0:01

Here's the thing, science isn't only

0:03

for the PhDs of the

0:05

world, it's for everyone. On

0:07

the Shortwave podcast, we dig

0:09

into the latest research with a humorous

0:11

touch. In under 15 minutes,

0:13

it's serious science, even if it

0:15

doesn't sound like it. Listen now

0:17

to the Shortwave podcast from NPR.

0:31

I'm Steve Strogatz. And I'm

0:33

Jana Levin. And this is The

0:35

Joy of Why, a podcast from

0:37

Quantum Magazine exploring some of the biggest

0:40

unanswered questions in math and science

0:42

today. Here we are. Hey,

0:44

Jana. Hey, Steve. I've got

0:46

something really fun queued up for you

0:48

today. Good. I'm looking forward to hearing

0:50

about it. OK. Now, I think this should

0:52

be kind of in your wheelhouse. It's about

0:54

gravity. Hmm. That's definitely within my wheelhouse. And

0:57

now let me add one more word, quantum

0:59

gravity. Yeah. And now it's in no one's wheelhouse. That's

1:03

interesting, isn't it? Because it's such

1:05

a hard open problem in physics. So

1:08

I had the chance to speak

1:10

to a great young physicist named

1:12

Monica Schleyer -Smith. She's at Stanford. And

1:15

she is taking an approach I've never

1:17

heard of anywhere else, which is to

1:19

try to build a kind

1:21

of toy model of quantum gravity

1:24

in the laboratory. Funny thing,

1:26

right? I mean, you think of quantum

1:28

gravity as purely theoretical pencil and paper

1:30

stuff. Yeah, absolutely. It

1:32

seems maximally hard. Maximally

1:34

hard, right? Do you ever

1:36

hear this idea that gravity

1:38

might emerge from entanglement? Oh,

1:41

yes. It's one of my favorites. Is

1:43

it? Yeah. I find

1:45

it really intriguing. I kind of

1:47

think of it as the entanglement as

1:49

like threads at the quantum level

1:51

and it embroiders a world that from

1:53

afar looks like it's smooth and

1:56

continuous and you look up close

1:58

and you realize it's really these entangled threads.

2:01

Very poetic. I like it. Well, that

2:03

is sort of the spirit of

2:05

what we're doing in this episode with

2:07

Monica. She's going to talk to

2:10

us about the ways that

2:12

she tries to entangle thousands

2:14

of atoms. That

2:16

she has maintained a very low temperature

2:18

so that they can express their

2:20

quantum mechanical nature and get entangled. But

2:23

it's many body quantum entanglement we're

2:25

talking about. entangled with each other as opposed

2:27

to just pairs? Right. That's the new

2:29

wrinkle here. I'm only used to

2:31

the idea of, you know, you

2:33

hear a lot about entangling two

2:35

atoms or something like that in

2:37

the old, like Einstein, Podolski, Rosen. thought

2:40

experiment and then later recreations of

2:42

that in the lab. This

2:44

is many body entanglement, thousands of

2:46

atoms. And I think she's trying

2:48

to stitch together some kind of

2:50

fabric of space and time like

2:52

you just described. Wow. I mean,

2:55

I'm not sure how she would get

2:57

gravity out of that quite yet, but that's

2:59

fascinating. I mean, I always

3:01

thought there was a kind of monogamy of

3:03

entanglement. So if one particle was maximally entangled

3:05

with another, It had to

3:07

be monogamous. It could not also be entangled

3:10

with a third partner. Well, that's

3:12

interesting. I've never heard that idea.

3:14

So maybe each one's partially entangled

3:16

with another, right?

3:18

So they're not maximally entangled one

3:20

to one. Yeah. It's a kind

3:22

of polyamory of entanglement. Well,

3:25

I knew you would get interested in

3:27

this. And so I think we should

3:29

just hear from Stanford physicist Monica Schlier -Smith.

3:33

Hey, Monica. Welcome to the show.

3:35

Thank you. Before we get rolling on

3:38

the question of what you've been doing

3:40

in quantum physics experiments, I read somewhere

3:42

that you got an early start even

3:44

as a high school student doing nanotechnology

3:46

in a lab. So yeah, I

3:48

was very fortunate as a high

3:50

school student to get to do summer

3:52

research internships at a company called

3:54

the MITRE Corporation. I wasn't yet working

3:56

in the lab, but I was

3:58

getting to really grapple with forefront issues

4:00

of cutting edge research and nanotechnology.

4:03

And that was really remarkable being a

4:05

16 year old getting to read

4:07

scientific papers and to be in a

4:09

research group. Amazing. When you say you

4:11

weren't in the lab, did they have

4:13

you doing computer simulations or what kind

4:15

of thing? That's right. I was

4:17

doing computer simulations. I

4:19

was developing some ideas that

4:21

actually led to a patent. We

4:23

were able to collaborate with a group

4:25

at Penn State that had the expertise

4:27

and was able to take it to

4:29

the next level and led to actually

4:31

publishable results. And after I left, they

4:34

actually started a lab based on some

4:36

of the ideas we'd been brainstorming about

4:38

when I was there. What a

4:40

great start to a scientific life. Should

4:42

we picture you as a little kid

4:44

with scientific parents or going out in

4:46

the woods looking at bugs or what

4:48

was your deep background? Definitely

4:51

going out into the woods exploring. My

4:53

mother actually, her background

4:55

was in linguistics and really

4:57

more in the humanities,

4:59

but she always wished she had

5:01

gone into science. And so that

5:03

was a very strong influence. And I

5:05

have an older brother who took the first

5:07

steps in studying physics in college. And

5:10

I guess I was inspired and followed in

5:12

those footsteps. quite a family. Well,

5:15

you're part of this

5:17

giant enterprise of modern physics,

5:19

and there's this so -called

5:21

standard model. I mean,

5:23

it's a super successful theory, but even

5:25

the most ardent proponents of the standard

5:27

model would agree there are a few

5:29

things that have to be filled in. Can you

5:31

tell us a little about what is that model? Yeah,

5:34

so the standard model describes

5:36

a wide range of particles

5:38

that make up matter in

5:40

our universe, electrons, protons, and

5:42

their constituent quarks, photons. And

5:45

part of the standard model,

5:47

for example, is also that some

5:49

of these particles are responsible

5:51

for electric forces between charged particles,

5:53

electrons, and protons. There

5:55

are additional particles that are responsible

5:57

actually for mediating those forces. A

5:59

photon is actually responsible for mediating

6:01

these electric forces, kind of bouncing

6:03

between the charged particles. that are

6:05

interacting. One thing that's

6:08

missing actually in the standard model,

6:10

one significant omission is a

6:12

particle that mediates the force of gravity, right?

6:15

So in the same way that

6:17

charged particles can attract or repel

6:19

depending on their charges, gravity seems

6:21

On the face of it, very

6:23

analogous. The mass of an object

6:25

is kind of the equivalent to

6:28

the charge in the case of

6:30

electric forces. The mass determines how

6:32

strong the attractive forces in gravity. But

6:34

the standard model doesn't have that equivalent

6:36

of the photon for electromagnetism. It doesn't

6:38

have something like a graviton that would

6:40

mediate the gravitational forces. It's possible to

6:42

have this theory, the standard model, that

6:45

is extraordinarily well -tested but has this

6:47

sort of glaring omission of gravity. Right.

6:49

So you've mentioned it doesn't have the

6:51

counterpart of the photon, the graviton. I

6:53

mean, we talk about gravitons, but they're

6:56

not part of the standard model. I

6:58

hear there are a few

7:01

other things like neutrinos, these tiny

7:03

neutral particles shouldn't have any

7:05

mass, but they do. Yeah,

7:07

and there are big mysteries in our

7:09

universe also about what we call dark matter

7:11

and dark energy. you know, 70 %

7:13

of the energy that should be there

7:15

is missing in the form of this

7:17

dark energy that we can't account for.

7:19

So lots of big mysteries in the

7:22

universe, even though the standard model is

7:24

experimentally extraordinarily well tested. Okay. But you

7:26

put your finger specifically on this missing

7:28

part of gravity in the standard model. And

7:30

so that's what we'll be talking about

7:32

mostly. So given that gravity is

7:34

not in the standard model, even though

7:36

we know it's a real important force,

7:38

it's keeping us both in our seats

7:40

at the moment. How is

7:42

that? hurting our current understanding of the

7:44

universe. We do have a very

7:46

excellent theory of gravity. Newton's

7:49

theory, or then if we want to get

7:51

fancy here, Einstein's general relativity. That's

7:53

right. We do have an excellent

7:55

theory. Einstein's general relativity is

7:58

also extraordinarily well tested. If

8:00

I go back to the analogy

8:02

with electromagnetism... the sort of

8:04

classical pictures that electromagnetic forces, they're

8:06

mediated by electromagnetic waves or

8:08

light. And in gravity,

8:10

we've by now even detected

8:12

gravitational waves, right? So in

8:14

its own right, gravity is

8:16

also very well understood and

8:18

tested. At some

8:20

level, the challenge is that

8:22

the microscopic description is really kind

8:24

of quantum mechanical, quantum mechanics is

8:26

a great theory for describing systems

8:29

at very small scales. But gravity

8:31

is a theory that works very well

8:33

in the regime of massive objects, you

8:36

know, the motion of planets. And

8:38

these are tested in very different regimes.

8:40

It's hard to get into a

8:42

regime where actually both gravity and quantum

8:44

mechanics matter, it's mostly so

8:46

far in thought experiments that

8:48

we realize we don't have a

8:50

unified theory and that there's

8:53

something missing when we can't connect

8:55

the rules of quantum mechanics with the

8:57

rules of gravity. That's interesting,

8:59

this last point that you

9:01

just raised, because there are parts

9:03

of relativity that play nicely

9:05

with quantum theory, right? Like we

9:08

do have special relativity. That's

9:10

right. And again, I think it's partly

9:12

this issue that one theory or

9:14

the other applies well, or one

9:16

can put in the minimal ingredients,

9:18

let's say special relativity of gravity

9:20

and combine that with quantum mechanics,

9:22

but somehow a full unified theory

9:24

is still missing. I mean, just to

9:26

give one other example that I find kind

9:28

of puzzling, in gravity, space and

9:30

time are treated on an equal footing. In

9:33

quantum mechanics, we actually don't treat them

9:35

on an equal footing. Systems evolve

9:38

in time and space is thought of

9:40

completely separately. And so

9:42

somehow there are these two inconsistent

9:44

ways of thinking about the universe. And

9:46

one has to start doing thought

9:48

experiments about things like what happens

9:51

to information that falls into a

9:53

black hole to start to realize

9:55

that actually to really fully understand

9:57

our universe, we need to reconcile

9:59

them. Hmm. I'm glad you

10:01

put your finger on space and time, because

10:03

that's really what we do want to be

10:05

talking about here. I mean, they're all linked

10:07

up, aren't they? Space, time, gravity, and then

10:09

this other whole story of quantum

10:11

mechanics. So let's talk about the

10:13

possibility that space and time

10:15

might not be as fundamental as we used to

10:17

think. Yeah. And so one

10:19

of the remarkable ideas that's

10:22

emerged from theorists who think hard

10:24

about this problem of reconciling

10:26

quantum mechanics and gravity is the

10:28

notion that perhaps

10:30

the fundamental building blocks

10:32

of gravity really

10:34

are quantum mechanical. A

10:37

number of years ago, I found a quote from,

10:39

you might say, the father of the atom,

10:41

Democritus, right? He was the Greek philosopher

10:44

who recognized that matter is

10:46

not just some smooth, continuous

10:48

thing. It has actually

10:50

fundamental building blocks that are

10:52

atoms and molecules. And

10:54

then he made this point

10:56

that phenomena such as

10:59

you know hot and cold, sweet

11:01

or bitter. taste, temperature, colors

11:03

emerge from the microscopic configurations of

11:05

individual atoms or molecules. I don't need

11:07

to think about the positions of

11:09

all the individual atoms to look at

11:12

an object and say it's red,

11:14

right? And so color is this kind

11:16

of emergent phenomenon. So

11:18

the question that has been explored

11:20

in recent years in this

11:22

effort to unify quantum mechanics and

11:24

gravity is, could it be

11:26

that gravity is actually also an

11:29

emergent phenomenon? So the microscopic

11:31

constituents are really quantum mechanical, and

11:33

gravity emerges as this sort simplified,

11:36

smooth description of what

11:38

fundamentally is really some complex

11:40

interacting quantum system. And

11:43

I find that idea

11:45

fascinating, and how might

11:47

gravity emerge from quantum mechanics? The

11:49

connection that's conjectured is

11:51

a phenomenon called entanglement. Go

11:54

on. I want to hear more because it

11:56

is incredibly fascinating. The first time I heard it,

11:58

my mind was blown. Tell us. So

12:00

entanglement is the idea that

12:02

I can store information, not

12:05

just in individual bits

12:07

or particles, but actually

12:09

in correlations. So, you know,

12:11

in your computer. you have information that's stored

12:13

in bits that are in like a one

12:15

state or a zero state, but that information

12:17

that's really stored locally in an individual bit. So

12:20

the quantum analog of a bit, we

12:22

call it a qubit, and it's

12:24

possible to have information that's not just stored in

12:26

a single qubit. If you look at the

12:28

state of a single qubit, it looks completely random.

12:31

In fact, randomness is an inherent aspect of

12:33

quantum mechanics. But if you look at the

12:35

states of two of these qubits, you would

12:37

find they're always either both 1 or they're

12:39

always both 0, even though each one individually

12:41

looks random. And so there is actually some

12:43

order in the randomness, some information that can

12:45

be stored in a way that you can

12:48

only access if you look at both of

12:50

these qubits. So this idea of correlations and

12:52

information that are sort of hidden in this

12:54

randomness, that's this notion of entanglement. And

12:56

one of the sort of challenges

12:58

this brings up is that

13:01

describing a quantum

13:03

system It's actually much more complex

13:05

than describing the bits in your classical

13:07

computer, because you need to keep track not

13:09

just of the states of the individual

13:11

qubits, but of all of these correlations between

13:13

them. So sometimes

13:15

I like to sort of visualize a

13:17

graph where I have my row

13:19

of these cubits. But then I want

13:21

to sort of draw some lines

13:23

that indicate something about the structure of

13:26

which ones are correlated with which

13:28

ones. And that's still an overly simplified

13:30

description. But roughly speaking, these correlations

13:32

I can visualize as some connections between

13:34

the cubits. And now the idea

13:36

is that perhaps actually this notion of

13:38

gravity being an emergent phenomenon,

13:40

the idea is actually describing

13:42

those correlations. And I

13:44

kind of think of it as there's

13:46

this one additional dimension that allows

13:48

me to capture extra information that describes

13:50

the structure of correlations. There's

13:53

some mapping from the

13:55

quantum mechanical system to actually

13:57

a geometrical description in which the

13:59

distance between the qubits says

14:01

something about how strongly they're correlated.

14:04

This notion has also

14:06

been given the name of holographic duality.

14:09

So why holographic? A hologram is

14:11

something that has two dimensions, but

14:13

actually it looks like a three -dimensional

14:15

image, right? It has this sort of

14:17

additional dimension. So there's this

14:19

notion that somehow once one accounts

14:21

for the entanglement between all these

14:23

degrees of freedom, gravity may emerge

14:25

as a description of

14:28

those microscopic quantum building blocks, a

14:30

sort of smooth macroscopic description in

14:32

terms of space -time curvature and geometry.

14:34

Okay, so that is a lot.

14:37

a lot, a lot going on

14:39

there. That's

14:41

fine because you've given us a lot to

14:43

chew on now. You said this really deep,

14:45

interesting thing that if I can paraphrase it

14:48

and correct me if I'm not hearing you

14:50

right, it's sort of

14:52

like saying distance is an illusion. What

14:55

really is meaningful is correlation,

14:58

right? That's sort of the idea that things that

15:00

look like they're a certain distance apart That's

15:03

our macroscopic way as big

15:05

creatures of thinking about what

15:07

microscopically is about strong correlations

15:09

or maybe weak correlations. Yeah, exactly.

15:11

Like a long distance would sort of

15:14

correspond to a weaker correlation, roughly speaking,

15:16

exactly. Okay. So we'll have

15:18

to come back to that, this idea

15:20

that space and distance... Is really

15:22

just an emergent way of talking about

15:24

what's really going on under the

15:26

hood, which is correlations of different strengths.

15:28

But so you spoke about that

15:31

there can be information in the relationship

15:33

between two things that are otherwise completely

15:35

random on their own. Right.

15:37

I kind of like to use the analogy of

15:39

a coin toss, right? And so

15:42

like imagine I'm here and

15:44

I'm tossing. coins and every time

15:46

I toss one, you'll also toss a coin.

15:48

And when we look at the outcomes

15:50

of those coin tosses, I'll see something completely

15:52

random. You'll see a random sequence of

15:54

heads and tails. And classically, that's

15:56

all there is to it. And there's

15:58

no correlation. But quantum mechanically, we could have

16:00

a situation where every time I get

16:02

heads, you get tails. And every time I

16:04

get tails, you get heads, despite

16:07

the fact that I'm here in California

16:09

and you're in Ithaca. Yeah,

16:11

exactly. And so that would be very

16:13

weird, right? Right, especially where I'm

16:15

far enough away that you couldn't

16:17

possibly get a signal to me fast

16:19

enough to influence me. Exactly.

16:22

There were more and more experiments over

16:24

the past 20 years or so trying

16:26

to really make sure that we verified

16:28

entanglement in the setting where these two

16:30

measurements were far enough apart that there

16:32

couldn't be any information traveling between them

16:35

and things like that. Okay, so as

16:37

it started, it was a very theoretical

16:39

idea going back to the 1930s

16:41

or something, right, from Einstein and Podolski

16:43

and Rosen and Schrodinger. Exactly. But

16:45

now, fast -forwarding to almost a century

16:47

later, it's not obvious to

16:49

me how you would maintain the

16:51

entanglement over great distances. Does it

16:53

take tremendous care to keep them

16:55

entangled? It's tremendously challenging and to

16:57

bridge long distances. There are different

16:59

choices you could make of photons because

17:01

they travel at the speed of

17:03

light across long distances. But even

17:05

so, There's some possibility that

17:07

the photon, if it's sent through an optical

17:09

fiber, that it's lost along the way,

17:12

or also if it's sent through free space,

17:14

there's still some chance it'll get absorbed

17:16

along the way. Sometimes there are tricks where

17:18

you can do what's called heralding that

17:20

maybe you don't succeed every time, but

17:22

there's a way to know actually

17:24

whether you successfully created an entangled state.

17:26

Okay, but now it sounds like you

17:28

and your students are doing this. Every

17:30

day now so maybe you should tell

17:32

us what are you entangling in my

17:34

lab? We work on smaller length scales,

17:37

you know the particles that we

17:39

entangle our atoms and an atom is

17:41

an angstrom scale object ordinarily you

17:43

would think that if I have two

17:45

atoms that are let's say a

17:47

millimeter apart they won't interact they won't

17:49

become correlated but that's actually a

17:51

length scale where we are able to

17:54

generate entanglement and the way that

17:56

we do it is in fact actually

17:58

using light I use this notion of

18:00

mediating interactions we use a very

18:02

engineered setup in the lab where a

18:04

photon can bounce between two atoms

18:06

or between two clouds of atoms and

18:08

introduce correlations and

18:10

entanglement between them. And that's

18:12

not the only way that one can generate

18:15

entanglement among atoms. I'll focus on this one

18:17

because one of the nice things about photons

18:19

is they can quickly bridge long distances

18:21

and they can give a lot

18:23

of flexibility. Naively, you would think what

18:26

will naturally happen is atoms will

18:28

maybe bump into other atoms that are

18:30

near them and you'll sort of

18:32

generate strong correlations between neighboring atoms. And

18:34

what we like to be able to do is have some network

18:36

where we can actually control the

18:38

structure of correlations and decide by

18:40

some knobs in the experiment, the

18:42

atoms that are most strongly correlated,

18:44

that's a way of using photons

18:46

to program the

18:48

graph of correlations in our

18:50

case, an array of clouds of

18:53

atoms. So I do want to

18:55

hear about the clouds of atoms

18:57

and the programmable networks that you're

18:59

building or engineering. But can

19:01

you give us a oral picture?

19:03

Like if we were standing behind you

19:05

looking over your shoulder, what

19:07

would we see? And

19:09

even lab in my research group, you would

19:11

see something like two to three optical

19:13

tables. So each of these something like four

19:15

foot by eight foot, maybe even a

19:17

bit bigger. There's usually one table that has

19:19

a bunch of lasers, because I mentioned

19:21

we need laser light as our tool for

19:23

manipulating atoms. So you'd see

19:25

these lasers, you would see lots

19:27

and lots of mirrors and various

19:29

other optical elements to steer the

19:32

lasers into the right places. And

19:34

then all of these laser beams

19:36

get steered into optical fibers going

19:38

from one table to another

19:40

table, which carry that

19:42

light to where our science

19:44

experiments actually happen. That second table

19:46

has on it an ultra

19:48

high vacuum chamber, which we

19:50

need in order

19:52

to have particular atoms in

19:54

one of our labs, it's rubidium atoms

19:56

that are well isolated from anything

19:59

else in the lab. We want

20:01

to be operating in an ultra

20:03

high vacuum environment where I can

20:05

just create a cloud. or

20:07

a few clouds of atoms that

20:09

are at very low temperature, and

20:11

that are essentially suspended by laser

20:13

beams in the middle of this vacuum

20:15

chamber. These rubidium atoms are in

20:17

what you're calling high vacuum soap, meaning

20:19

they're not bumping into any oxygen

20:21

or nitrogen. There's no air in there.

20:24

They're just rubidium atoms, which I don't

20:26

even really know how to think about

20:28

rubidium. I've heard of rubidium. If you

20:30

remember your periodic table. Yeah, no, I

20:32

don't. Tell me. I'll just say

20:34

it's in the first column. And

20:37

what that means is basically there's

20:39

one valence electron, so one outer

20:41

electron that is relatively well isolated

20:43

from all of the other electrons.

20:45

And that actually turns out to

20:48

be convenient for making it a

20:50

relatively simple atom to control and

20:52

manipulate with lasers. I see.

20:54

And why do you want to have a cloud of them?

20:57

So I'll say that I might actually

20:59

prefer not to have a cloud,

21:01

but we work with a cloud for

21:03

the class of experiments I described

21:05

where we use light photons as our

21:07

means of generating some network of

21:09

interactions between atoms. We can

21:12

actually generate stronger interactions if we

21:14

use many atoms rather than

21:16

just one. If I go back

21:18

to the wave picture of light,

21:20

there's constructive interference. Photon

21:22

bouncing off one atom and hitting another

21:24

atom to make them interact if I

21:26

have many atoms the waves that they

21:28

scatter can interfere Constructively and that can

21:31

actually enhance the strength of interaction cool

21:33

and so at least for kind of

21:35

first experiments It's been convenient for us

21:37

to work with let's say clouds where

21:39

each cloud has a thousand atoms and

21:41

then we have an array of such

21:43

clouds but there is a path that

21:46

we're interested in actually going towards single

21:48

atoms where each atom interacts more strongly.

21:50

Okay, but so for now we could

21:52

think of a cloud of 10 ,000 atoms

21:54

or something, but you mentioned low temperatures,

21:56

so you want to tell us how

21:58

low? Yeah, so typically

22:01

we work with temperatures that are

22:03

on the scale of tens

22:05

of microkelvin, a thousandth of a

22:07

degree above absolute zero would

22:09

be a millikelvin. We're often a

22:11

factor of 20 or 50

22:14

below that in temperature. And

22:16

so that sounds extraordinarily cold. One

22:19

thing to keep in mind is actually

22:21

the room isn't cold. That vacuum chamber

22:23

isn't cold. If you touch it, it's

22:25

at room temperature. But it's just this

22:27

cloud of atoms suspended by laser light

22:29

that we are able to actually bring

22:31

to very low temperature using tricks of

22:33

laser cooling. There's some way of using

22:35

the laser to hit the atoms in

22:37

just the right way that sort of

22:39

knocks the wind out of them. Exactly,

22:41

that's a great analogy. Okay, I see.

22:43

So I'm getting the picture now. You've

22:46

got these clouds, 10 ,000 atoms, you

22:48

get them very cold, not to the

22:50

point where they're a single quantum object

22:52

in the sense of Bose -Einstein condensate, but

22:54

still They're constructively interfering

22:56

in the wave picture enough that it's

22:58

sort of like a strong edge in

23:00

this network of interactions that you're trying

23:02

to build. And maybe I can also

23:04

just add the reason we don't

23:06

need to get them down to this

23:09

state of matter of Bose -Einstein condensation. It

23:11

turns out if the atoms are moving

23:13

around a little bit, that's okay for

23:15

the experiments that we do. There's still

23:17

a sense in which we place all

23:19

of our atoms in a given cloud

23:22

into the same quantum state. So they're

23:24

not all at the same position, but

23:26

what we care about most in our

23:28

experiments is some internal state of the

23:30

atom. So there's that electron we talked

23:32

about. And we control which

23:34

state that electron is in. We

23:36

control which way its spin is pointing.

23:39

And so we have actually very good control over

23:41

the internal states of these atoms. And those

23:43

will all be identical in the given cloud. So

23:45

is the entanglement that you're trying to

23:47

set up at the level of spins

23:49

then? Because I mean, I know in

23:51

the traditional old thought experiments about entanglement,

23:53

they used to frequently talk about the

23:55

spin of. two different particles. Exactly. And

23:57

so the spin is sort of the

23:59

real physical implementation of what I talked

24:01

about before with the heads and tails

24:03

of the coin, right? These two possible

24:06

states of the coin are like a

24:08

spin that points up or down. And

24:10

actually in quantum mechanics, a spin

24:13

that could point anywhere in three

24:15

dimensions. But when we decide

24:17

to do a measurement, we have to

24:19

choose what we call a basis. We

24:21

can measure it as it point up

24:23

or down, as it point right or left.

24:25

But we actually can't determine both of

24:27

those things at the same time. So these

24:29

are examples of incompatible observables. A

24:31

famous example is position and

24:33

momentum of a particle. Two

24:36

different components of the spin, the vertical

24:38

or the horizontal, that would be another

24:40

example. Okay, so now we've

24:42

got the visual of you in

24:44

your lab and the optical table

24:46

and all the lasers and mirrors

24:48

and cables. But then, now that

24:50

you have this ability to entangle

24:52

these clouds, you can make whatever

24:54

networks you want if I'm hearing

24:56

you right. You know, you're doing

24:58

this very fundamental research about what

25:00

in the jargon might be called

25:02

something like many body entanglement. So

25:05

that's sure to be important. Right.

25:07

So having control over entanglement can

25:09

be a resource for making better

25:11

precision measurements when what limits you

25:14

is quantum uncertainty, the sort of

25:16

randomness inherent in quantum mechanics or

25:18

another one quantum computation. The idea

25:20

is if you have a sufficiently

25:22

well controlled quantum system. where you

25:24

can really program in the interactions

25:27

in the same way that you

25:29

program your classical computer, but now

25:31

the building blocks are quantum bits,

25:33

then one place that seems very

25:35

natural to give us an advantage

25:37

is precisely in describing quantum mechanical

25:40

systems, be it the behavior of

25:42

electrons and materials, or be it

25:44

actually problems from chemistry, for example.

25:49

I'm kind of dumbfounded a little bit. I

25:52

mean, you have these many body

25:54

systems. She's somehow managed to entangle

25:56

them in this kind of a

25:58

network. How do I get

26:00

from there to gravity? Why would I

26:02

think it's gravity and not some other

26:04

complex system that emerges from the many

26:07

body problem? That's really a

26:09

good question. I think she is

26:11

looking for signatures of something

26:13

that would be like a discrete

26:15

analog of curvature of a

26:17

continuous space. So networks

26:19

can have properties like curvature

26:21

the way that smooth manifolds

26:23

can have curvature. Amazing.

26:26

So she's trying to make a

26:28

space time. or like a manifold or

26:30

something. It's something like you said

26:32

with your beautiful analogy of the embroidered

26:34

fabric that it might look like

26:36

a nice smooth dress, but you look

26:38

up close, it's a lot of

26:40

threads stitched together. Right. Oh, fascinating. So

26:43

the network itself is an emerging

26:45

space time in some sense. Something like

26:47

that, but it's controversial what she's

26:49

doing. And as she says, It's

26:51

also possible that this won't lead to

26:53

a deeper understanding of gravity, but maybe it

26:55

will help with precision measurements or maybe

26:58

it will help with quantum computing. We're going

27:00

to get right into that after the

27:02

break. Welcome

27:19

back to the Joy of Why. We're

27:21

speaking with Stanford physicist Monica

27:24

Schleyer -Smith, who's been telling us

27:26

about her toy model of gravity.

27:29

So there is a certain amount

27:31

of controversy attached to this idea

27:33

that gravity and space time emerge

27:35

from quantum entanglement. We haven't really

27:37

said that out loud, but maybe

27:39

we should. Right, absolutely. So like,

27:42

even if that doesn't pan out,

27:44

you're not wasting your time, I

27:46

think, in the lab. I like

27:48

to think not, again, it's not

27:50

the only thing I'm working on,

27:52

but also for me, I'm fascinated

27:54

by the idea that gravity in

27:56

our universe might be an emergent

27:58

phenomenon where the building blocks are

28:01

quantum mechanics. But there's also another

28:03

way to think about this whole

28:05

field, which is to say, there

28:07

are certain cases where one sees

28:09

this so -called duality. So there's a

28:11

strongly interacting quantum system that has

28:13

an equivalent description in terms

28:15

of equations that look like gravity. And

28:18

whether or not gravity in our

28:20

universe is a manifestation of quantum mechanics,

28:23

these theoretical tools of taking a

28:25

strongly interacting quantum mechanical system and mapping

28:27

them to a description in terms

28:29

of curved space and gravity, maybe that

28:31

gives you new ways of calculating

28:33

things about the quantum mechanical system or

28:35

new insights into how it will

28:37

behave. Okay, I'm sold. This is good.

28:40

So maybe you should tell me

28:42

then about this kind of toy model

28:44

of gravity. What do these networks

28:46

have to do with quantum gravity and

28:48

spacetime? Yeah, you know, this effort

28:50

at understanding quantum gravity has been almost

28:52

purely the domain of theory until

28:55

recently. And so my thought was

28:57

if gravity might be an emergent phenomenon,

28:59

if curved space might be something that

29:01

can emerge as a natural description of

29:03

quantum correlations, Can we build

29:05

a system where we start to

29:07

see this phenomenon of something that looks

29:09

like curved space emerging as a

29:11

description of the quantum correlations? Before

29:13

we did an experiment, rather

29:16

serendipitously, we began talking with a

29:18

theorist at Princeton named Steve

29:20

Gubser, who actually tragically passed away

29:22

in a climbing accident a

29:24

few years ago. But

29:26

Steve was working on this effort

29:28

to reconcile quantum mechanics and

29:30

gravity. He had developed, I would

29:32

say, a particular version

29:34

of this holographic duality.

29:37

So holographic duality was this notion that

29:39

I have a quantum mechanical system

29:41

that I can think of as kind

29:44

of living on the boundary where

29:46

the higher dimensional space has gravity and

29:48

the gravity gives rise to curvature

29:50

in that space and the distances within

29:52

that higher dimensional space say something

29:54

about correlations in the quantum system on

29:56

the boundary. What was really

29:58

valuable to us about talking to

30:00

Steve Gubser was that he was working

30:03

on a formulation of this holographic

30:05

duality where there's a very nice way

30:07

of actually visualizing bulk geometry. He

30:09

wanted actually a discretized theory. So what

30:11

do I mean by that? He

30:13

wanted there to be a shortest length

30:15

scale in his theory, motivated actually

30:17

by the fact that in our universe

30:19

there's a length scale called the

30:21

plank length, where you expect to start

30:23

running into major problems in reconciling

30:25

quantum mechanics and gravity. And anyway, so

30:27

Steve had this discretized version of

30:29

holographic duality, where the bulk geometry is

30:31

represented by a tree. and the

30:33

boundary where the quantum mechanical system lives,

30:35

it lives on the leaves of

30:37

the tree. Ah, nice. Yeah, so you

30:39

can kind of imagine if you're

30:41

at the trunk of the tree that's

30:43

somewhere in the middle of the

30:45

bulk, somewhere in the middle of the

30:47

part that's described by gravity, and

30:49

then that trunk has two branches, and

30:51

each of those has two more

30:53

branches, and each of those has two

30:55

more branches. In just a few

30:57

steps, the number of leaves actually grows

31:00

exponentially with the distance that you

31:02

go out from the trunk. If you

31:04

think of the tree as kind

31:06

of radiating outwards so that the leaves

31:08

end up on a circle, you

31:10

have this weird thing where the

31:12

circumference of that circle is actually exponentially

31:14

larger than the distance measured and

31:16

how many times you need to branch

31:18

to get to the leaf. So

31:20

we have this graph, and again, I

31:22

kind of think of it as

31:24

I can visualize it as the leaves

31:26

live on the circumference of a

31:29

circle. The circumference is exponentially larger. than

31:31

the diameter instead of being larger

31:33

by a factor of pi. And

31:35

so somehow, like, this isn't just

31:37

a flat disk. It's curved. And it

31:39

actually has negative curvature. Oh, I

31:41

see. You're saying that's why it's curved.

31:43

Because if it were just a

31:45

flat disk, it should only be the

31:47

two pi times the radius. But

31:49

this is way more circumference than that.

31:51

Exactly. If anyone in the audience

31:53

has seen some prints by Escher, so

31:55

like he'll tile a disc with

31:57

fish. And in the middle, there are

31:59

big fish and around the circumference,

32:01

there are lots of really tiny fish.

32:03

And that actually really is a

32:05

representation of this hyperbolic geometry. Yeah.

32:07

And why hyperbolic? I mean, there's also

32:09

this nice picture of like, if you're

32:11

trying to flatten out a rug, it

32:14

may be fine. But if you try

32:16

to flatten out a piece of lettuce,

32:18

or something that looks like a saddle.

32:20

Right, yeah. It keeps popping up, right?

32:22

You can't flatten it out easily, and

32:24

it's because it has too much circumference

32:26

for its distance from the center. So

32:29

that's what negative curved space is

32:31

kind of like. Exactly. And

32:33

so this tree graph is a

32:35

discretized version of this negatively curved space,

32:37

or what is called anti -decider space

32:39

in the context of general relativity

32:41

and gravity. So...

32:44

We kind of asked ourselves, is there

32:46

a quantum mechanical system we can build that

32:48

would have correlations that make it sort

32:51

of look like it lives on the

32:53

leaves of a tree graph? And

32:55

we realized there is a toy model

32:57

where a given site in my array

32:59

can talk to its nearest

33:01

neighbor, its second neighbor, its

33:03

fourth neighbor, its eighth neighbor. But

33:06

it's not a lot of connections that you

33:08

need to build, but they give you an

33:10

efficient way of getting information from one point

33:12

to any other. So we've actually been thinking

33:14

about that. And then we got in touch

33:16

with Steve Gubser, and he pointed out, if

33:18

you tweak this model just a little bit.

33:20

so that the longest range interactions are the

33:22

strongest, like the atoms that

33:24

are far apart physically actually have the

33:26

strongest interactions. That might make it

33:28

look like your system of atoms lives

33:30

on the leaves of a tree

33:33

graph. And so what would

33:35

you have to measure to see

33:37

a signature that the counterpart of

33:39

space is being curved? Essentially, we

33:41

measure spin correlations. If a given

33:43

spin is pointing to the right, how

33:46

likely is it that another The site of our

33:48

array also has the spin pointing to the right.

33:51

So my students started thinking about what's

33:53

a clever way to plot these

33:55

data. You have to put some

33:57

constraints on how you plot it. It's nice to

33:59

plot things on a page, so in two

34:01

dimensions. And so he ended up

34:03

with some plot where the sites ended up

34:05

arranged around the circle, the ones where

34:07

we measured the strongest correlations were close to

34:09

each other. And the order of the

34:11

sites is actually very different from what it

34:14

is in our. physical system, but it's

34:16

a nice way of representing things as near

34:18

each other if the correlations between them

34:20

are strong. And then

34:22

he also started drawing lines. So pairs of

34:24

sites with the strongest correlations, he drew

34:26

a line between them. And then

34:28

he sort of treated those as a

34:30

new bigger site, asked what's the average

34:32

direction the spin is pointing on that

34:34

bigger site. And by iterating that process,

34:36

the picture that popped out was precisely

34:38

this tree graph. Let me

34:40

make sure I understand one aspect though. Since

34:42

you have so much ability to program

34:45

who's interacting with who, what part of the

34:47

results is a surprise to you? How

34:49

much is built in and how much is

34:51

not built in? Yeah. So at some

34:53

level when we did the experiment, you know,

34:55

we did know what we thought should

34:57

come out. Right. We didn't know yet how

34:59

we would analyze the data and this

35:01

way of actually visualizing the tree graph. came

35:03

up in discussions between me and my

35:05

students when that wasn't something we had thought

35:07

about before we did the experiment. In

35:10

general, we often start with

35:12

something where we know what should

35:14

happen, and the goal is eventually, and

35:16

we're not there yet to be

35:18

honest, but the goal is eventually to

35:20

get somewhere where we actually don't

35:22

know what will happen. There are some

35:25

proposals for more kind of agnostic

35:27

methods of trying to go from a

35:29

quantum mechanical system to determine does

35:31

it have an emergent description in terms

35:33

of some curved higher dimensional space. What

35:36

is the geometry? What is the

35:38

metric on that higher dimensional space? So

35:40

I think that's an important direction for future

35:42

research is to go beyond just we build a

35:44

toy model where we know what should happen. There's

35:47

lots more one can do quantitatively

35:49

and actually since then we've been developing

35:51

tools where we actually do not

35:53

in the tree graph but in simpler

35:55

settings really. probe the spatial structure

35:57

of entanglement. So I really think

36:00

it's a first step to even start to connect. Part

36:02

of it is even connecting two

36:04

different communities, right? Figuring out a common

36:06

language for even talking to theorists

36:08

who think about gravity when all I

36:10

know is quantum mechanics. Well, I

36:12

hope this is not a too sensitive

36:14

question, but you mentioned Steve Gubbs

36:16

or who I didn't know. Was

36:19

he alive to see the triumph

36:21

of you know, his insight

36:23

about how to do the setup

36:25

that it did really work. He

36:27

was not alive to see us

36:29

do the experiment. Yeah. So we

36:31

had a theoretical proposal published in

36:34

physical review letters and it was

36:36

really shortly after we wrapped that

36:38

up that he tragically passed away

36:40

in this climbing accident. I'm

36:42

sad that he didn't get to see the experiment

36:44

and I still find it really motivating

36:46

to try to kind of push forward

36:48

and continue. He clearly had some great insight

36:50

because, as you say, it didn't seem

36:53

obvious that that was the way to set

36:55

things up. Well,

36:57

the whole thing really seems very high risk

36:59

to me. I have to say

37:01

high risk, high reward. I'm just wondering, is

37:03

that kind of who you are? Are you

37:05

that type of scientist? I'm somebody

37:07

who looks for some of the problems

37:09

I work on to be ones that

37:11

everybody else isn't also working on. My

37:14

sense is, you know, if everybody

37:16

else is already doing a thing, like

37:18

I might not need to do

37:20

it. And yeah, maybe that does draw

37:22

me into things that are a

37:24

little bit more on the risky side.

37:26

And this direction of simulating quantum

37:29

gravity is one that I find high

37:31

risk, high reward. But that same

37:33

toolbox has applications and enhanced precision measurements

37:35

in quantum computation. Even if

37:37

it helps give some new insight to the

37:39

theorists that then take that to the

37:41

next level to learn something about gravity in

37:43

our universe, that would be amazing. If

37:45

it even just gives us new ways of

37:48

thinking about quantum mechanical systems, that can

37:50

help us in that effort to engineer and

37:52

understand quantum many body systems, be it

37:54

for applications in precision sensing, computation, understanding

37:56

the design of materials. I

37:58

think there's a broad effort

38:01

to control and understand entanglement

38:03

that will benefit from

38:05

research in this area. And

38:07

if it answers questions about gravity in

38:09

our universe, that's amazing. But even if it

38:11

doesn't, we won't have wasted our time.

38:14

I agree. It seems sort of surefire in

38:16

a way. It can't possibly be as

38:18

risky as it sounds, because as you say,

38:20

how could it hurt to learn more

38:22

about how to control quantum systems and manipulate

38:24

them with entanglement? That has to be

38:26

good. That's what I think. Yeah. Well, okay,

38:28

so let me close with kind of

38:31

an emotional question. What is it that really

38:33

fires you up? What are you really

38:35

trying to do? What's your motivation? I

38:37

really think that it does help

38:40

get me motivated to know that something

38:42

we learn in our experiments may

38:44

have Practical application, but it's not

38:46

important to me personally that that application

38:48

is tomorrow if it's 50 years

38:50

down the line and actually it's not

38:52

the application We thought it would

38:54

be and it's a different one.

38:57

That's fine with me So I think

38:59

I do believe deeply that fundamental

39:01

science will ultimately have technological impact

39:03

But I enjoy it also just for

39:05

the kind of thrill of being

39:07

at the frontier of the unknown

39:09

and trying to push that frontier forward

39:11

well It's so exciting to hear

39:13

about this. I really appreciate you taking the

39:16

time to talk to us today, Monica. And

39:18

thanks for being on the Joy of Why.

39:21

Thank you. It was a pleasure to chat. I

39:25

love this high -risk statement. I absolutely

39:27

love that she said she likes

39:29

to at least do some of her

39:31

work in an area which is

39:33

not crowded that nobody else is working

39:35

on. I love both those things. It

39:38

really feels to me and I think

39:40

it's where a lot of the great

39:42

stuff happens. But I wonder if your

39:44

reaction is a little bit of a

39:46

reflection of who you are. Oh, absolutely.

39:50

You know, I don't think this is for

39:52

everybody. It's not. It's not. And obviously,

39:54

if you're going to build the James Webb

39:56

Space Telescope, you don't want to be

39:58

a high risk thinker. I mean, it's a

40:00

very risky project, right? But it was

40:02

clear and well -defined. And the reason it

40:05

was late and over budget is because they

40:07

were minimizing the risk. So you definitely

40:09

don't want everybody being like that. You want

40:11

people to collaborate on big projects together

40:13

and move in the same direction. But yeah,

40:15

absolutely. My whole life. people were telling

40:17

me, don't work on that. Nobody's

40:20

doing that. Well,

40:22

I think the secret is sort of

40:24

like in relationships where you're looking for the

40:26

match. Yes. And so I feel like

40:28

there's a lot of ways to be a

40:31

good scientist or mathematician and you have

40:33

to know yourself. clearly

40:35

is not scared of risk. She's thinking

40:37

about the long game. You know,

40:40

she reminds me of Gaudi, that

40:42

Sagrada Familia in Barcelona. You

40:44

know, he didn't get to live to

40:46

see it finished, but he had a great

40:48

vision and they're still building it as

40:50

far as I can tell. Yeah, an interminable

40:52

project. But it's not for everyone. Not

40:54

everyone has that kind of nerve like she

40:56

could hit the jackpot or she might,

40:58

you know, get a nice little payout. Mm

41:00

-hmm. You know, it sounds like there's a

41:02

lot of potential unanticipated consequences. I

41:04

find that really inspiring. We throw something out in

41:07

the world and we have to see who else is

41:09

there to pick it up. Well, it's

41:11

always a pleasure to chat with you about these

41:13

things. And you, Steve. Thanks, Jenna. All right. Well, we'll

41:15

see you next time. We'll see you next time. Thanks

41:21

for listening. If you're enjoying

41:23

the joy of why and you're not already

41:25

subscribed, Hit the subscribe or

41:27

follow button where you're listening. You

41:30

can also leave a review for the show. It

41:32

helps people find this podcast. Find

41:34

articles, newsletters, videos and

41:36

more at Quantamagazine .org. The

41:39

Joy of Why is a

41:41

podcast from Quantamagazine, an editorially

41:43

independent publication supported by the

41:46

Simons Foundation. Funding decisions

41:48

by the Simons Foundation have no

41:50

influence on the selection of topics, guests,

41:52

or other editorial decisions in

41:55

this podcast or in Quanta

41:57

Magazine. The Joy of

41:59

Why is produced by PRX Productions. The

42:01

team is Caitlin Folds,

42:03

Livia Brock, Genevieve Sponsler,

42:05

and Merit Jacob. The

42:08

executive producer of PRX Productions

42:10

is Jocelyn Gonzalez. Edwin

42:12

Ochoa is our project manager. From

42:15

Quanta Magazine, Simon France and

42:17

Samir Patel provided editorial guidance with

42:19

support from Matt Carlstrom, Samuel

42:22

Velasco, Simone Barr,

42:24

and Michael Kenyungalo. Samir

42:26

Patel is Quanta's editor -in -chief.

42:29

Our music is from APM Music.

42:31

The episode is by Peter Greenwood

42:34

and our logo is by

42:36

Jackie King and Christina Armitage. Special

42:38

thanks to the Columbia Journalism

42:40

School and the Cornell Broadcast Studios.

42:43

I'm your host, Steve Strogas. If

42:46

you have any questions or

42:48

comments for us, please email

42:50

us at quanta simonsfoundation .org. From

43:02

PRX.