Quantum Computing 101

What exactly is quantum computing? How does it relate to quantum mechanics? And what sort of problems could quantum computing solve in the future?

Learn more about your ad-choices at https://www.iheartpodcastnetwork.com

See omnystudio.com/listener for privacy information.

Episode Transcript

Speaker 1 (00:04):
Get in touch with technology with tech Stuff from how
stuff Works dot com. Hey there, and welcome to tech Stuff.
I am Jonathan Strickland, the host, an executive producer with
How Stuff Works and a lover of all things tech.
And this is another episode in the little mini series.

(00:27):
I'm recording while I'm attending the Think two thousand eighteen
conference in Las Vegas, Nevada. It's sponsored by IBM, and
IBM does this big conference. It's sort of a an amalgamation,
a gloming on to several different smaller conferences that IBM
has been holding for several years. They kind of pushed

(00:48):
them all together and turned it into a giant, mega conference.
And I emphasized giant. I mean there are tens of
thousands of people attending this conference. It feels like more
than that when you're trying to get through the Mandalay
Bay Conference Center, because holy cats, lots of executives, a
lot of blazers, a lot of blazers out there, folks.

(01:09):
I gotta watch my what I say because pretty much
everybody in there is a giant stakeholder in some big
business or another. And chances are if I if I
say something rude, I've just insulted a millionaire, and I'm
not in that tax bracket. But let's talk a little
bit about one of the topics that got a lot

(01:29):
of coverage here at IBM think two thousand and eighteen,
and that is quantum computing. It's a big deal, and
that's because quantum computers are beginning to emerge from the
realm of experimental science into practical applications. In fact, you
could argue it's already there and has been for a
couple of years, but it's still relatively new and I

(01:51):
think very mysterious for a lot of people. And I've
talked about a little bit in previous episodes, but I
really wanted to dedicate a an entire episode kind of
quantum computing one oh one and really talk about the
principles behind it, the history behind it, what it might
be used for, why it's such a big deal in

(02:12):
the first place. So this is our full episode on
the topic, and I'm going to reference some of the
things I've learned while I've been at this conference. Let's
do what I love to do. This is like a
good old traditional episode of tech stuff. We're gonna dive
into the history of quantum computing and quantum mechanics and
quantum theory. So this all begins before the computer age.

(02:34):
We have to discuss the history of quantum mechanics itself. Now,
I'm not going to go into exhaustive detail, because to
do that would require an entire podcast series, not just
an episode, but a series of episodes to kind of
talk about all of the developments in quantum mechanics. And
not only that, but it's a messy history filled with

(02:57):
a lot of scientific debate and our humans and uh
experiments and counter experiments, thought experiments, aim calling. There was
some adultery in there too. I mean, it's it reads
like a soap opera at times, and and like I said,
it's just it's so deep and dense that to really

(03:18):
cover it would require multiple episodes. So this is kind
of like a an introductory a bird's eye view of
the history of quantum mechanics. So let's talk about the
developments around the turn of the last century, the twentieth century.
In nineteen hundred, it only been a couple of years
and scientists had even discovered the existence of electrons at

(03:38):
that point. No one was even sure in nineteen hundred
if electrons were even part of the atom. They didn't
know our electrons actually a component of atoms or are
they something else? So do they coexist with atoms but
they're not bound to atoms? They weren't sure. In nine hundred,
there was general agreement that atoms were in fact a

(04:00):
kind of a fundamental particle, but beyond that, there wasn't
a whole lot of agreement on them. No one was
really sure what made the atoms of one element different
from another, and therefore they weren't sure why elements were
different in the first place. They could identify elements, they
could identify the qualities of elements, but they couldn't explain

(04:22):
why they were different from each other. Well, in there
was a smarty pants physicist, Max Planck, who was trying
to work out some reasons behind a curious observation that
people had noticed for centuries but didn't They couldn't explain it.
And that was the nature of heat radiation and the
light that it can produce. So let's say that you're

(04:45):
a blacksmith and you've got some iron, and you put
in the forge and you heat the forge up. Eventually
that iron, as it grows hot, will begin to glow,
and it first will kind of glow red, and then
that red will get brighter and brighter, kind of turned
into an orange. And if it gets hot enough, it'll
glow white. If you could get it hot enough before

(05:05):
it melted, you could make it even glow blue. These
different colors would represent different energy states, but no one
knew that at the time. No one was able to
explain why iron would change color as it got hotter.
So Plank was working on this problem. He was trying
to figure out, well, what is what explains us, or
what at least describes this, and eventually came up with

(05:28):
a formula that fit the observations he made in experiments.
He had figured out a formula that that seemed to fit,
But why did it fit? Why did that formula describe
what was happening? He couldn't tell. He wasn't sure, No
one was sure at first. He kept working on it,
so eventually Plank figured out that the atoms could apparently

(05:52):
only take on certain quantities of energy, So it could
take a certain amount of energy, and then any above
that it could not accept until it got to the
next specific allowable energy level. So you could think of
it as steps of energy. You could accept a certain amount,

(06:12):
and then you could step up and accept a new
larger amount, but anything in between those two steps didn't
fit the formula. And this was very curious. It wasn't
something that was continuous, right, This idea of steps of
energy levels was really perplexing at the time. You might

(06:34):
think of it more like a continuous string, but it wasn't.
It was this broken series of steps. So this really
got people wondering what the heck was going on. Um,
how could materials take on specific increments of energy rather
than any arbitrary amount. Planck didn't know. He didn't know
why it was happening. He only knew that it was

(06:55):
happening based upon his observations, and that the explanation he
had fit what he observed. He just couldn't explain why
it worked. He announced his findings on December four, nineteen hundred.
Now some people trace that as the origin of the
study of quantum mechanics, though of course at that time
it wasn't yet called quantum mechanics. It did, however, formulate

(07:20):
the foundation of what some would refer to as old
quantum theory. Now that theory stated that these acceptable energy
increments were specific quantities, right quantities of energy, and that
any phenomena that would only accept certain values of a
physical quantity fell into this category, and it typically was

(07:43):
stuff on the atomic scale, tiny tiny scale, not classical scale,
which seemed to follow the rules of classical physics. These
things didn't seem to follow the rules of classical physics.
The rules were different for some reason. So scientists said
that the values of this physical quantity of energy, uh,

(08:04):
we're said to be quantized. That's the values of this
energy is quantized. It was generally believed that you'd have
to do lots of experiments and make lots of observations
to kind of suss out the rules for that quantization
or perhaps even uncover a set of universal rules that
would work in all situations. So there were scientists like

(08:25):
Albert Einstein who seized on this notion, and they began
to apply this idea to other areas of study. He,
for example, Einstein, that is, proposed that the total energy
of a beam of light was quantized. Several other big
thinkers were looking into similar fields. But then the First
World War broke out and that really slowed down progress

(08:48):
in the sciences because a lot of the leading scientists
at the time we're all in Europe, so obviously Europe
being heavily affected by World War One meant that a
lot of that work was put on hold. However, at
the war's conclusion, things picked up again at that stage
after World War One, but before World War Two, you

(09:09):
had scientists like Max Bourne and Werner Heisenberg who were
extending our understanding of the quantized world. Now Born and Heisenberg,
along with Pascal Jordan's, wrote an extremely complicated but consistent
theory of quantum mechanics. Meanwhile, you had another smarty pants

(09:33):
Irwin Schrodinger or Irvin if you prefer, that would be
of Schrodinger's cat fame. He was working on his own
theory to describe quantum mechanics, and for a while, those
two theories were the focus of a pretty nasty war
within physics in which both sides were kind of disparaging
the ideas of the other side. And essentially one group

(09:58):
is saying, you guys are full of it. My theory
describes what's actually happening Here's is a mess, and the
other side saying, nah, our theory is far more descriptive
of what is actually going on your theory is nonsensical.
But then in Schrodinger and Carl Eckert, who was working

(10:18):
completely independently of Schrodinger, both proved that these two seemingly
different approaches were actually describing the same thing. They were
just doing it from completely different points of reference. So
on the surface they superficially seemed like they were at
odds with one another, but underneath that it turned out

(10:39):
they were. They were in alignment. As one book I
read on the subject said, it's like comparing how you
add Arabic numerals to how you add Roman numerals. The
two processes look very different from each other, but if
you do them each correctly for the same two values,
you'll always arrive at the same answer, no matter what

(11:00):
method you use. Now that's not to say that everything
was smooth sailing from that point forward. Many scientists had
problems with aspects of quantum mechanics, such as it's probabilistic nature.
That is, much of quantum mechanics concerns itself with probabilities
rather than certainties. In fact, lots of things and quantum
mechanics become inherently uncertain the more you try and nail

(11:23):
it down, the more uncertain other elements will become. That's
partly what Heisenberg's uncertainty principle states. Heisenberg was specifically talking
about a quantum particles position versus its momentum. Heisenberg stated
that the more precisely you measure one of those two values,
the less you can know about the other one. So

(11:45):
if you measure a quantum particles position with great precision,
you won't know very much about its momentum, and vice versa.
And that this is just a fundamental feature of our universe,
so it's tough if you don't like it. The probabilistic
side of a quantum mechanics is tied also to measurement.

(12:06):
This was a central focus of a debate between two
great physicists, Neil's Bore and of course Albert Einstein. Einstein
was not keen on the probabilistic nature of quantum theory.
Uh He has often been attributed the phrase God does
not play dice with the universe, although that is a
paraphrasing of what he said. And then Niel's Bore was

(12:28):
paraphrases saying God doesn't care what you think he's doing.
Um so that was kind of the back and forth.
Although both of those statements were paraphrase, neither of those
were actually what the scientists were saying, just kind of
was a an interpretation of what they said. Quantum mechanics
experiments wouldn't really produce a definite solution. So we're used

(12:51):
to things like calculations coming up with a specific answer.
Right even let's just take simple arithmetic. If you say
two plus two equals for then you know you realize that,
all right, well, that that makes sense to pless two
equals for that's a that's a certain value. It's a
definite answer. Whereas with quantum mechanics you would get results

(13:11):
that would be listed in terms of probabilities, not in
terms of here is the answer. You would get a
probabilistic distribution of possible values. So that means every single
value you would get would get assigned a probability, and
if you were to measure a quantum state, that would
actually cause it to collapse into one of those probable

(13:32):
values that it possibly could have been. This is also
related to that concept of quantum tunneling I mentioned earlier
this week. The idea of an electron could potentially inhabit
one of any positions that are within a certain field,
and because there's that probability, it means that sometimes the

(13:54):
electron will inhabit that position. And if that position happens
to be on the other side of a barrier, just
because the zone the electron could exist in happens to
overlap that barrier, then that means sometimes the electron is
on the other side of the barrier, even though it
did not physically pass through the barrier. It's it's part
of the weird nature of quantum mechanics and probabilistic distribution. Again,

(14:18):
it's not a certainty, it's a probability. Another concept of
quantum theory that ends up being very important with quantum
computers is that of superposition. This is a pretty tricky concept,
as it is so counterintuitive that it prompted Schrodinger to
create what he thought was an absolutely bonkers example so

(14:39):
that he could illustrate how whacka doodle this idea was
on the macro scale. But today that example is widely known,
or at least it's known by name. That would be
Schrodinger's cat. So what is superposition and what the heck
was that famous thought experiment. Well, superposition refers to quantum
particles as inhabiting all sable states simultaneously. So a state

(15:04):
is really just a feature, something that the quantum particle possesses.
So let's take electron spin as an example. All right,
So electrons can spin in different directions, and for this
particular example, let's just talk about spinning up or spinning down.
So electron can spin up or it can spin down.
According to some versions of quantum theory and its quantum state,

(15:26):
that electron can be said to be both spinning up
and down simultaneously. It's both states at the same time.
It inhabits them while it's in this quantum state. But
when you measure the electrons spin, when you observe it,
the quantum state collapses down into one of the two
possible states. So you're never going to observe an electron

(15:49):
spinning up and down simultaneously because the act of observing
changes that which is observed at the quantum level. This
is the argument some people make that you know, measuring
doesn't matter because if you measure, you have changed the
thing that you were measuring. Now, that is true on
the quantum scale, but as you move up to the
classical scale, it's not really something you need to concern

(16:13):
yourself with. So, uh, you can't confuse quantum mechanics with
classical mechanics. It they are rules that define two different universes,
really the quantum level and then the classical level. So
it's not like classical physics need to be thrown out
the door. They still apply just two things that are

(16:34):
on the classical scale. When you get to the quantum scale,
that's when you have to look at quantum mechanics, and
that's when you start seeing these seemingly weird and counterintuitive rules.
And I say seemingly because the only reason they seem
weird to us is because we cannot observe them directly.
We don't exist on the quantum level um and in
the way that we can perceive it. We can just

(16:57):
work out the math and figure it out, and then
we can design experiments, and through those experiments we can
we can actually look for evidence that supports these theories.
And in fact, that has happened over time. People have
designed experiments to test these ideas and found through the
results of the experiments that those ideas seemed worthy, they

(17:18):
seemed valuable, and and real. Now, Schrodinger's cat is a
way of exaggerating this superposition effect, kind of in an
effort to show how crazy it sounds. So here's the
thought experiment. Let's say you've got a cat, and you
put the cat in a metal case. Inside that case
with the cat is a device that contains a radioactive particle. Now,

(17:39):
that radioactive particle could undergo radioactive decay within the next hour,
or equally, it could not decay within an hour. So
there's an equal chance that it could decay or that
it could remain whole within the span of an hour.

(18:00):
If the particle does decay, the energy it gives off
will cause a glass vial containing a poison to break,
and that will release the poison in the cage and
kill the poor kitty cat. The whole experiment is completely
sealed away. The cat is unable to interfere with the device,
because if you interfere with a quantum state and then

(18:20):
it decoheres, the whole experiment falls apart. So you have
to have this is a thought experiment anyway, but you
have to have it set up in a way so
that the cat's not going to interfere with the quantum state.
So an hour goes by with the cat inside this
cage and the radioactive element in there as well. And
the question you have to ask yourself before you open
up the cage is is the cat dead or is

(18:43):
it alive? Now? According to the super position theory and
Schroedinger's interpretation of that theory, you would have to say
that the cat is both alive and dead at the
same time. That exists in this quantum state where it
is alive and dead. It is only when you open
the cage and you look in and you are essentially

(19:03):
measuring the system this way, because you're making an observation
that the entire system will collapse into one of the
two possible outcomes, And at that point the cat makes
the transition into either being perfectly fine or very much
an ex kitty cat joining the choir invisible, running up
the curtain, kicking the bucket, shuffling off the mortal coil.

(19:25):
You get the idea. This is where you get all
those jokes about the cat being half dead. But here's
the crazy thing. While Schroedinger's thought experiment did make superposition
sound really bonkers, experiments supported the notion of superposition. Now Granted,
we're talking about effect on the quantum level, not something
that's observable in our macro world. Schrodinger would argue that

(19:48):
because the the whole premise of the experiment relied upon
a quantum particle, whether it decayed or not, it doesn't
violate this. The consequences of the at quantum event would
be on the macro level, but that the actual event
itself would still be in the quantum level. Uh. There's
some people who dispute that, so it kind of becomes

(20:11):
a philosophical argument. But the point is that the experiment
started to support this idea of superposition, and it's one
of the few, one of a couple of principles of
quantum mechanics that makes quantum computing such a potentially powerful
tool and a possible revolution in computing in general. The
other big concept in quantum theory that is of particular
importance with quantum computers is called entanglement. Now, this is

(20:35):
the strong correlation between two quantum particles that link those
two particles together, no matter how much physical distance might
separate the particles. So you could take two entangled particles,
and if you could do it in a way where
you're not disturbing the entanglement. You could move one particle
to the other side of the universe from the first

(20:56):
particle and they would still remain entangled. Einstein would call
this spooky action at a distance, and entangling particles means
that these two particles are always going to complement one
another in some way. So let's take electrons again. Let's
say you entangle to electrons so that their spin is correlated,
and if one electron is spinning up, the entangled partner

(21:19):
is always spinning down. This is just one example of
a way you could entangle particles so that means no
matter how much distance separates these electrons, if electron A
is spinning up, then electron B is spinning down. If
electron A starts to spin down, then electron B will
start to spin up, and he'll do it exactly at
the same time. There's like no delay, and this will

(21:41):
happen no matter how far apart those electrons are. It
seems impossible, and yet that is in fact what seems
to be happening with entanglement. However, once you observe one
of those two electrons, then the entanglement is broken and
you will know at the moment of observation, the moment
of measurement, what that other electron was doing, But you

(22:02):
don't know what it's doing anytime after the moment of observation.
You can only say, at this precise moment, the other electron,
wherever it may be, was doing this particular activity. At
that point, the system decoheres, and so it gives you
information but nothing. Some people have argued that this is

(22:23):
a way that you could potentially have faster than like communication.
Others argue no, because all it does is tell you
information that previously existed. The information didn't travel, just your
realization of what that information was occurs to you. It's
another fine distinction that gets into philosophical arguments, and its
outside the scope of this particular podcast, but it is

(22:46):
a fascinating discussion. So together, super position and entanglement are
two of the factors that really make quantum computers so
potentially revolutionary. And it's weird to say potentially, because today
they are actual working Quantum computers just have a somewhat
limited scope right now, but they're getting better all the time,
and in fact, some of the prototypes are really impressive

(23:07):
already before we get to actual quantum computers. There's a
little more history I need to cover. In nineteen seventy three,
Alexander Hollevo argued that for any given number of cubits,
which are quantum bits, you could not possibly carry more
information than that same number of classical bits. So, in

(23:28):
other words, if you have eight quantum bits, those eight
quantum bits could carry only as much information as a
classical bite, bite being eight bits, and of course a
bit being a basic unit of information, either a zero
or a one. However, the eight cubits through superposition could
represent all possible states of that bite. So it's not

(23:51):
carrying more information, it's just carrying Uh. It's hard to
hard to put this in a way that makes sense.
It's not carrying more information than a bite, it's just
carrying every single variation of information that bite could represent. Again,
anotherir fine distinction. This gets really fuzzy and wibbly wobbly
timey y me to me. In the early eighties, people
begin to theorize about the possibility of quantum computing and

(24:15):
talking about how you might use quantum particles to represent bits.
So again, like electrons, you could use electrons in their
spin And this is a quantum quality that electrons have,
and if you were able to put those into a
quantum state, you could use the electron spin to represent
what would normally be a bit in a classical computer.
That's just one possible example, mind you, because you could

(24:38):
use all sorts of different stuff to represent these bits.
You could use photons and their polarization if you wanted to,
or other quantum particles and other qualities. In Richard Fineman
presented a talk in which he lamented the fact that
classical computer systems would be incapable of simulating the evolution

(24:59):
of a quant um system, because quantum systems would just
be far too complicated for a classical computer to do
this in any reasonable time frame. He did, however, hypothesize
that if you were able to create a quantum computer,
you could potentially simulate the evolution of a quantum state.
Theorists began to flesh out what a quantum computer might

(25:21):
look like, and how it might operate, and even how
you might try to go about making one. This was, all, however,
still within the realm of the theoretical In the mid nineties,
and engineer named Peter Shore discovered an algorithm that would
really put a fire under the bottoms of quantum computer researchers.
His algorithm was a set of rules that a quantum

(25:43):
computer could theoretically be able to follow and allow it
to factor large integers much more quickly than a classical computer. Now,
the reason this posed both an exciting opportunity and a
terrifying realization was because factoring large numbers is what most
modern day cryptography is based off of. Uh, you take

(26:05):
numbers that are hundreds of digits long, prime numbers specifically,
so these are numbers that are only divisible by themselves.
And then you take two of those numbers that are
both hundreds of digits long, like five hundred digits long,
and they're both prime numbers, and you multiply those two
prime numbers together, you get an even larger number that

(26:26):
ends up being sort of your public key, your your
key that you used to encrypt stuff. But the only
way you can decrypt the information is if you know
what those two numbers were, those two huge numbers you
started off with were, which is hard to determine. It's
really hard. If you're using a classical computer. It would
take years or more, depending upon how long the number

(26:48):
was to brute force the answer if you're following classical
computer science. But Shore's algorithm was a short cut that
a quantum computer, not a classical computer. A quantum computer
could run and run that same calculation in a fraction
of the time. So, in other words, a quantum computer

(27:08):
following this algorithm that was discovered by Shore could reverse
the process we use to make all of our data secret. Well,
by the late nineties, the first rudimentary quantum computers were
being constructed in the laboratories. They were really primitive. They
could not run very many operations before they would decohere uh,

(27:29):
and then you'd have to start all over again. They
were delicate systems. They were consisting of just a couple
of quantum bits of processing power. But it was the
beginning of the revolution. So how can quantum computers be
so powerful compared to classical computers and exactly what sort
of problems would quantum computers be good at solving? Well,

(27:49):
i'll tell you about that in just a moment, but
first let's take a quick break to thank our sponsor.
All right, so let's talk about bits now. As I mentioned,
a bit is a basic unit of information and it
is binary, meaning it can have only two states. So

(28:13):
we express bits as a zero or a one, and
you can think of that as being off or on,
or down or up. Just as an electron spin has
different states or a photons polarization, so too does a
bit machine. Language is made up of strings of bits.
A collection of eight bits makes up a bite, and

(28:33):
a single bite can represent up to two hundred fifty
six different states. I talked about this recently in the
I p V six episode. I did the numbers in
an ip V four address or just a regular old
IP address. Those are based off octets or bites. Each
number in that IP address can have a hypothetical value

(28:54):
between zero and two hundred fifty five. I say hypothetical
because some numbers are off limits due to the rules
of Internet protocol. But if you didn't have those restrictions,
each of those four numerals in that address could have
a value between zero and two inclusive. Those would be
the two six potential values of that bite. A classical

(29:16):
computer relies on these bits. It's the form of information
the processor takes in and the form it spits back
out again. The information does get translated into formats where
humans can find useful or initiate some action that is
useful to us in some way. Humans have made a
series of computer programming languages, starting with assembly code or
a similar code really, which is just a step above binary,

(29:39):
up to high level programming languages that abstract those zeros
and ones so that we can structure programs in a
way that's more natural for us to understand. It's still
it can look like complete gibberish to you if you
don't know computer languages, but in fact it is far
easier to read than just zeros and ones. So it's

(30:00):
hard to think of any kind of information just in binary.
But in the classical computer, a bit has to be
either a zero or a one, it cannot be both,
and classical computers will run processes in sequence. So you
can speed that up a little bit by using a
couple of different strategies. One is just to make more

(30:20):
powerful processors that can handle more information and smaller amounts
of time. That will help. You can improve bus speeds,
You can improve the speed that a CPU can draw
information from memory or put information back in memory. That
will help too, but eventually you run up against the
upper limits of what we can accomplish with today's technology,

(30:42):
and of course that keeps on improving, but you still
will run up against those limits. You can use a
multiple core processor. Multi core processors are great. You can
even use an array of processors. That's useful if the
computer problems you're working on can be split into smaller
problems that can be solved in parallel. Not all problems
fall into that category, however, and even if your processor

(31:05):
isn't the fastest, if you're talking about a parallel problem,
multiple core processors might be a better choice than a
single powerful core processor. I usually use this particular analogy.
Imagine you've got a math test, and the math test
has ten problems on it. You also have a math
class and it has eleven students in it. One of

(31:27):
the eleven students is a super math genius, and she
has an innate sense of math. It's almost spooky. It's
like she can visualize mathematics all around her, and she
can solve any one problem faster than anyone else in
the class. Just doesn't matter. But the teacher gives the
super smarty genius all ten of the math problems on

(31:49):
the test, whereas each of the other students in the
class each of them, being good at math but not
at genius level, gets only one of the ten problems.
So student one gets problem one, Student two gets problem too,
and so forth. You have ten students working to solve
one problem each, and a supermath genius working on all

(32:10):
ten problems. Who's going to finish first? Well, the super
genius is going to solve each of her ten problems
faster than any of the individual students will finish their
respective problems. But chances are the group of ten will
finish first because they each only have one problem to
work on. They're able to divide and conquer, as it were.

(32:32):
And some computational problems are like that. But there are
classes of mathematical problems that are too tough even for
the fastest classical computers running scores of processors. There so
difficult as to be practically unsolvable. Now I say practically
on purpose. It's not that a classical machine can't solve

(32:53):
these sorts of problems. They just can't do it in
any sort of reasonable amount of time. It could take
years or decades or centuries, depending upon the complexity of
the problem. So what kind of problems am I talking about? Well,
there's a class of problems around the concept of optimization,
and that's a big part of what quantum computers could tackle.

(33:14):
These are problems they get very hard to solve, particularly
as you add more components to it. Now, I'll give
you a very simple example. Let's say you're throwing a
big dinner party. You've rented out a swanky joint. You've
got five tables. Each table has seating for ten people,
So you've got fifty people on the way to your party,

(33:36):
and it's your job to assign seats for each of
the people who are coming. However, there's a problem. Not
all of your friends are crazy about each other. So
let's say you've got a buddy named Sally, and she
would absolutely hate to sit next to Jim. Jennifer would
love to sit next to Sally, but she definitely doesn't

(33:56):
want to sit next to Sally's cousin, Darryl. But m
and Darryl are best friends, so they definitely want to
at least sit at the same table, if not next
to each other, and so on and so forth. You've
got all these different conditions that exist, and you want
to find the best possible seating solution to the problem
of who sits where in order for you to have

(34:18):
a nice, lovely dinner and not have a breakout into
a three stooges pie throwing routine. Well, here's the thing.
The problem of sitting just ten people around a table
is factorial. There are three point six million possible configurations
for ten people to sit at a table. That's just

(34:38):
ten at one table. And remember you have five of
those tables, and you have numerous rules you want to
do your best to follow to ensure that it's a
pleasant party and no one's gonna go home with punch
and pie spilled all over their outfits. So how do
you solve this problem? Well, a classical computer would choke
on this kind of problem because they would have to

(35:01):
run every single possible scenario, and then it would have
to check the results of all those scenarios against the
rules that you had given it, saying all right, well
don't put so and so next to so and so,
and then it would have to tally up all of
those different scenarios, analyze the whole thing, and determine which
one out of all the different scenarios that could come out.

(35:22):
And remember there's three point six million per table, that
which one is the best. By that time, half your
friends have moved away, or had kids, or have become
honored ancestors to generations that follow because it just took
way too long for this classical computer to work out
the problem, and your party was a bust because you
never got the invitations out in the first place. Now,

(35:45):
another problem in this class is called the traveling salesman problem.
This is a classic problem and it goes like this.
Given a list of cities that a salesperson has to
visit to do his or her rounds, what is the
shortest possible route that the salesperson can follow that will
allow them to visit every single city and return home

(36:07):
to their point of origin the shortest possible route among
all those cities. This one's pretty easy to understand, but
it's actually fiendishly difficult to solve, especially as you add
more cities to the problem. So this type of problem
is called an MP hard problem, and the more cities
you add, the harder it gets. So how could quantum

(36:27):
computers do a better job than classical ones with these
sorts of problems? And it comes down to a basic
unit of information in the world of quantum computers, The
quantum bid or the cubit you know, I mentioned it
a couple of times, and a cubit can be placed
in superposition, meaning that in its quantum state, it is

(36:48):
behaving as if it's both a zero and a one simultaneously.
You can also entangle cubits with one another, so that
the state of one cubit and it's entangled cubit are
highly correlated. So you could encoded in such a way
where you say, if cubit A is a zero, do
nothing to cubit B. If cubit A is a one,

(37:10):
flip cubit B to one. That would be an example
of entanglement. With these properties, it's possible to solve these
traditional unsolvable problems in a very short amount of time
if you have a quantum computer with a sufficient number
of cubits. That's because the cubits in their quantum state
can essentially run all possible solutions to a problem simultaneously

(37:32):
rather than sequentially. I'm oversimplifying here, but that's the general principle.
And as you add more cubits, your ability to process
information grows exponentially. Now, how does that work? Well, if
you have a single cubit, but it can potentially be
two states at the same time. Because of superposition, that
cubit actually represents two states, not one. Remember a bit

(37:55):
can only represent one state at a time. If you
have two cubits in superposition, that can represent four states
at one time. So does that mean three cubits is
going to be six states? No, No, three cubits would
be eight states. So one cubit can be two states,
Two cubits can be four states. Three cubits can be

(38:18):
eight states. That means there's eight possible values of the
three cubits, and I'll give them to you right now
so you can see that I'm right. You've got zero zero, zero, zero, zero,
one zero, one zero zero, one one one zero zero,
one zero, one one, one zero, and one one one,

(38:39):
So eight potential values. Every time you had a cubit
you have to you have you end up going with
two to the power of number of cubits you have
for potential states. So in other words, with three cubits
you have two to the power of three. That's eight.
That's how many potential states you could represent. Right now.

(38:59):
I B. M has a prototype quantum computer that has
fifty cubits, and that's a prototype. It's not one that's
rolled out to for anyone to in general to use,
but they do have it. So that means you can
represent two to the fiftieth power in states. So that's
in case you wanted to know. If you knock that
out too, to the fiftieth power, that is one quadrillion,

(39:22):
one hundred twenty five trillion, eight hundred ninety nine billion,
nine hundred six million, eight hundred forty two thousand, six
hundred twenty four states. It's a lot of potential states.
And if that's not enough, if you build a sixty
cubit computer, you just add ten more cubits, you'd have

(39:43):
one capable of representing one thousand quadrillion states. It's insane.
In IBM announced it would make a five cubit computer
available for people to run calculations and experiments on using
a cloud based interface. Uh. This is necessary because in
order to create a quantum computer, you have to take

(40:04):
a really special, extreme precautions to not just create the
quantum state, but to preserve it. So how special am
I talking about? Well, the quantum computers that IBM uses
are cooled to ten millie kelvin in other words, or
fifteen millie kelvin, depending upon which source I was looking at.
Both of them came from IBM, but once at fifteen

(40:26):
and one said ten millie kelvin is incredibly tiny. You're
talking about a fraction above absolute zero. Absolute zero is
the point at which there is no molecular movement, which
is quote unquote colder than space itself. To achieve this,
IBM has to use liquid nitrogen to get the computer
down to a low temperature, and then liquid helium to

(40:48):
get it to an even more insanely low temperature. And
what did IBM used to create the cubits? Did they
use electrons or photons? Nope, they created what they called
artificial atoms. They used a superconducting Josephson junction. What. Well,
it's a superconductor that's coupled to a second superconductor over

(41:10):
a weak link. And I really wish I could go
into more detail and explain how this works, but frankly,
it goes well beyond my understanding, and I feel I
would need to take a college course to get a
handle on it in order to explain it properly. So
I'm not going to try because I'm afraid that if
I did, I would mis explain it to the point

(41:30):
where I would just be giving completely wrong information. Suffice
to say, it's a man made component on a microchip
that's paired with a microwave resonator. The microwave resonator is
what is used to communicate with the cubits, and it's
housed in this crazy looking metal contraption that reminds me
of a super fancy espresso machine, and that in turn

(41:52):
is encased in a cylinder that is a giant refrigerator
to cool it down to these insane low temperatures. Now,
to make it more complicated, if you were to interfere
with this computer in any way, and that could be
electromagnetic interference, it could be heat, it could be motion.
It's very sensitive, you would cause the quantum states to

(42:12):
collapse and decohere, which would turn your expensive quantum computer
into a pretty pathetic excuse for a classical computer until
you can repair the quantum states, and typically you have
a very short amount of time on the order of
milliseconds to complete your operations before either the error rates
get out of hand, which makes all the results look

(42:34):
like they were truly random as opposed to probabilistic, or
the system itself will collapse. The impracticalities of quantum computing
mean that only a few select organizations are ever likely
going to have an actual quantum computer. They're just too
complicated and too sensitive for the general person to have. However,
if they follow the the methodology of IBM and make

(42:56):
it available for other people to use through cloud based
systems where you know you're able to control the quantum computer,
you're just doing it remotely through an interface that they've designed,
then they can make quantum computing more accessible. You won't
own one, but you will be able to access one.
It's pretty crazy, really. Uh. The IBM methodology is called

(43:18):
IBM Q. You can actually go and join that program.
If you want to learn how to program quantum computers,
you can use IBM Q to do it. They have
guides on how to program. They have a very simple interface, UH,
so that you can learn how to program on the
five cubit machine. They also have access to a sixteen
cubit machine through this system, so you can start designing

(43:40):
uh programs to run on a quantum computer. If you
want to check that out. It's frankly, it's beyond my
capabilities to actually do this, at least with my current
level of understanding. But then I'm not really a programmer.
So the program is out there, should definitely take a
look into it and see if you're interested. Well, in
ten you could have access to a live cubit computer

(44:01):
that would give you the potential to have a superposition
of thirty two states simultaneously. So when you encode a
problem onto a quantum machine, what is actually happening. You're
applying a phase to each of those states. So you
can think of the phase like a wave. Some phases
will amplify others and some phases will cancel out others.

(44:25):
This is just like a wave and how waves work
when they encounter other waves. So, for example, if you
have noise canceling headphones, those work by producing sound waves
that are out of phase with the sounds you're surrounded by.
So if you have a perfect tone of a certain frequency,
the sound wave visualization will be one of those lovely

(44:46):
curves has a regular hills and valleys that rise and
fall at a perfect curve and in a particular frequency
that's depended upon whatever the tone is, and it'll look gorgeous. Now,
if you were to produce a second tone where the
sound wave has its peak at the same point on

(45:09):
that wave form that the first wave the first tone
has its valley. So the highest point on your second
tone matches with the lowest point on your first tone,
and vice versa, and they are exactly the same amplitude
and same frequency. They'll cancel each other out. It will
be as if you can't there's no noise at all,
because these two sound waves cancel each other out, and

(45:30):
it's it's as if there's nothing there. That's how noise
cancelation headphones work. They have a microphone that takes an
all incoming sound and then they generate a sound in
the headphones that is out of phase with the sounds
that are around you. They cancel it out. It's not
just muffling sound, it's canceling it by generating this out

(45:53):
of phase sound wave. It's kind of interesting, Well, quantum
computers are doing the same sort of thing with there
the various represented states of the quantum state, like all
those potential combinations of zeros and ones. So the problem
you encode onto the cubits applies those phases, and as

(46:14):
long as you have enough cubits to handle the problem
you're trying to solve. Everything should work out pretty well.
Some answers get amplified, some get canceled out, and you'll
arrive it's your solution, or it's a little more accurate
to say you'll arrive at a probabilistic distribution of solutions.
So better solutions will occupy a higher percentage of probability

(46:34):
than not so good answers. So you can think of
it as like each answers on a pillar, and the
most likely answer is on the highest pillar and the
least likely answer is on the lowest pillar. Does that
mean that the answer is always the right answer is
always going to be the one that's on the highest pillar. No,
that's not how probability works. It's likely, but it's not

(46:55):
always going to happen. That's where you can run into errors. So,
like I said, you're gonna have to look at those
error rates, quantum engineers are gonna have to keep a
close eye on error rates. If we are able to
build more powerful quantum computers, that's great, but if error
rates are high, we can't trust the results we get.
And the more operations you try to run in sequence,

(47:18):
the more opportunities you have for error rates to have
an effect on your results, until again, your probabilistic results
will start to look more like randomized data. Now I've
talked a bit about the sorts of problems quantum computers
can tackle the theoretical problems, but that's mostly in the
thought experiment world. What could quantum computers do in the

(47:40):
real world. Well, i'll tell you right after we come
back from this break for our sponsor. All Right, so
you got your quantum computer. What the heck are you
gonna do with it? Well, one thing you could do
is follow Richard Feynman's suggestion back in the early eighties

(48:02):
and use your quantum computer to simulate the evolution of
quantum states. Actually, simulations in general would be a really
useful application of quantum computers, because, unlike a classical computer,
a quantum computer with a sufficient number of cubits remains
undaunted by the exponential difficulties those simulations pose. So take

(48:23):
chemistry for example. If you want to simulate chemistry down
to the molecular level and you want to work with
long chain polymers, that gets really complicated very quickly because
you've got all these interactions going on at the sub
atomic level that you have to account for. So electrons,
for example, are negatively charged, and they repel one another

(48:44):
because like charge repels like, but they also will be
attracted to the nuclei of the atoms because the nuclei
contained protons those have a positive charge and opposite charges attract.
So you've got these really complex interactions that are going
on at the molecular level, and it gets even more
complicated every time you add another atom to the molecule chain.

(49:07):
And it's that complexity that makes simulating molecules such a
huge challenge for classical computers. In a presentation at think,
an IBM researcher named Talia Gershon, who was part of
the Science slam as well, talked about iron sulfide and
modeling an iron sulfide molecule, and she said that the
largest iron sulfide molecule that the most powerful classical computers

(49:31):
can simulate right now would be a molecule that had
four iron atoms and four sulfur atoms. That would be
a very small iron sulfide molecule. But you couldn't go
bigger than that because the classical computers just couldn't handle
all of those sub atomic interactions accurately. Uh, that's a
severe limitation. If we could shed that limitation, we could

(49:55):
run simulations and all sorts of chemical compounds, and we
could potentially learn the properties of those compounds and think
of potential uses for those compounds. This could revolutionize multiple industries,
a material science, a medicine, those two. In particular chemistry
in general, the chemists could simulate the properties of a

(50:15):
theoretical drug long before ever moving to clinical trials, perhaps
eliminating false leads and saving vast amounts of time and efforts. So,
in other words, you could, based upon your knowledge, create
simulations of various molecules to see how they would play
out in various scenarios, and anything that looked promising, you
could then go forth and try and synthesize and move

(50:39):
forward with clinical trials or at least you know the
earliest stages of testing. That way and narrow down the
limitless possibilities much faster and uh potentially make much more
effective medicine. Arvin Krishna, who's an s VP senior vice
president over at IBM, also mentioned that quantum computing could
be used for financial risk analysis. I imagine it would

(51:02):
also be good for running other types of simulations, ones
that classically are really difficult to manage. For example, it
could be really useful for weather forecasting. That's similar to
the traveling salesman problem I mentioned earlier. Quantum computers could
also be used to help plot out the most ideal
travel routes, not just for a single vehicle, but a
fleet of them. That would be useful in multiple industries,

(51:23):
from transportation to shipping. More efficient travel means fewer delays,
which in turn means cost savings, not to mention fuel conservation.
So you might first think that shaving some miles or
minutes off of travel is a trivial use of so
powerful a computing device, But when you start to think
of the ripple effects the things that that implies, you

(51:44):
start to see the bigger picture. Now I mentioned weather
forecasting that is a really challenging science. Actually, there are
a lot of factors that impact whether you may have
heard my podcast about weather forecasting and how insanely difficult
it is. You've got these big components of weather that
we're all familiar with, things like temperature, humidity, air pressure,

(52:04):
that kind of thing. But there are also other factors
that influence weather patterns, like geography. The topography of the
area you live in affects weather, how it plays out,
the presence of air pollution. Other variables can all affect weather,
and there's so many different variables that shape the weather,
and those variables can have an effect on other variables

(52:25):
that in turn can have an effect on other variables.
In other words, there becomes the sort of domino effect
that can happen in ways that are very difficult to predict.
Simulating the weather with enough data points to ensure precision
is really difficult. Classical computers struggle with this. We use
a lot of supercomputers to crunch the numbers now, and
even then we have to make tough choices. We have

(52:47):
to make allowances for this. So, for example, you could
create a weather model that has a really high resolution,
but it covers a relatively small region. Or you can
have a weather model that covers a much large arger
region but has much lower resolution, so you have lower
amounts of accuracy within that larger model. Uh you also

(53:09):
can have models that predict weather out further into the
future than others, but again with a compromise to either
the size or the resolution or both, So quantum computers
might allow for unprecedented scaling of these weather models, perhaps
one day even leading us to the gold mine, which
would be a global weather model that has high resolution

(53:31):
for any point along the Earth, or at least any
point in those regions where we have enough reliable weather
sensors to provide the data points necessary to create the
simulation in the first place. Now, one thing that I
mentioned earlier that quantum computers would definitely change is how
we protect information. Using Shore's algorithm and a quantum computer
with a sufficient number of cubits, you could determine the

(53:53):
prime number factors of any large number relatively quickly, which
puts all of our encryption at risk. Well not all
of it, but but but the vast majority of our
of the way we encrypt things would be at risk.
And I'm not just talking encryption for stuff like email
or online shopping. Credit Card transactions would be at risk.
They rely on large number factoring, so that would be

(54:16):
a problem, as would numerous otherwise secure data exchanges. They
would also be at risk. All the secrets would no
longer be secret, so this would be like someone creating
the perfect skeleton key that fits all the locks in
the world, and at that point, there's not really a
reason to use a lock because you already know someone's
out there with a key that's going to open it.
So you've got to figure out a different way to

(54:37):
lock stuff. So rather than give up, it just means
we have to come up with a post quantum encryption strategy. Now.
I mentioned that in the episodes are recorded about the
IBM Science Slam to Chilia, Boscuini mentioned a lattice based
cryptography strategy, which would use a plotted point within a
realm of dimensions multiple dimensions as many as like a

(54:59):
hundred dimensions as an alternative to factoring large numbers. I
can only sort of pretend like I understand what she's
talking about, because it goes way over my head. But
according to Buscini, this could pose a problem so difficult
that even a quantum computer might have trouble working it
out and thus end up securing our data. We would
just be switching our encryption strategies. So quantum computers do

(55:22):
have the potential to make a tremendous impact on our world.
Though it is important again to note that they aren't
going to replace classical computers for all tasks. Quantum computers
are ideally suited for a subset of computational problems, including
ones that are really hard for classical computers to tackle.
But there are other tasks that classical computers will be

(55:43):
just as good at, or even better at, than quantum computers.
So I don't mean to suggest that in twenty years
everyone's going to have a quantum computer sitting on their
work desk, unless you have to work in a quantum
computer laboratory, in which case you might because you might
have to do repairs or something. Anyway, that wraps up
this quantum computing one oh one episode. I hope you

(56:04):
guys enjoyed it. If you have any suggestions for future
episodes of tech Stuff, make sure you write me and
let me know what those are. It could be a technology,
it could be a person, could be a company, or
maybe it's a suggestion for an interview or a guest
co host. I am happy to hear all of those.
Just SIMI a message. The email address is tech Stuff
at how stuff works dot com, or drop me a

(56:24):
line on Facebook or Twitter. The handle for both of
those is text stuff h s W. Check out our
Instagram account to see some behind the scenes stuff that
we post occasionally. And remember you can watch me record
the regular episodes live on Wednesdays and Fridays over at
twitch dot tv slash tech stuff. We've got a chat
room in there. You can join in there and and

(56:44):
chat with me. I'll be happy to talk with you
during breaks, and I will talk to you guys again
really soon. For more on this and thousands of other topics,
visit House staff sat com.