[MUSIC] With classical computing,

you reach some limitations, and you run up against physical barriers. And one of those is just related to

the ever decreasing size of the things that are actually bits,

or doing the computation. At these small length scales, you start

to affect atoms, individual atoms. And you need to worry about this sort

of different regime of physics which is typically related to the smaller systems,

which is quantum mechanics. [MUSIC]>>Okay,

let’s get started as soon as possible. My pleasure to have Eric today

defending his thesis and we are excited to be able to ask some questions.

>>[LAUGH]>>Thank you.>>Good call.>>I guess I’m into my sixth, starting my seventh year at UCSB

in the physics PhD program. So I’ll be talking today about some of my

work in the group, and the title here, computing prime factors using a Josephson

phase-qubit quantum processor. So I thought I would start with just sort

of introducing the ideas of classical computing, can look at where we are. So if you think back to around the 50s

when we had this tabletop transistor, which occupied a meter squared, all right? And here we are today,

with Intel Sandy Bridge technology, which fits a billion transistors in about

a chip this size, 100 millimeters squared. As you continue to press these smaller and

smaller sizes, you no longer can do the same things that you were able to do,

and you can either compete with that, or you can decide to harness that power.

>>And the copper powder filters are up here.>>[MUSIC] All right, there we go.

>>Wow, I haven’t seen this before.

>>Yeah, the copper can dude. It’s looking hot. I’m going to pull off the bias tees with

the box and just leave the attenuators. So I’m going to pull from the attenuators. Does that make sense? So pull from here.

>>Why?>>So I wanna bring the bias to you in the box.

>>Again in the box. Okay, that’s exactly what I was not doing.

>>It’s not just that we’re gonna try to make it a better laptop or

a better new phone, or whatever. It’s actually a disruptive technology

in that it would be a different style of computing. It solves problems that on a classical

computer are retractable and you just cannot do.

>>Let’s first put on the safety can here.>>[INAUDIBLE]>>[MUSIC]>>Let’s come down.

>>I just want it flush up top. [MUSIC]>>Stop.

>>The o-ring is falling down.Yeah. It’s fine. Perfect.

>>Yeah, we tore a little nylon. Yeah, hey hey.

>>Yeah, and definitely here.

>>Well, that’s a fixable problem.

>>Thanks guys, looks great.>>As an experimentalist, within the group that I work at, the

expectations are to be able to basically propose an experiment and be able to

do every part of that experiment and ideally, successfully

execute the experiment. Not only in the sense of taking

the correct data analysis and all that but also being able to write it

up in scientific journals. That’s kind of important for

your career and what people, typically,

are gonna be interested in. Of course, the exciting parts about

quantum computing is even the underlying fact that we’re using quantum

mechanics to do this computation now. And rather than these really certain

position or momentum operators and things we can actually say a particle is

actually located here, becomes this wash. It’s actually a probability distribution

where these things can be found. And what we do in quantum computing is

it’s actually a number of experiments are run to come up with a probability. The quantum bit, this qubit, we measure it’s gonna collapse in

one of this things, either 0 or 1. But as you repeat the experiments you get

these probabilities of the system, right? And it can be actually in a super position

of states, where it’s both 0 and 1. The question is whether this is

really something useful to have. And then in practice,

can we actually build such a device? So all right, a practical use of a quantum computer

might be, say to compute prime factors. Let me frame the picture. Multiplying two large

numbers is challenging for most humans, but relatively

straightforward for a classical computer. Recall that computers used

to be used as calculators, not just social networking devices. Finding the prime factors

is the reverse problem. The challenge here is,

say I gave you some composite number, N, where N is composed of p times q. And we seek these prime factors p and

q, okay? Sounds easy enough. Well, let’s maybe try something

like the RSA number of 2048, which is about a 617 decimal digit

composite number that looks like this. So if you’d like, I can give you guys some

scratch paper and wait a few minutes here for everyone to try to factor this.

>>[LAUGH]>>And we’ll start a clock. And let’s see how long it takes

to actually compute this. So what is the time to compute? Well, if I arm you with sort of

the best classical known algorithm, which is the General Number Field Sieve,

and I’ll take Peter Shor’s algorithm, the quantum algorithm, and

we’ll press the go button, right? We’ll wait.>>[MUSIC]>>So it’s gonna take about

the age of the universe for you to solve it with the kinda

classical General Number Field Sieve, whereas it will be on the order of

seconds here with Peter Shor’s algorithm. All right, it’s this image that

really motivated me to actually do this experiment. So why is this relevant,

or why should anyone care? Well, say you wanna send secure

information on the web, right? And you give the credit card number,

bank account, whatever. We encrypt that information between

you and the seller, the vendor, whoever, with basically

this RSA encryption scheme, which relies fundamentally on using these

large composite numbers that are composed of two very big prime numbers multiplied

together to give you that number. That’s kind of the key with which we use

to pass this information back and forth. However with Peter Shor’s algorithm,

the quantum computer, one could crack this very quickly, in order of seconds rather

than the edge of the universe right? So then the question is well,

what security do we have left? What’s great is there’s actually quantum

encryption, which one can actually find out if someone is eavesdropping on their

transmission via quantum entanglement. And it’s a much stronger

test of your information, of how well you secure your information. So the experiment that I’m working

on is to actually map this idea of using Peter Shor’s algorithm

to factor a composite number. One would need a lot more resources

to actually factor larger numbers. So I’ve been working about 5 years

to factor 15 into its prime factors. Got into the clean room,

I’ve made a device. Have a quantum processor composed

of nine quantum elements. Then we insert one of these devices

in a superconducting cavity, wire bond it to make

electrical connections. [MUSIC] And mount it in a helium three,

helium four dilution refrigerator. Thanks guys. I think that’s it, the rest,

I’m gonna take this can off and finish wiring the boxes. Appreciate it. And in order to remove thermal noise, and enter the regime dominated by quantum

mechanics, we evacuate the chamber and cool the quantum processor

to just above absolute 0. [MUSIC]>>[INAUDIBLE] [MUSIC]>>And it is with this quantum

processor that I’m gonna again try and find the factors of 15. [MUSIC] How’s it going?

>>Okay, we should talk.>>Okay.>>We’ve been talking about this. Is this folklore and

myth, or is this science? So if we want to keep this as folklore and

myth, we don’t do anything. [MUSIC]>>And we return that we get, indeed that

15 equals 3 times 5, 48% of the time.>>[APPLAUSE]>>Sir, 48% success rate that was what you were supposed to get?

>>Ideally, you should get 50%, right?>>That’s the best that Shor’s algorithm will do.

>>Okay.>>It’s 50%>>Okay.>>Yeah.>>So, this is clearly just beautiful work. But it does make me wonder, so what’s

the number of qubits we’re aiming for? A thousand? 10 million?

>>10 million would be great.>>[LAUGH]>>And that will be next year, do you think?

>>Yeah, I think that will be next year. Are you recording that?

>>[LAUGH] [MUSIC]

## 100 Comments

## I haven't thought of a funny name yet.

I understand but that analogy is pretty poor. Imagine its a car without wheels, it can move and function, but at a very poor and ineffective way.

## betamale3

Looks like its going to be a real boon to mankind, worth every penny.

## Ralph Walters

15=3(5) 48% of the runtime. The more interesting question is what was 3(5) 52% of the runtime?

## aaron4820

Yeah I was just about to ask that then I saw your post, it would be very useful if 48% of the time, 15=3*5, and 52% of the time it's all over the place, so that you know whatever the constant is almost 50% of the time, then that is the right answer… but if 52% of the time you get the same wrong answer (say 5*2)… then there's no way I could tell if 5*2 is the right answer or 5*3 is the right answer in practice..

## Ralph Walters

Yes, very very interesting, and a little disturbing. Maybe the proposition that we live in a multiverse has some legs. When we get down to the very fundamental level of nature, we arrive at a sort of gateway to other universes with different sets of physical/mathematical principles. 48% of the time the experiment shows that we are "here", and 52% of the time we are somewhere else, wherever that may be.

## sigmapsicharlie

This is cool and what not, but I'm not sure how this is going to help starving people around the world not starve. Or help "3rd world" communities have electricity. All this seems like a new way to make a new Iphone or computer to play Call of Duty. what a waste of money and brain power. I think Tesla had it right to want to make power free for all the world. And until we have these guys dropping quantum everything and working on getting everyone a light to read with, this is all useless.

## MASTERmw100

two things that really bother me at the end of the video…why do all of these fucks have strawberries and also whats wrong with the old dude's head? is he a fucking reptilian??

## TheDoctorRulesPSN

First of all, quantum computing will not be used for video games or for personal use. This technology has the potential to solve extremely difficult problems in various sciences- like medicine. Eventually, someone will be able to use a quantum computer to fully understand cellular processes and in turn, manufacture powerful medications.

## Gaurav Arya

And Northrup bought a Q comp for "Unknown purposes" 🙂

## badpanda84

""First of all, quantum computing will not be used for video games or for personal use""

LOL it would be pretty funny if the first thing people did was play minesweeper or solitare. If calculating 5 x 3 = 15.. is a breakthough… then playing minesweeper would be amazing

## simplecrs13

Can it play Crysis?

## james woodfine

although you say that but didnt the computers we used today started just for very high places like governments, medical, science and big companies. so maybe in 40 to 50 years time we could have quantum computers in are houses playing crysis 23.0

## SebMeister

No, no..

## Kekekeje Ajsgshdj

Well considering how much computing power we could have compartively now, the worlds fastest computer will be like a calculator if we can achieve this processing power, which i would think help the medical field find cures vastly quicker,

"However, the computational basis of 500 qubits, for example, would already be too large to be represented on a classical computer because it would require 2500 complex values (2501 bits) to be stored.[1"

## Laharl Krichevskoy

4:08 Python?

## Laharl Krichevskoy

remember capitalism? our computers are shit compared to what they could make, they dont make them better becuase they can still sell us expensive crap. Most of the people know shit about technology anyway. Last 3 or 4 years have been a pita talking about cpus, neither intel nor amd have made any mayor progress. amd is crap and intel stoped getting better after the first FX (bulldozer). Unless they could make more money i dont see them developing something they cant profit.

## James Sasinowski

How could quantum computing be reliably used with an ideal success rate of 50%?

## DragonsREpic

10 million*

## Smokestack

Because silicon computers, what we are currently using, will reach max computing power in 15 years or so. Quantum computers have no real limits on computing power that we know of. It is just a matter of discerning effective methods to turning that higher power into an effective and accurate means of computation.

## isodoublet

Theirs isn't a quantum computer in the same sense. It performs a procedure called quantum annealing which is appropriate for certain optimization problems but not general quantum computer solvable problems.

## isodoublet

Yes you can, because checking if a factorization is correct is straightforward: you just have to multiply the numbers together. If you get the wrong result, run the algorithm again. The probability that you'll never get the right answer is about (1/2)^n, which gets very small very quick.

## polishfish

strawberry basket refreshments!

## Kyle Mallard

True, but by computing a calculation in seconds compared to 10^9 years, you can definitely afford multiple tests until the right key is obtained.

## QuakerOats

Closest thing to black mesa..

## Eralp Bayraktar

Well let me explain it to you, 15 = 3×5 is integer factorization and those kinds of problems are 'easy-to-check', meaning if you are given a solution you can easily check whether they are correct or not. Keep this in mind. If I told you 987=3*7*47, you can immediately multiply 3 numbers and check whether I was right or wrong. 'Immediately' is compared to the time you would spend if I asked you the question "What is the integer factorization of 987?" Do you see how hard is it?

So coming to the problem this quantum computer gives us the right solution 48% of the time, wow this is fantastic! I would be glad if I could get the right solution 0.001% of the time! Do you see why?

Because I could feed the output of the quantum computer to a classical computer and let it check if the solution was correct. How long would it take? 1 ms? Well I certainly could run the quantum algorithm 1000 times, and check the output 1000 times, only 1 second overhead with these numbers. (Of course you would stop if you got the right solution inbetween 0-1000)

The good thing is that integer factorization problem is easy-to-check as I just said. Consider TSP problem, given a path, can you easily check if it is the solution? Nah, you can't, as much as I know.

## Gavin Deulufount

@superstrongholdkapo- "What exactly does this mean? That math is a partial illusion?"

It's still all based on math. A problem that's hard from one approach is solvable from another. Especially when you're changing the whole basis of computing from the 'mechanical' macro-world to probabilistic single particles.

## Gavin Deulufount

Gotta know, what happens with the other 52%? Is it just noise, some kind of imaginary solutions, a product close to 15, non-prime numbers, or what? It would help to understand the whole jig if we knew.

## Zycho

Right now I'm hoping quantum computing will fail because if it succeeds it will be the end of the internet. Maybe that is a good thing in the end though..

## Felix

This will ruin all online security. All sensible data have to be taken offline if this is accessable to anyone.

## mukul dhiman

Seriously, people here are so hateful, these all people are like the dude who said,"There might be a market for only 10 computers." People don't realise how useful this technology is. This can enable you to exhaust near to infinite probabilities, like finding a new stable drug, simulating possible outcomes to complex problems, etc. These computers might not make it to your home, they are going to replace the supercomputers.

## Neale Scott

I don't get it. What was the point of all that?

## ZionJW

did he go to slvhs because the stool he was sitting on looks like the ones we make

## Mariksel Azemaj

nuk mora vesh asnje gje

## Martyn Norman

"But what…is it good for?" — Engineer at the Advanced Computing Systems Division of IBM, 1968, commenting on the microchip.

## Tim Teatro

I saw iPython and matplotlib! Woot!

## AndiotiZ

I still think Quantum computing still has a LONG way to go. It's more likely that Nano or molecular Computers will precede current generation of computers.

## Jeff Bond

If we had 10 million qubits working optimally we'd have enough computing power and data to store every electron, photon, and atom in the universe in so many different states that writing down the number of different state universes would take longer to write than the age of the universe.

## Silverheartsmusic

Erik & crew – think about ESD protection when you handle the silicon. static clothes > movement > kV's charge > fries or weakens the silicon. You're carefully carrying the processor – but you're generating a charge – it's looking for the easiest discharge path. Been there done that. ~Bob, EE

## monicaxireland

This is such a complex subject and this is what is being made public about what is being discovered. It makes me wonder at times are even more amazing breakthroughs being made that we are unaware of. The problem of course is that some of these teams are being doomed to going down blind alleyways that will never work without realizing they have been visited before. Has anyone created a list of the different kinds of Quantum research techniques that is currently going on. (i.e like Seth Lloyd's work at MIT or the different approach Dwave are using) to try and define clearly methodologies and outcomes?.

## SignatureCha0s

There's actually a programming language being developed that should work with quantum computers (at least according to simulations and theory) called qBasic.

## Artyte B.

p=np 🙂

## Jeremy Kilroy

So in order to make it a functional computer. It needs an algorithm that runs what ever you are calculating several times, two arrays with each being run let's say 5 times , if the same value is returned 2 times out of 5 it returns that value to the first array. It does the same thing again and if it returns the same value twice out of five it returns that value as the correct value to the second array. Then it compares the arrays if they have the same value it returns it as the correct calculation and then proceeds to the next calculation, else it reruns the calculation. I think that would work.

## Alex Rosales

Can this run crysis 3 at max settings or will I have to go to low?

## Alex Moen

But HOW did this happen? If you're able to do a basic computation like this, haven't you solved the decoherence problem to some extent (although not perfectly, since it was 48% of the time versus 50)?

## mr.mixleplix

I believe they need to attach the T-1000 module to the flux capacitor to for voltron fusion to be possible.

## Nick Chahley

This will revolutionize the Ti-83 graphic calculator. FINALLY!

## Valentina Sherren

I'm just trying to get my head around electrons, molecules and mathematical formulae and wondering how Quantum computing would be beneficial to human beings. Then I spot a headline about how it could grant immortality – which would be quite an achievement – and I can better understand why it would be worthwhile for these very clever scientists to continue their research into the world of Quantum Physics and the creation of Quantum Computers.

## Leggo My Ego

Call me when it can be used to display porn, even if it's 48% of the time.

## Clorox Bleach #1

Apple will sell it for 1.3 million with only 4gb memory and integrated graphics as the base standard.

## Lucas Citolin

I am loving the comments on this, love science and fun !

## Henry S

amazing achievement

## Unfiltered Reality

So weird, handing out strawberry baskets and coffee.

## GlennBen

I think they should add the vitron 5000 to the colex capacitor to achieve a high velocity radrant so that the regenitive retrospective invulator will transistally reconsultrate the flux hydron so fusion may be finally possible.

## Declan

Where can I find the music?

I've looked around for it and can't find it.

## MIchael Wright

But does it blend?

## JadendayZero

calculating 3*5 = 15 40% of the time? lol thats not reassuring.

## upplsuckimcool16

If the Qbit can be 1 AND 0 wouldn't that make for an almost infinite amount of storage space?

## Bruce McKinlay

Nice Whalers crest on that guys shirt near the end!

## Bill Nichols

Why is the solution of 1 x 15 not an option? Does this account for the reason 3 x 5 was the solution 48% of the time?

## nat copeland

very interesting

## littlegodpan1

When something is completely understood it can be explained in the simplest terms. I will continue to do research. Because I don't understand what it is for. It seems to be a really fast calculator. What is their projected goal?

## Youn Kim

10:02 The questioner's scalp looks weird.

## fcycles

I am guessing that qbits are use in super-impose states to compute all possibilities at once… but to get the answer out of the super-impose states do you need an operation which will make the answer you are looking for come-out? Can someone give an example?

## Computer Fraud and Abuse Act of 1986

Does it have a cathode ray tube output?

## Alex Sander

weeeird dude i am pretty sure 3×5=15 100% of the time

## Sky Sevastian

Dont our brains work like quantum computers?

## Bethis GG

whats the starting sound? its pretteh chill m8-

## MrWr99

3*5=15 40%!

## SycheRyder

I feel so incredibly stupid trying to follow this. So instead I'll focus on that guy's pants in 3:41 . What up with those?

## joe chu

this might be a stupid question, but does the success rate go down when the input number is larger?

## Joseph Stokes

The purpose of the horrible distracting shit music is?

## Anto Caballero

Lotto uses the same computing vector of 15. This took me less than a min to fig out. 0 thru 9 from dust to ash. Think about that for a moment.

## yoyofargo

3:43 Is…

😐

Is that windows xp running in 2012?

## Jan

most overhyped technology ever.

## gustavodude1

I must say, it was lot of money/time spent for 48% probability. Best wishes !

## Danny H

What if your device is working just fine and quantumly speaking? 48% of the time the results are indeed 3 and 5. 52% of the time it is 1 and 15 and -3 and -5

## salzee

Fascinating.

## Hansel Rainer

Why they don't use AI like DeepMind to solve the programe to run the quantum computing?

If using human itself to programme such things, it's gonna take time.

## Eric Askew

not gonna lie, was down with the intro music

## James Neave

Just… please make sure there's one in my phone for encryption before there's one at GCHQ or the NSA listening to my phone. Ta.

## Fuh Que

What good is a computer where it gives a right answer (prime factors or whatever) 50% of the time and wrong answer 50% of the time? Sounds like just guessing is just as effective and far cheaper.

## R Robin

If the algorithm works on a digital computer already what is the significance of 3×5 = 15 with a 52% error, nothing ground breaking here

## Denuris Turgryn

8:30 ever heard of armaflex, guys?

## CC Smooth

Im assuming that the times it produces a different answer that thats due to the qbits not behaving as they should? Meaning that eventhough theyve supercooled and evacuated the chamber that some outside influences are still affecting the subatomic particles?

## Ray okeedoki

In my imaginings of better computing, I always thought that a knowledge base in 3D space, would become the standard. It would contain all of mankind's knowledge to date. A clone of Artificial Intelligences would have it as their base.

But something I was reckoning on was 3D positioning could effect things by the interactions of actual electro/magnetically at some static level, would influence each other… but I did fail to say at the beginning this was for AI in 3D space… with the concept put forth that all knowledge has its place in logic compared to all other knowledge. And therefore would have its best positioning in 3D structure, a bit like a brain, but better. As it can be logically "wired" before it is designed to mass produced by machine into the solid unit. That is it evolves a Virtual add on for all its learning, so that in the next major upgrade, all of the AI's combine their data, so all get the new upgrade, either virtually or in the 3D mass produced new base model.

Then again, it could be colours and crystals at some point, making binary a fraz of billions of colours…

## Vlad

A quick update from 2017, we are at 2000 qubits with D-Wave's 2000Q 😀

## Kirk

man i wish i had this kind of intelligence and know how..

## The Thugger

« And then we cool it down to almost absolute zero » wtf does absolute zero even mean?

## TheUnown4

That fidget spinner t-shirt

## John Merrell

so modest….

## American Citizen

Pocket size version any day now?

## American Citizen

Your bank balance is only $15. But only 48% of the time? This is going to make accounting difficult!

## American Citizen

If you have to "verify" the result using a traditional computer then you will have a problem. Consider Gödel's incompleteness theorems — if you can solve a problem with a quantum computer that same problem will not be solvable on a traditional computer. If a quantum computer is only "sometimes" correct then you will never be able to verify an quantum answer by checking it on a conventional computer.

## David Wilkie

Hypothetically (?), the self-organization of the Quantum Fields Mechanism, is what atomic structures in the Standard Model and Periodic Table are, all resonant summed history, phase-locked together in chemistry and physics, and are arranged according to the languages of mathematics in our perceptions to behave in particular "computed" ways of psudo-randomness. Ie if we can remove the "noise" of the annealing process, then a predetermined result of superposition will match the actual state output by the device, ("The Babbage accuracy of construction problem").

It's the ultimate parallel processing by design using an effect similar to the growth of spacetime-elemental grains (of primes) in crystal-coordinate resonance?

Sum of all histories, now to now=>eternal constant/resonance.

If the analog computation, (another name for an analogy by simplified comparison), is continued, then it's the readout of results that is the determining factor, (literally), of a successful operation, and so the methods and devices used in MRI, tuned to the desired range of solutions, is critical? (And possibly a compound set of scanning frequencies like WiFi in reverse..)

It's analogous to the "hard problem" of mind in body, because it's a phase-locked distribution of QM-Time, Q-chemistry by Fields Modulation Mechanism e-Pi-i resonance embodiment. If coherence is disturbed, the content in context connection goes very wrong, because the singularity synchronisation of the Quantum principle is also the Quantum Operator of the Exclusion Principle.

The nearest "device" analogous to the imagined operations of a Quantum Computer is possibly emulated by Human Savants, and that seems to be based on intense focus and vast memory storage with bio-logical Random Access (?)

If structures, =events, within the Universe are resonance grains of spacetime, and bio-logical organisms are integrated multi-oscillator quantum information computation, insect colonies are hive-mind distributed computation, then brain-body organisms are integrated cellular hive-mind, up to the levels of self-recognition and "consciousness".

Everything is phase-locked in resonance and timing spacing coherence/exclusion of cofactor-primes, AM-FM phenomena.

One Multiverse ghost of coherence in one Quantum Computational Machine?

## 蔣國棟

if we seek the prime factors of N = p x q,

“N equals post wrong question” can’t seek a “prime n quick” solution with no me in we.

no equal prize nor request for prime n priwe then → pri

the answer is pri

## James Markco

Fuck im dumb.

## Abd Aziz Abd Hamid

Is there time frame?

Will there be possibility of total waste of time and energy?

## crazieeez

"I have been working about 5 years to factor 15" 😀

## PhatGar

We're on our way, I say.

Remember, the first models of our Computer was just like this. One computer required one room, it was only able to compute numbers and nothing else, it has a separate room for the CPU, it's loud, it takes more than one person to make a computer work before. People thought it was impossible to finish. But not anymore.

With quantum computers, we are doing the same thing again, but more ambitious.

## John Middleton

15=6 112358134711236915