • big enough to do this.

  • But what we do now, is that if you don't update your

  • cryptography now, all the messages you send over the next

  • few years and the ones in history could potentially be

  • read. What this means, for example, is if you're a Cisco

  • selling networking equipment, you're going to offer

  • quantum-safe encryption as an option in the very near

  • future. Becayse even though it doesn't look like you need

  • it right away. If your product doesn't have it and a

  • competitor does, guess which product gets bought?

  • One big issue facing quantum computing, other than

  • increasing the number of cubits while keeping things

  • stable, is that no one actually knows the best way to build

  • a quantum computer. Yet the Quantum computers, a Google of

  • IBM and other companies show off are very much still

  • experiments. There's also a big education gap.

  • Not many people are studying quantum computing yet.

  • China is pouring billions into quantum computing education,

  • and the U.S. Congress passed a law in 2018 called the

  • National Quantum Initiative Act in order to help catch up

  • watching people get rid of him.

  • Which means that you want to invest in them now.

  • You want to be hiring people with quantum computing

  • knowledge. Not necessarily to do quantum computing, but

  • because you want that intelligence in your organisation so

  • you can take advantage of it when it shows up.

  • Now China, with its promised $10 billion in it, is really

  • upping stakes in terms of the number of Chinese quantum

  • physics PhDs that are going to start appearing.

  • And you know if that hair restoration or life extension

  • drug happens to be property of the Chinese government, what

  • does that do to the world economy?

  • That's much more powerful than making war Other experts

  • have compared Google's announcement to Sputnik, the Soviet

  • satellite launched into orbit in 1957.

  • The beach ball sized satellite was the first manmade object

  • to orbit the Earth. But Sputnik didn't really do anything

  • useful other than prove launching something into space was

  • possible. Many people are surprised that where exactly we

  • are. For those who are just getting started, they like to

  • make noise about vacuum tubes and Sputnik and things like

  • this. But let me give you some numbers.

  • IBM has had quantum computers on the cloud for three and a

  • half years since May of 2016.

  • We're not in any sort of Sputnik error.

  • We're not landing on the moon.

  • But for those of you who like space history, I think we're

  • probably well into Mercury or Gemini.

Quantum computers use the natural world to produce machines

字幕と単語

ワンタップで英和辞典検索 単語をクリックすると、意味が表示されます

B2 中上級

量子コンピュータの誇大広告、説明されています (The Hype Over Quantum Computers, Explained)

  • 10 2
    李柏毅 に公開 2021 年 01 月 14 日
動画の中の単語