0:00

It looks like an Audi. It drives

0:02

like an Audi. It's crafted like an

0:04

Audi. It exhilarates, excites,

0:07

and thrills like an Audi. And

0:10

after all that, it electrifies.

0:13

Because it's a fully electric Audi e-tron.

0:16

How you get there matters. And with five

0:18

fully electric models available today,

0:21

getting there has never been easier. Audi,

0:24

progress you can feel. Learn

0:27

more at audiusa.com

0:29

slash electric.

0:31

I've been having a rough

0:34

week, I have to tell you. What happened? Well,

0:36

there's this joke that I want to use that I can't use,

0:38

and it's driving me crazy. What's the joke? Okay,

0:41

so as we record this, as you probably know, Chairman

0:44

Xi Jinping of China is in San

0:46

Francisco. Yes. He's having a meeting with

0:48

President Biden. It's this really big deal.

0:50

So I thought, what could be better while

0:52

Chairman Xi is in San Francisco than to

0:55

change my Grindr bio to Chairman Xi? Like,

0:59

just imagine, you're scrolling through the grid, you want

1:01

to see who's cute in the neighborhood,

1:02

and then you come across the official Communist

1:05

Party portrait of Xi Jinping, and the bio

1:07

just says, visiting for work, looking

1:11

for cuddles. I, if I

1:13

saw this, it would make my whole life. And

1:15

so I want to give that somebody, that was that experience,

1:18

but I can't because the platform has a rule

1:20

against impersonation. No. Yes,

1:22

and I was like, as much as I like this joke, I

1:25

am not willing to give up my Grindr account for it. Well,

1:27

and isn't Grindr now owned by a Chinese company?

1:29

Well, it was, but the US got

1:32

very concerned about this because there's

1:34

a lot of what I guess we could call sensitive data

1:37

being exchanged on Grindr,

1:38

and they legitimately worried

1:40

that it would have like national security implication

1:43

if you had a bunch of US military service

1:45

members running around the world with Grindr on their phones

1:47

while it was owned by a Chinese company, so

1:49

they had to spin it out. Well, I'm sorry that your Grindr

1:52

gag didn't work, but I am happy

1:54

that you're now definitely on our no fly list for attempting

1:57

it.

2:02

I'm Kevin, and this is Tech Columnist from the New York Times.

2:05

I'm Casey Neu from Platformer. And this is Hard Work. This

2:07

week on the show, there's an AI pin

2:10

that has the tech industry talking, and

2:12

I tried it. Then YouTube opened

2:14

its doors to deepfakes.

2:16

And finally, Sci-Fi Foods CEO

2:19

Josh March serves us a little Thanksgiving

2:21

dinner that was grown in a lab.

2:24

Yeah, the sider waiver. We did.

2:39

Casey, a couple weeks ago on the show, we talked about these

2:41

AI wearables that have been announced

2:44

over the past few months. And recently,

2:46

we got a big update on one of them.

2:49

Just last week, the Humane AI pin

2:51

was launched and demoed, and you

2:54

actually got to see it up close,

2:56

I understand. That's right. It was the sequel

2:58

to the Inhumane AI pin, which sort of

3:00

committed many fewer human rights violations,

3:03

and I appreciated that. But no, listen, imagine

3:05

you have $700 and you

3:07

want to answer for yourself the question, what if Siri

3:10

was good? Okay, you might be

3:12

interested in the Humane AI

3:14

pin. This thing is a little

3:16

computer and has a little magnet, and

3:18

you snap it together and you wear it on

3:21

your shirt or your jacket. And I know

3:23

what your next question is going to be, which is, well, what the heck

3:25

does it do? What the heck does it do?

3:27

Well, I'll tell you. So it's got a camera, can take pictures. Okay,

3:30

if you want to talk to it, you tap it. You

3:32

can ask it questions that you might ask

3:34

an AI. It has a speaker,

3:37

what they call a personic speaker. That's

3:39

a combination of personal and

3:41

sonic. And this speaker

3:44

will talk to you. You could say,

3:46

hey, what's going on with the weather? Or give

3:48

me some ideas for a recipe? Or is there a grocery

3:51

store around here? And now you don't have to get your phone out

3:53

of your pocket. You just tap your little pin and

3:55

you move on with your day. Oh, and one more

3:57

thing, Kevin, because I know what you were going to say. You were going to say,

3:59

Casey. Did they build a laser projector

4:02

into this thing so that you can project

4:04

a user interface onto your palm

4:07

in lieu of a screen? You were about to ask me that. I

4:09

was. It's true. Let me tell

4:11

you, they built a laser projection system, okay? So for the first

4:13

time in your life, I know every time you're

4:15

looking at a screen on your smartphone you're saying, why

4:18

isn't this projected onto my hand? Exactly.

4:20

Those humane people answer that question. Okay, here are the other

4:23

things we know about it. It costs $699. Yeah. It's

4:26

available in three, what they call colorways.

4:28

Why this company chooses not to use the word color

4:31

and opts for colorways instead. I'll never

4:33

know. Colorways and Silicon Valley speak for

4:35

colors, okay? We have a unique culture

4:37

here in our own language. So

4:40

you also have to buy a data plan that costs $24

4:43

a month on the T-Mobile network and it

4:45

gives you a new phone number and you

4:47

can use it to send and receive text messages.

4:50

You're a green bubble if you use the humane AI pin.

4:52

Great, which is the color of life, Kevin. So

4:56

you get this pin, it is sort of being billed

4:58

as not necessarily a smartphone replacement,

5:01

but something that may over time come to do

5:03

more of the things that you would currently use your smartphone

5:05

for. Yeah, I mean, I think when these folks started

5:07

at Humane, they were very curious about

5:10

what would be next generation hardware.

5:12

People have very different answers to this question,

5:14

right? The meta people think it's going to be some kind of helmet

5:17

or glasses that you put on your face. The snap people

5:19

think it's definitely glasses. The Humane people said, we're

5:21

going to build a pin, but everyone

5:24

is trying to figure out, well, is there kind of something

5:26

beyond the simple smartphone and maybe

5:28

that thing doesn't have a screen? Right. So

5:31

that's one sort of reason for this device to

5:33

exist is that I think everyone in Silicon

5:36

Valley is thinking about what comes after the

5:38

smartphone. What is the sort of logical next step?

5:40

And a lot of companies are also thinking about, well,

5:43

these AI tools, they're very powerful

5:45

and very cool and very potentially useful,

5:48

but using them on a smartphone just

5:50

feels a little bit anachronistic. Maybe there's some device

5:53

that should be custom built for this. And so this is,

5:55

I would say, the first major release that

5:57

we've seen that tries to answer that question of like, what would a

5:59

device. that was built for AI look like.

6:02

Yeah, and I think it also tries to answer the question

6:04

of like, if AI gets good enough, maybe you don't

6:06

need a screen anymore. Maybe you don't really need apps

6:08

anymore, right? Maybe it is just a purely

6:11

conversational interface that does whatever

6:13

you want it to do. So I was not

6:15

invited to the launch of the Humian AI

6:17

pin. What did you do to those people? Jesus. And

6:21

but I did watch the launch video, very

6:24

strange launch video, I got to say, it takes

6:26

place in this like empty office with

6:28

nothing on the walls. It's like two people

6:30

who sort of look like they're doing Steve Jobs cosplay

6:33

like all dressed in black. Kevin, OK,

6:35

this style has a name. And

6:37

I would say it's like a kind of low

6:40

energy presentation of what they

6:42

say will be like a world changing device, which

6:44

I was a little confused about. But aside

6:46

from the aesthetics of the video, there were a few

6:49

things that I did want to like flag as being potentially

6:51

cool because I don't want to just crap on

6:53

this idea out of the game. You kind of do. Well,

6:57

I have some questions about it, but I will say

6:59

there are a couple of things that struck me as pretty cool. One

7:01

is the device itself looks cool. It's

7:03

like a very sleek sort of iPhone

7:06

looking device, which makes sense because a lot of

7:08

the people at Humian came from Apple.

7:11

And I thought that a couple of the features were cool.

7:13

One is it can summarize your

7:15

text messages to you over voice,

7:18

which is always something I don't know if you've ever tried to

7:20

use Siri or something while you're driving

7:22

to like send and receive text messages.

7:24

But it is infuriating. Absolutely.

7:27

Because it reads it will read like not only

7:29

the text messages that you want, but like also

7:32

the like the security code that

7:34

you got when you tried to log into your bank and it'll

7:36

read that to you. You know, it'll

7:38

read like all the emoji like tap

7:41

backs to your messages. It

7:43

just doesn't really do a good job of getting you the

7:45

information that you're looking for in a succinct way. But yeah,

7:47

it doesn't seem like the people who built the feature use

7:49

it. It's I think how I would describe it. So but this this

7:52

humane AI pin, you can ask, you know, catch

7:54

me up. This was something that they showed off in the

7:56

demo video, and it will sort of

7:58

summarize using AI. all of the texts

8:01

and emails that you've gotten since you last asked. Yeah, and

8:03

by the way, I will say this is a very good pitch

8:05

for the sort of early adopter tech crowd because

8:07

what could be more flattering to your own

8:09

ego than the idea that you're receiving so

8:11

many messages that you actually need an executive

8:14

summary of the messages that you

8:16

receive. It's like, yeah, let me let me dig into

8:18

the details. It's like, oh, well, your friend shared another

8:20

meme and that was your executive summary

8:23

of your messages. Totally. So

8:25

one other cool feature I thought was this instant translation

8:28

feature, which was in this demo video where they basically

8:30

show two people talking. One of them is talking

8:32

in Spanish and the other one is wearing

8:35

one of these humane AI pins and they

8:37

can just sort of tap it and it will

8:39

instantaneously take that Spanish that

8:42

someone is talking and convert it into

8:44

English and speak it out loud to you in English in

8:46

sort of something that resembles their voice. Yeah,

8:49

that could be a fun thing to take around

8:51

the world and meet new people and talk

8:53

to people you might not otherwise be able to talk to. Totally.

8:56

So if you're in a restaurant in a foreign country, you could just tap your little pin

8:58

and it could sort of translate what you want. There

9:01

you go. So those are some of the cool things.

9:03

I do have some questions about this device though,

9:05

because I would say the demo that

9:07

it sounds like you and some other reporters got

9:10

was fairly limited. Like people weren't actually allowed

9:12

to try this on for very long. My

9:14

colleague, Aaron Griffith was able to try it

9:16

for like 10 minutes, which is always kind of

9:18

a red flag when a company releases something

9:20

and is like, we can use it. We can show

9:22

you, but you can't use it yourself. I always kind of

9:24

wonder like, what are they hiding? Yeah, yeah.

9:27

Let's just say that reporters were never asked to try

9:29

the Theranos for themselves before that

9:32

whole thing went down. So here's

9:35

one of my questions. Is the experience of

9:37

using it or seeing it used, did you feel like this

9:39

is actually a step forward for computing

9:41

or did it just kind of feel like a new gadget

9:44

that was kind of cool, but maybe not all that much more

9:46

useful than your smartphone? Yeah, I think

9:49

there are sort of two different ways

9:51

of using this thing that I have very different feelings

9:53

about. There is the laser projector

9:56

thing, right? And when I was watching the

9:58

humane employees show this to me. me. There

10:00

was part of it that was just quite fascinating

10:02

because they have developed these gestures like

10:05

to zoom out to the home screen, you pull your

10:07

hand back from the projector. And

10:09

to do this picking, you make these gestures

10:11

like pinching your fingers together. And it's

10:14

like watching someone who has learned

10:16

a new form of sign language to

10:18

interact with a computer. And there is

10:20

just kind of a weird spectacle in that. At

10:23

the same time, I looked at that and

10:25

I thought, this thing does not do

10:27

enough to get me to learn a new language yet. So the

10:31

projector stuff, it's unbelievably

10:33

cool. Just from a technology perspective,

10:35

I don't think it is going to be the easiest

10:37

way to do anything that they showed me. But then there's

10:40

a second question, which is, do I

10:43

want to wear AI? And

10:45

I think there are cases where

10:48

the answer is probably definitely going

10:50

to be yes. We've seen

10:52

in sci fi, a lot of the

10:54

ways that the creative types have been predicting

10:57

an AI future is that you wear an earbud

10:59

and you're just sort of able to converse with that all day

11:02

long. Her the movie from 2013 by

11:04

Spike Jones, the most famous example of that we

11:06

talked on the show about Mrs. Davis, a show

11:08

from this year that has the same kind of metaphor.

11:11

And I do think people are going

11:13

to be doing enough computing during their day that the ability

11:15

to just kind of like tap a thing on their ear and be

11:17

like, Hey, like, give me directions to the grocery store.

11:20

Or Hey, you know, what emails have I gotten

11:22

since I was in that meeting? That makes a lot

11:24

of sense to me. Now, this is a brooch

11:27

that you put on your chest, not a thing that

11:29

you put in your ear. But 1.0 hardware

11:33

is almost always bad, right? Like, I

11:35

did you buy the first iPhone? No, no, neither did

11:37

I. Right. And yet we can both agree the iPhone

11:39

was a really good idea and they got there.

11:42

So to me, the interesting question is not like, should

11:44

everyone go out and buy a $699 device with a $24 a month

11:47

plan? Because the answer for

11:49

most people is obviously going to be absolutely

11:52

not. To me, the interesting question is like,

11:54

well, is there a direction here? Is there

11:56

a path to something? And does the path

11:59

wind up being AI on your chest? Does

12:01

it wind up being AI somewhere else? Or

12:03

do we have it all wrong? And we really just are going to

12:05

use smartphones forever. But if I had to make the bet

12:07

right now, I would say that yes, there is something

12:09

beyond the smartphone. Yeah, speaking of that path,

12:12

I want to bring up one more thing about this

12:14

company that I just find totally fascinating and

12:16

entertaining, which is their origin story.

12:18

This was a story that my colleagues, Aaron Griffith and Tripp

12:20

McColl wrote in the New York Times about humane.

12:23

And it is truly my favorite detail about

12:25

this company, which is that it owes

12:27

its existence to a Buddhist monk who

12:29

goes by the name Brother Spirit. Did

12:32

you read the story? I did. I did was

12:34

very interested to learn this. It didn't come up during

12:36

the presentation. I'll say that. So I'll just

12:38

read this paragraph to you. Quote, a

12:40

Buddhist monk named Brother Spirit led

12:43

them the founders of humane to humane.

12:45

Mr. Chowdhury and Mrs. Bonjorno had

12:48

developed concepts for two AI products,

12:50

a women's health device and the pin. Brother

12:52

Spirit, whom they met through their acupuncturist,

12:55

recommended that they share the ideas with his

12:57

friend, Mark Benioff, the founder of Salesforce.

13:01

A more San Francisco paragraph has

13:03

never been written in the pages of a major newspaper.

13:05

I would submit to you. I love that.

13:07

I mean, look, Mark Benioff, he's a

13:09

he's a designer of a lot of stories in San Francisco.

13:11

You know, he's he's a man about town. He's making connections.

13:14

He's wheeling a deal. Brother Spirit, I need that. I need

13:16

like a 3000 word profile of Brother Spirit

13:18

and how he's become the tech advisor to Silicon

13:20

Valley. Got him on the show. Yeah. Yeah. So since

13:22

this launch, there has been a lot of

13:25

hubbub. People are saying this is a great idea.

13:27

I'm very excited about this. Other people are saying this

13:29

will never work. Who's going to pay $700 plus $24 a

13:33

month to wear a thing on

13:35

your shirt that can't even do the things that your

13:37

smartphone can. And then there was there

13:39

were some funny little details in

13:41

the promotional material for

13:43

the launch itself that turned out to be

13:46

wrong in the ways that AI products sometimes

13:48

do get things wrong. So there was a part

13:50

in the video where they

13:53

showed off this AI pin being

13:55

used to take the nutritional contents

13:57

of a snack that the one of the founders was

14:00

preparing to eat, so he taps his pin and

14:02

is holding some almonds in his hand and

14:04

he says, how much protein is in these

14:06

almonds? And the Humane AI pin

14:09

uses its camera to analyze it and says

14:11

there are 15 grams of protein in

14:13

these almonds. People who

14:15

watched this video later pointed out that it would take

14:18

about 60 almonds to get 15 grams

14:20

of protein, so many more almonds than the small

14:22

handful that was shown in this video. So

14:24

if you are relying on the Humane AI pin for

14:27

your nutrition facts, you may be getting things wrong.

14:31

Another discrepancy with the video, they

14:33

showed off the AI pin being asked where the best place

14:35

to watch the next solar eclipse would be and

14:38

suggested watching it from Australia.

14:40

It turns out the upcoming solar eclipse

14:42

actually not going to be visible from Australia. Well

14:44

I thought maybe the pin was just having fun with you because

14:46

imagine you book a trip to Australia, you get down there and then

14:49

the pin is like, walk out, walk out. But

14:51

you know, something similar happened with the Google Bard

14:53

launch where there was like a screenshot that contained this

14:56

factual error. So yes, I think

14:58

if we've learned nothing else from these AI launches, it's

15:00

that you really want to get a fact checker for your promotional

15:02

materials. Totally. So aside from these

15:04

sort of small details and some

15:06

of the skepticism around this product category

15:08

in general, like did this strike

15:11

you as an AI wearable that you would actually want

15:13

to wear? No, I wouldn't. But

15:15

at the same time, I do want to try to bring peace

15:17

to Silicon Valley because as you said, there are really

15:19

two camps here. There is the camp of folks, particularly

15:22

in the sort of hardware reviewer crowd that are just

15:24

extremely skeptical saying, nobody wants this, this

15:26

thing is going to flop, it's the next juice arrow. And

15:29

then you have people that I would just describe as

15:31

like technologists, people who work on

15:33

product who are engineers. And they're looking at this

15:35

and they're like, there is an insane amount of cool

15:37

technology in there. So I just want to say like, both

15:40

of these people are actually right. It is both true

15:42

that most people should not buy this device. And

15:44

it's true that there's some incredible technology

15:46

that they built in there. Now, what does this mean

15:49

for the humane company? I don't know. There

15:52

is a very, very difficult thing to

15:54

get right. They're going to need to hope that they

15:56

sell enough or are going to be able to raise enough

15:58

additional capital that they They can make a version two

16:01

and three and four and hope they're able to

16:03

find that product market fit with what people actually

16:05

want. But in terms of how good

16:08

a start are they off to, I don't

16:10

know. I gotta say probably

16:12

a C minus. I would have to give this whole thing. I

16:14

will say what this demo and

16:16

hearing about this AI pin really

16:18

made me wish for was not an AI

16:21

wearable. It was a better Siri. I

16:23

would give so much money for

16:25

a Siri that could actually do the things

16:27

that I want Siri to do that could

16:29

do this kind of instant translation that could

16:31

summarize my text messages and not just read

16:34

every little emoji out to me. That

16:37

is what I want from Siri. But

16:40

I will also say like I think there is something to

16:42

this idea of the screen.

16:45

Screens get a bad rap mostly from you,

16:47

mostly from me. But screens have

16:49

one thing going for them, which is that they are a very dense

16:52

way to consume information. I don't

16:54

know if you've ever tried to order something from

16:56

an Alexa device. Have you ever tried to do this? I

16:59

would never even try. Honestly. So

17:01

I haven't tried it a long time. But that's because

17:04

when I did try, back when this feature first came out,

17:06

they would say you can reorder dog food

17:08

through your Alexa device. And you would

17:10

try doing this. And what it would spit back

17:13

was a list that would take about two

17:15

minutes to read. It would say, OK, we've got Purina

17:17

Kibble. It's a 30 pound

17:19

bag for $46. We've

17:22

also got I am's Kibble.

17:25

And it's got lamb flavor and chicken

17:27

flavor. And it comes in

17:29

a 28 pound bag. And it was like five

17:31

minutes later, you have this list of

17:33

things that you can order from. Whereas if I'm doing

17:35

this on my screen, I can consume all of that information

17:38

very quickly, select the dog food that I want

17:40

and move on with my life. Audio is

17:43

just not a very information dense medium.

17:45

And so, yes, I think there will be times when

17:47

you want to talk to an AI and have an AI

17:50

talk back to you. But I do not think that that

17:52

is how we will go about doing things like

17:54

ordering products or buying plane

17:57

tickets or anything like that, because it's just so. slow.

18:00

You're totally right. And I do think the question of what

18:02

if Siri were good is a good one. I

18:05

mean, I think another very potential outcome

18:07

here is that Apple buys humane. Like

18:09

after the first round of sales, no good, you want, but

18:11

they built all this cool technology that is often

18:13

a time where hardware startups look around and say, well, who

18:16

might this hardware be useful to? Totally. Okay,

18:19

that's the humane AI pin. We will keep tabs on

18:21

this category and this story going forward. And

18:23

humane, if you want to send us some demos to use,

18:26

we will we will try them out. Until then, we'll

18:28

put a pin in it. Hey,

18:29

Hey,

18:47

there's progress.

18:50

And then there's the book,

18:53

the double looks, the

18:55

craftsmanship,

18:56

the sound that's both serene and

18:59

invigorating the fast

19:01

charging capability, the electric

19:03

performance, the belief that

19:06

how you get there matters in

19:08

the feeling. There's progress. And

19:11

then there's the fully electric Audi Q8

19:13

e-tron.

19:14

Audi

19:16

progress, you can feel. All

19:19

right, Casey, speaking of AI,

19:21

we had some other news this

19:23

week about AI and content

19:26

moderation. You had a newsletter

19:28

this week about YouTube and some policy

19:30

changes that they have made to deal with AI generated

19:33

content on their platform. That's right.

19:35

And if you're already saying, I don't know, Kevin,

19:37

that sounds a little boring. Here's how I would frame

19:39

it. If you have spent the

19:41

past year wondering, when is the

19:44

flood of deep fakes truly going

19:46

to arrive and make it difficult for me in

19:48

many cases to tell what is true and false? Well,

19:51

we are getting ever closer to that day, my friend. Totally.

19:53

So let's just recap what happened with YouTube

19:55

this week and then talk about what it means. So on

19:58

Tuesday, YouTube put up a blog post. outlining

20:00

what it calls its approach to responsible

20:03

AI innovation. And in practice, what

20:05

they did was give people permission

20:07

to post a lot of what I like to call synthetic

20:10

media, what are often called deep fakes.

20:12

These are videos that have been created using

20:14

generative AI, or their video that

20:16

maybe you shot on your camera, but you use some sort

20:18

of AI tool to manipulate

20:21

it, right? And a lot of platforms have been coming out with

20:23

these kind of generative AI policies trying

20:25

to wrestle with what do we do when people

20:27

start making fake videos or

20:29

posts that are manipulated in some way

20:32

using AI to claim that something

20:34

happened that actually didn't happen. So we're starting to see

20:36

platforms starting to grapple with this. But what YouTube

20:39

did is not just throw open the doors

20:41

to deep fakes, right? Because they actually did add some

20:43

disclosure requirements and require

20:45

creators to label videos that

20:47

show what they call realistic, altered

20:50

or synthetic content. So what did they

20:52

actually announce? And then let's talk about what it sort of means

20:54

between the lines. Yeah, well, so as you point out,

20:57

there are these disclosure requirements. So if

20:59

you are going to upload a synthetic

21:01

video, and it looks realistic,

21:03

whatever that means, we don't have details on that, you

21:06

are going to need to tell YouTube

21:08

when you're uploading the video. And YouTube

21:10

is then going to put one of two labels

21:12

on it. If you're doing something that's sort of silly

21:15

and fun, like maybe you use an AI tool

21:17

to create like a dog chasing a cat,

21:19

and it's very cute, and no one cares about the societal

21:21

implications of that, then it's just

21:23

going to sort of be in the metadata, when you sort of

21:25

click into a detail view of video, it'll say,

21:27

hey, this was made using an AI tool. If

21:30

you are doing something a little bit more

21:32

sensitive, and again, we don't have a lot of details about

21:35

what this is going to mean, but I don't know, maybe

21:37

you make synthetic video

21:39

about an election, and

21:41

it feels a little bit edgier, then there's

21:44

going to be an overlay on top of the video. So as

21:46

people watch it, they're going to see a little box that

21:48

says, hey, this is altered or synthetic

21:50

content. And is it sort of working on the honor

21:52

system, like YouTube is trusting that creators

21:55

will check this little box when they do

21:57

upload stuff that was generated using AI?

22:00

much. You know, there are no reliable

22:02

tools for detecting in many cases

22:05

what videos were made using generative

22:07

AI. And so YouTube is going to ask

22:09

people to just be honest about that.

22:11

Now, at the same time, they have AI

22:13

systems of their own. And I'm sure that over time,

22:15

those will get better at detecting what was

22:17

made with AI. So there's this little overlay

22:20

if your video is deemed to be realistic,

22:22

or maybe deals with some sensitive topic. What

22:25

else did they announce in this blog post? Well,

22:27

look, maybe you're listening to this and you have the

22:29

assumption that if I woke up

22:31

one day and said, I'm going to make a deep fake

22:33

of Kevin and the deep fake is going to say, I'm

22:35

a big dumb dumb and I'm bad at podcasting.

22:38

And I made that just the information that

22:40

I say in the mirror every morning. How'd you get that video?

22:43

Back into my Dropbox. So let's

22:46

say I make this video and I

22:49

go to upload it to YouTube that

22:51

that would just sort of automatically be against

22:53

the rules, right? Maybe YouTube might even have an automated

22:55

system to say like, I don't know, did Casey have Kevin's

22:58

permission to do this? That is not what they're

23:00

going to do. They're going to let me upload

23:02

that. And then if you Kevin, do

23:05

not like the video, you can go into

23:07

YouTube and you can file a

23:09

request under YouTube's privacy policy to

23:11

say, Hey, I don't like this, take

23:14

it down. And then YouTube will consider a variety

23:16

of factors, including whether

23:18

you are a public figure, and then

23:20

they will decide or their automated systems will decide

23:23

whether they want to honor your request. But the answer

23:25

might be no. And the I'm a dumb dumb

23:27

video might stay up on YouTube. And it now has 4 million

23:30

views and counting. I just

23:33

bought a house with the AdSense money. So if

23:35

a deep fake video is made of me

23:37

and uploaded on YouTube, I do have some recourse

23:40

I can ask for it to be taken down and YouTube will consider

23:42

that. Do you think that's YouTube trying to sort of get

23:44

ahead of some problems that it sees

23:46

coming? Or do you think this is already happening? Well, I'm

23:48

sure that in some cases it's already happening, although

23:51

I don't think it's happening that widely. But

23:53

I think YouTube sort of arrived at

23:55

a fork in the road where they had to decide do

23:57

we want to be really restrictive about what we enable?

24:00

people to post using generative AI or

24:02

do we want to be more permissive and

24:04

say hey go nuts and we'll

24:07

just sort of remove bad things on a case-by-case

24:09

basis. And to be honest with you I sort

24:11

of assumed they would err on the more restrictive

24:13

side of things just given all of the backlash

24:16

to social platforms in general over the past half

24:18

decade or so the sort of misinformation they've

24:21

enabled the hate speech the cyberbullying the

24:23

harassment but instead YouTube just said

24:25

uh you know we're gonna trust you guys and you have

24:27

to tell us if you made it using generative AI but

24:29

otherwise our policies are mostly gonna stand as they are. And

24:32

why do you think they did take that permissive

24:34

approach? I mean I have not been able

24:36

to get them to give me a straight answer to that

24:38

question. I think they would say that they

24:40

have always operated in a tradition

24:43

that tries to enable the maximum amount

24:45

of speech. I think there are some good reasons

24:47

to do that. If you believe that we should

24:49

be able to do parody and satire

24:51

of our world leaders and upload those videos

24:54

to YouTube and you want to make a parody or satire

24:56

using generative AI then YouTube is

24:58

enabling you to do that. That's the same time I would

25:00

just note if I were trying to design the

25:03

cheapest way for YouTube to do

25:05

content moderation in a world of generative AI

25:08

this is the system I would pick because

25:10

the onus for reporting bad things it

25:12

is not on the platform it's not on me who

25:14

made the video calling you a dum-dum it is now

25:17

on you the possible victim of

25:19

my deepfake to go in and just hope

25:21

that the little form that you fill out reaches

25:24

somebody who agrees with you. I guess my big

25:26

question about this is how is it going to work

25:28

when people with a lot of clout

25:31

with YouTube when advertisers when big YouTube

25:33

creators when celebrities start

25:35

getting deepfaked into videos that

25:38

they didn't appear in and that they're not happy with

25:40

is that just going to be sort of treated on a case-by-case

25:42

basis or will they actually lobby YouTube

25:45

to change its policies what do you think? Well it's

25:47

a great open question do you remember

25:49

the video of Nancy Pelosi

25:52

in which she appeared to be slurring her

25:54

words. Yes it was this is a big controversy

25:56

from 2019 and this was not actually

25:59

a deepfake someone taken a video of Nancy

26:01

Pelosi speaking and they slowed it down

26:03

a little bit to make it seem as if she were slurring

26:06

her words and maybe potentially drunk. And this was

26:08

a huge controversy. Nancy Pelosi was

26:10

super mad about this. They were lobbying Facebook

26:12

to take it down. It was a big thing inside of Facebook.

26:15

Ultimately, Facebook decided to leave that

26:18

video up. But again, we're

26:20

so early in these days that most

26:22

famous people have not yet had the experience of

26:24

seeing themselves deep fakes. And I have

26:26

to imagine that pretty soon a lot

26:28

of them are going to see these deep fakes and some healthy

26:31

percentage of them are going to hate it. Right. They're gonna be super

26:33

mad and they're not going to be sort of placated

26:35

by like a little overlay on the video, even

26:38

if the creators are honest

26:40

and forthright and check the box when they're uploading

26:42

their deep fakes. And of course, Kevin, there is another

26:44

twist, which is that there is a different

26:46

system if you are a musical artist. Really?

26:49

Right. Yeah. So we talked about this a little

26:51

bit on previous episodes when we were talking about AI

26:53

music stuff. But what YouTube made clear

26:55

with this most recent blog post is that there is going

26:58

to be a different process for its music

27:00

partners. We know that YouTube

27:02

is heavily dependent on its relationships

27:04

with the major labels for all sorts of things. Music

27:07

videos and music in general is a

27:09

huge percentage of the listening

27:11

and the watching that YouTube gets, right? So

27:13

YouTube does not want to lose these deals. It

27:16

also really wants to create new creative

27:19

tools so that you or I could go into YouTube

27:21

or some new app and say, Hey, I want to make

27:23

a radio head song with Drake's voice. And that

27:25

would just be available to me. I think they'll probably

27:27

try to set it up so that the artist gets some sort of

27:30

revenue if what I create winds up making

27:32

any money, right? But so as

27:34

part of these negotiations, and as part

27:36

of this new blog post, YouTube has gone out

27:38

and said that if you're a record

27:40

label and you represent artists participating

27:43

in these early AI music experiences, you're

27:45

going to get a different request form

27:48

that you can fill out saying, Hey, my name

27:50

is Drake. I'm a famous recording artist. I've noticed

27:52

a bunch of people are making songs using my voice. I

27:54

don't like it. Take them down. And I assume

27:56

that those requests are going to get a lot

27:59

more attention than just Joe Schmo when he files

28:01

his request saying, hey, I don't like this deep fake. Right.

28:03

So this policy, I guess it feels

28:05

like maybe a 1.0 policy that's

28:08

going to get revised over time. Because right

28:10

now, the technology for creating

28:13

video deep fakes is just not quite there

28:15

yet. You can create little short clips,

28:18

but it's just not really. You can't

28:20

type in, like, make me a full length

28:22

music video of Drake

28:24

singing a radio head song. That is still beyond

28:27

the reach of these AI systems. But

28:29

I think they're probably thinking, like, it's not going

28:31

to last. These tools will get better. And

28:34

pretty soon, it will be possible

28:36

to make something like a full length music video. And

28:38

so it feels to me like this is sort of them wanting

28:40

to put something out, but knowing that it's going to change

28:43

and evolve over time. Is that your read on it? It is. And

28:45

by the way, that's good. You know, I should also say these

28:47

policies aren't even taking effect yet. They're

28:49

going to take effect in the coming months, which means

28:51

that YouTube is giving itself time to

28:54

refine them based on the feedback that it gets.

28:56

So that's good. I'm all in favor of

28:58

that. It's good in general to see platforms trying

29:00

to get out in front of these issues and not

29:03

just catch up once the crisis is already

29:05

metastasized. Yeah. I also think

29:07

it's probably them trying to get ahead of potential

29:09

deep fakes having to do with the 2024 election in the US. We

29:13

know that the platforms, they all sort of orient their

29:15

trust and safety work around US elections,

29:18

because those are the times when the stakes

29:20

are the highest for them and

29:22

when politicians are really paying attention to what's

29:24

going viral on some of these platforms. So

29:27

I understand the impulse to want to do something. You

29:29

had an interesting point in your newsletter, though, which

29:31

is basically about how this is content

29:33

moderation that is happening at sort of

29:36

a different level than we are used to. Explain

29:38

that and sort of walk us through your logic there. Yeah.

29:41

So in the first era of social media,

29:43

content moderation rarely happened at

29:46

the level of the tool. And what I mean by

29:48

that is if you open Adobe Photoshop and

29:50

you wanted to draw a figure of a naked human,

29:53

Adobe is not going to pop up a little wizard that

29:55

says, looks like you're trying to draw nudity. Knock

29:57

it off, right? But if you go to open AI.

32:00

that you're creating gets shared and distributed. But

32:02

we don't live in that world because the platforms are saying

32:04

absolutely share and distribute this kind of content. Now,

32:06

the platforms would say, Casey, we still have

32:09

all of the same rules, right? You cannot post hate

32:11

speech if you make it with generative AI. You cannot

32:13

bully people using generative AI. And

32:16

I hear that, but I'm looking at these

32:18

policies as nascent as they are, and

32:20

it just seems like they're leaving a lot of wiggle room

32:22

for people to cause mischief and mayhem.

32:24

So I would just say, let's keep our eyes

32:26

on this as these video-making tools get

32:28

better. Totally, all

32:29

right. When

32:32

we come back, we talk with a man who is trying to bring lab-grown

32:35

meat to the masses, and we get a taste. A

32:37

taste of the future.

32:53

An SUV this practical shouldn't

32:55

be this enjoyable to drive. One

32:58

of this size shouldn't be able to handle

33:00

so much.

33:02

One so functional

33:03

shouldn't be this sharp.

33:05

A model this innovative shouldn't

33:07

also be this intuitive. And

33:09

yet the fully electric Audi Q4 e-tron

33:12

is just that,

33:13

because how you get there matters every

33:15

single day. The fully electric

33:17

Audi Q4 e-tron.

33:19

Audi,

33:20

progress you can feel.

33:22

Learn more at audiusa.com

33:25

slash electric.

33:27

So Casey, we're coming up on Thanksgiving and the holiday

33:30

season, when a lot of people gather with their families and

33:32

eat a bunch of meat. And

33:34

this has been coming up for me recently,

33:37

because I have family members who are non-meat

33:39

eaters, they're vegetarians and vegans. And

33:41

so every year I'm confronted with the same horrible

33:44

question that gnaws on my conscience all year, which

33:46

is, why do I still eat meat? Yeah,

33:48

it's one of those things where if you sort of

33:50

just do any level of research, it

33:53

becomes clear to you that if

33:55

nothing else, you should probably be eating less meat as an

33:57

American than you're already doing. Yeah, I like

33:59

to call myself. an intellectually convinced

34:01

vegetarian because I know on

34:03

a rational level that eating

34:05

meat is indefensible. I know that our kids and

34:07

grandkids are going to be horrified that we ate

34:10

meat. And yet what we have to say to them is,

34:12

but it was delicious. And

34:14

that's the thing that so often gets left out of this conversation.

34:17

It's true. I mean, I don't have to tell you

34:19

there are all kinds of reasons that meat is

34:21

not something that we should be eating as much

34:24

of as we do. It has environmental

34:26

sustainability risks. It's

34:28

cruel to animals when you think about factory

34:30

farming, and it has public health

34:32

risks too. And so in the lead up to Thanksgiving,

34:35

these arguments about meat are always

34:37

rattling around in my head. And

34:39

so this week, I wanted to actually

34:42

do something about it and have a conversation that squares

34:44

directly on this issue of meat. You've

34:46

looked into the question of can we sort

34:49

of have our beef and eat it too, as

34:51

it were? Exactly. So

34:53

I've been fascinated for years with this issue of cultivated

34:56

meat as they're now calling it. It

34:58

used to be called lab grown meat, although I liked it better because then at least I knew where

35:01

it came from. It came from the lab. So

35:03

cultivated meat, if you're not familiar with it, is just

35:05

a term that applies to meat that

35:08

is grown from the cells of animals,

35:10

but it is not grown on an animal. There's

35:13

no living being that has to be slaughtered

35:16

to make this meat. Instead, it is grown in

35:18

a lab in these bioreactor

35:21

tanks. And this has been something

35:23

that tech companies and food science companies

35:25

have been working toward for years. And

35:27

I would say that it has been an area where the progress

35:30

so far has been disappointing. Yeah. For

35:32

decades now, scientists have

35:35

considered this a holy grail of

35:37

the food world. If they could cheaply

35:39

create meat that didn't require slaughtering

35:42

animals, that would have a lot of obvious

35:44

benefits. But we have unfortunately just seen failure

35:46

after failure in this space. Totally. There's

35:49

been so much hype around this industry. In the early

35:52

2010s, there were enough sort of alternative

35:54

meat projects getting underway that people gave

35:56

it a name. They called it Schmidt. Do

35:59

you remember this? S-C-H-M-U-T,

36:01

shmeet. I love the word

36:03

shmeet. So shmeet did not

36:05

stick, but the efforts to make cultivated

36:07

meat did. The first lab-grown

36:10

burger was unveiled at a taste test in

36:12

London in 2013. According

36:15

to my colleagues at the New York Times at the time, the

36:17

meat itself was a bit dry and

36:19

a bigger barrier was that it cost

36:21

about $325,000 for a single burger. It's

36:25

so hard to fit that into the average family's

36:27

budget. So obviously

36:30

the promise and the hope

36:32

of these companies was that over time, the

36:34

cost of making cultivated meat would come down

36:36

radically. Eventually, it would be possible

36:39

to mass produce it and it would be

36:41

sold to customers at a price that

36:43

approximates or even maybe is cheaper than

36:45

the meat that we would all buy in the grocery store

36:47

today. A beautiful dream for shmeet. A

36:50

dream that so far has not panned out. Today,

36:53

there are estimated to be about 100 companies

36:55

working on cultivated meat around the globe.

36:58

In June, just this year, the US

37:00

became the second country in the world to approve

37:02

the sale of cultivated meat for two

37:05

different companies. But the

37:07

dream of mass produced cultivated meat

37:10

that is plentiful and cheap

37:12

enough for people to use it as a substitute

37:14

for slaughtered meat is still

37:17

a ways off. Today, we're going to talk

37:19

about why it has been so hard for

37:21

cultivated meat to hit the mainstream. And we're going to be talking

37:23

about it with Josh March. Josh is

37:25

the CEO and co-founder of Sci-Fi

37:27

Foods, which is a startup that is making cultivated

37:30

meat burgers by combining beef

37:32

that it grows in a lab with plant

37:34

based ingredients. He is very

37:36

optimistic that this future of cultivated meat

37:38

is coming. And he actually was very helpful

37:41

in helping us understand why it

37:43

has taken so long for this damn cultivated

37:45

meat to arrive on our plates. Yeah,

37:47

well, I can't wait to shmeet him. Okay,

37:51

that's enough for the word shmeet. I'm putting a moratorium

37:53

on that. Let's bring in Josh.

38:00

Josh Marsh, welcome to Hard Fork. Thank you. Great

38:03

to be here. So one thing that's different about your

38:05

approach to lab-grown meat, I guess cultivated

38:08

meat is like the industry term because it sounds

38:10

less weird than lab-grown or something like that. So

38:13

one thing that differentiates your approach from

38:15

a lot of other startups in this area is that

38:17

you are combining lab-grown

38:20

or cultivated meat with more plant-based

38:22

ingredients. Your burgers are

38:24

about 90% plant-based and

38:27

only about 10% of the cultivated

38:29

beef. So why did you choose

38:31

that approach? Yeah, I mean, look, ultimately

38:34

we're working towards a future where we can create

38:37

any kind of meat products 100% cultivated

38:39

from cells. That future

38:41

of doing that in a really large scale

38:44

and affordable way is still a bit of a distance

38:46

away. The reality is the

38:48

technology does not exist to do that at scale,

38:51

at low cost today. Right. I'm

38:53

very skeptical of anyone who claims otherwise. I

38:55

will put some cards on the table here and say

38:58

that cultivated meat has been a

39:00

category that has been, I would say,

39:02

disappointing for me. I do this column

39:05

every year called the Good Tech Awards where I single

39:07

out a bunch of startups that I think are doing good things for the

39:09

world. One year I featured a bunch

39:11

of cultivated meat startups and I talked

39:13

to them and they were all telling me, our products are about

39:15

to be on store shelves. You're about to be able to

39:18

eat this in restaurants, get ready. Cultivated

39:20

meat is coming. And then it just didn't.

39:23

It's too expensive. There turn out to be production

39:26

problems. There are all these companies that

39:28

have gotten raked over the coals for

39:30

exaggerating their production details

39:33

or claiming that they could make stuff for cheaper

39:35

than they were actually making it. What

39:37

is the sort of advance that

39:40

you guys have come onto that you think is going

39:42

to allow you to actually bring this stuff to market

39:44

and not just be like another cycle of empty

39:46

promises? Well, first of all, you're completely right.

39:50

That's really the reason I started the company. I mean,

39:52

after spending a few years on the sidelines

39:55

of the industry, I felt there was a lot

39:57

of very bold claims

39:59

and a lot of very like

39:59

hand-waviness that costs were magically going

40:02

to come down. And the more I learned about the technology

40:04

and the science involved, the more I realized

40:06

that that's not necessarily true. There's

40:08

some real biological, physical limitations, and

40:10

you need a pretty clear plan for how you're going to get

40:12

an affordable product. And it was clear to

40:15

us from the get-go that, first

40:17

of all, the only products that would be viable today

40:20

would be blended products, where the majority is

40:22

still plant-based. And you're really just using the cells, essentially,

40:25

as a flavoring ingredient. That solves a lot of the

40:27

technical as well as the cost challenges. When

40:30

we created the very first sci-fi

40:33

burger three, four years ago now,

40:35

it cost about $20,000 to create. That's

40:38

more than you'll pay at any restaurant in America, except

40:40

for Salt Bay's restaurant. That actually is cheaper

40:42

than the Salt Bay burger. And why is it expensive?

40:45

Fundamentally, you take a cell from an animal.

40:47

That cell has all kinds of behaviors and characteristics

40:50

that are really optimized for growing

40:52

in a cow, not optimized for growing

40:54

in a big steel tank. And so you have

40:56

to change that cell's behavior

40:58

and adapt them to grow in a cheaper

41:01

and more scalable way.

41:02

And it

41:03

was obvious to us, based on

41:05

really the background of our team,

41:07

that the fastest and most reliable way to do that

41:09

would be to use the power of genetic engineering, really

41:11

enabled by tools like CRISPR, which have really

41:14

changed the game when it comes to being able to make

41:16

tiny edits to animal cells. And

41:18

that was really the route to shift that behavior and

41:20

get to a scalable and affordable product. I

41:22

mean, it's just dawning on me how incredibly complicated

41:25

this is. I mostly write about software.

41:27

You have to solve an insane number of technical

41:30

problems just to grow a little bit of

41:32

this beef.

41:33

Yeah, and what was really important

41:35

to us and one of our guiding

41:37

principles as a company was that we

41:39

only wanted one miracle. So

41:41

I think a lot of startups fail, especially

41:44

in biomanufacturing, when you're relying

41:46

on multiple miracles. You need to engineer

41:49

new cell lines. You need to create new bioreactors.

41:52

You need to come up with cheaper downstream

41:54

processing. It's too many things.

41:56

And we felt all of that would be way too risky. And

41:59

so we knew. that if we could just

42:01

use CRISPR to engineer ourselves to have certain

42:03

kind of core KPIs in terms of how

42:05

they grow, we could make a 10%

42:08

product at an affordable price

42:10

that would be profitable to us at scale.

42:12

But we really tried to reduce the amount of risks

42:14

and miracles down as much as possible. And what was the one

42:16

miracle that you needed? I'm not sure I caught that. What was

42:19

the miracle that you're betting on? The one miracle is basically

42:21

could we use CRISPR, engineer

42:23

our cell lines to shift their behavior so

42:25

they can grow really effectively. It seems like CRISPR and meat

42:27

should have a sort of bond. It isn't

42:29

like a nice space to do what we're doing. Josh,

42:32

I'm going to ask a question that I already know the

42:34

answer to because I am obviously a very studied

42:36

food scientist and know all of the things that you're talking

42:38

about. But for any of our listeners who might

42:41

have questions, could you just walk me through

42:43

like how you actually make

42:45

lab grown meat? Yeah, yeah, absolutely. So,

42:48

yeah, we first, a number of years ago, took

42:51

a tiny bit of muscle biopsy

42:53

from a living cow. A vet did it, not

42:55

us. And it was just a tiny bit of muscle. The cow

42:58

went back running around it sealed afterwards. We

43:00

then take that to our lab. We isolate the individual

43:02

cells. We then have done years

43:05

of R&D to adapt and

43:06

engineer those cells

43:08

to get them really comfortable growing in our efforts.

43:11

This is the CRISPR part. That's the CRISPR part. So,

43:13

CRISPR, from my basic understanding

43:15

as an expert food scientist,

43:18

is a way to kind of snip material

43:20

out of certain genes and insert it into others.

43:23

It's like a gene editing tool. It's

43:25

like a gene editing. But what's great about CRISPR is

43:27

it allows you to make very targeted tiny

43:29

changes. Like literally most of the things we do

43:31

is we delete a single base pair of DNA

43:34

and that just stops like a protein being expressed. We're

43:36

not putting any foreign DNA. We're not taking

43:38

like DNA from salmon and putting it into beef.

43:40

All we're really doing is just shifting

43:43

the expression of proteins in beef. So what

43:45

are you doing with CRISPR on these

43:47

beef cells taken from the cow? So,

43:49

as an example, when you take a cell

43:52

from a cow muscle, right, that cell only

43:54

wants to grow when it's attached to other

43:56

cells or surface.

43:58

So you can grow it in the lab and in Petri dish. It's

44:00

like a single layer of cells. It's a tiny amount.

44:02

Very, very expensive. But what

44:04

you really want to do for manufacturing is you want it to be

44:07

growing just floating in liquid in a big steel tank. You

44:09

want it to look kind of like

44:10

you're fermenting wine or bacteria, yeast, that

44:12

kind of thing. Just a giant floating mass

44:15

of beef and water. Sounds very appetizing. Yeah.

44:18

Almost like a soup. Tasty, really

44:20

tasty beef soup. It's a sous vide. Yeah,

44:22

exactly. But if you just take like a normal

44:24

cow muscle cell and put it into that condition, it doesn't

44:26

like those conditions. It will die. So

44:29

we figured out why we could use CRISPR to make a few

44:31

different tiny changes that in combination would

44:34

make the cell comfortable growing in liquid

44:36

in suspension, it's called. Interesting. So

44:38

then you have these cells that can grow in liquid.

44:41

You let them grow. Yeah. So

44:44

basically, what is a bioreactor? A bioreactor is basically

44:46

just a big steel tank that

44:48

is extremely clean, perfectly sanitized,

44:50

not a single bacteria in there until you put the

44:53

cells in. And then it allows you just to carefully

44:55

control the conditions. It allows you to add different

44:57

nutrients and feeds. You're feeding the feeding the

44:59

cells, sugars and amino acids and vitamins,

45:02

all the stuff that life needs to exist.

45:04

And it's just the right conditions so that

45:06

the cells keep growing. And so essentially,

45:08

once we've done those years of R&D, we then have

45:11

a cell line that grows really well. And

45:13

each time we want to manufacture it,

45:15

we start from a tiny variety of cells and we

45:17

start expanding it in larger and larger vessels until

45:20

eventually it's in a big production bioreactor and

45:22

it keeps doubling every day or so. And then

45:24

eventually after like a week in the big production bioreactor,

45:26

we can harvest it, send the future down,

45:29

and we have a lot of tasty B cells.

45:31

Got it. And how much meat can you currently

45:33

produce this way in like a week? Out

45:36

of our biggest bioreactor, we could produce

45:38

like

45:39

two and a half thousand grams or so a

45:40

week. So two kilograms? Yeah,

45:43

a couple of kilograms a week. Great math, Kevin. Yeah.

45:46

That's how it goes. We love math

45:48

challenges on the show. So that is not

45:50

a lot of meat. But we have a very small pilot

45:52

plant here in the Bay Area that

45:54

we just finished construction of. You

45:57

know, this is not a large scale commercial facility yet,

45:59

but we will be able to do that. commercializing out of the pilot

46:01

palm, which we're excited about. And when you do commercialize,

46:03

when you're at scale, what do you think a pound

46:05

of lab-grown meat will cost you to produce?

46:08

Like, what is it going to take? My understanding

46:10

is even though the prices have come down, it's still very

46:12

expensive to make cultivated meat in

46:15

these bioreactors. I think

46:17

a lot of people see prices, the main

46:19

barrier here, regular beef, if you go

46:21

to the supermarket, you're paying, you know, five,

46:23

seven, maybe 10 bucks a pound for that. What

46:26

does it cost you to produce a pound of meat

46:28

this way currently? And then what do you need to get

46:30

the price down where it's maybe equal to or

46:33

almost equal to the price for regular beef? We like to think

46:35

of everything as like cost per burger. You

46:37

know, I mean, quarter pound of burger, obviously. Out

46:40

of our pilot plants, we just

46:42

finished construction of, we'll be launching

46:44

commercially out of that end of next year. And

46:47

we think we'll be at about $30 per

46:49

quarter pound of burger. That's the 90% plant, 10% sales

46:53

out of the pilot plant. From a larger

46:55

scale commercial facility, our goal is to get it down

46:57

to a dollar a burger. And we know exactly

46:59

what we need to do. It's mainly just around the performance

47:02

of the cells. And we know that that kind of performance

47:04

is possible from

47:05

what people have done with other cells and other species. But

47:08

we think it's very biologically possible. And it's just

47:10

a couple of years more of development to get us down to about the

47:12

dollar burger cost. Now that's our cost.

47:14

So we'd be selling about $2 burger, so it's

47:17

a wholesale price of about $8 a pound. And

47:19

then what if any sort of regulatory

47:21

hurdles are there? Because my understanding is that the

47:24

regulators are being very

47:26

slow to kind of give approval to cultivated

47:29

meat to be sold in grocery stores and

47:31

at restaurants. So what needs to happen

47:33

for this to be sort of legal

47:35

and sort of widespread? You know, I think we're actually

47:37

in a really great place in the US in terms

47:40

of how forward thinking and supportive

47:42

the regulators are. The FDA and the USDA

47:45

have this joint regulatory framework where

47:48

the FDA regulate the cell lines and the

47:50

upstream process of growing the cells. Once

47:52

you harvest it, the USDA regulate

47:54

the harvest process and the facility.

47:57

They basically treat it like you've sorted an animal

47:59

at that point when you harvest it.

47:59

cells. And you

48:02

have to have

48:03

a pretty developed process. You

48:05

have to demonstrate that you can consistently

48:08

grow the same cell line at

48:10

a reasonable scale, and you have to be able to do

48:12

a lot of food safety tests and show that there's

48:14

no contamination. You have to submit safety

48:16

dossiers, and the regulators want to look

48:18

at the final product that you're creating and check you're

48:21

saying it's beef, is it actually, is it fats

48:23

and all kind of stuff similar to beef. And

48:25

so you have to submit quite a lot of information. Once

48:28

you've actually submitted it, although generally the

48:30

FDA will give you an approval letter in nine to 12 months.

48:32

They've already done that for two companies, and

48:35

then the USDA will come approve the facility.

48:37

Honestly, I don't think the delay has been on the

48:39

regulatory side. The delay has mainly been companies

48:42

getting to the stage where they're ready to submit.

48:45

I want to ask about the culture piece

48:47

of this too, because I'm very

48:49

optimistic that you all will

48:52

eventually come up with something that,

48:54

you know, Casey or I or others sort

48:56

of like coastal, you know, our

49:00

snobs will eat and feel virtuous

49:02

about. But I want to read you a tweet

49:05

that I saw, I guess we're calling them posts now

49:07

from Ronnie Jackson, who is President Trump's

49:10

former doctor who is now in Congress in

49:12

Texas. And one of your favorite ex accounts to follow.

49:15

He said, and I quote, I will never

49:18

eat one of those fake burgers made in

49:20

a lab. Eat too many and you'll turn

49:22

into a socialist Democrat. Real

49:25

beef for me. You know,

49:27

that's funny as eating real beef turned me into a socialist

49:29

Democrat. So I think it's just different effects

49:31

on different people. So but I think there's a serious

49:33

question, which is like, it is clear

49:36

to me and to a lot of other people that the barrier to

49:38

a culture where we all eat cultivated

49:40

meat is not just the technology or the cost.

49:42

There's also something sort of intrinsically

49:45

linked to kind of a feeling that people have where

49:47

things that are grown in labs are weird and beef

49:49

that comes from cows is natural. And

49:51

so why would you swap out the thing that is natural

49:54

for the thing that is weird? So how do you market

49:56

something like this? Or how do you think about positioning this

49:58

so that, you know, someone's someone like

50:01

Ronnie Jackson, maybe not him specifically, but

50:03

so that... No, let's focus on Ronnie Jackson. If

50:05

he can get Ronnie Jackson, we're winning. Yeah. What's

50:07

your Ronnie Jackson strategy? Yeah. Well, first

50:09

of all, I think this is a really important topic. More

50:12

than almost any other food, and especially red meat, has

50:14

a lot of emotional, symbolic

50:17

connotations. And so I do think you

50:19

have to be quite careful. And I think this is one of

50:21

the challenges that plant-based meat has had by

50:23

just being like, meat can be plants. And people are like,

50:26

no, it can't. And so

50:28

I think, again, that's part of the reason

50:30

for taking a different tact from a branding. And

50:32

the fact that we can be more pro meat, I think, is really

50:34

important. And I think there is this message

50:37

that like, hey, yeah, if you can

50:39

go and hunt all your meat, great. But

50:41

let's be honest, most of us do a hunter

50:43

gathering in the grocery store. Right? And

50:45

if you're doing a hunter gathering in the grocery store, we'd

50:48

rather be able to create meat in this new,

50:50

awesome, fun, exciting way without factory

50:52

farming and all the other crap that comes with it.

50:55

And I think it has to be fun. We

50:58

think about this a lot. Even

51:00

though a lot of my motivation may be like, climate

51:03

and animal welfare, that's not most people's motivation

51:05

when it comes to eating a burger. And in fact, even

51:08

for most people who intellectually agree

51:10

with those things, when it comes to ordering

51:12

a burger on Friday night, they're

51:14

not thinking about those things. It's a very lizard brain,

51:17

emotional decision to get

51:19

the tasty burger. And so we think

51:21

it's really important to just be fun. This is

51:23

part of the reason for our branding. Yeah, call

51:26

it sci-fi foods, because we think that people

51:28

are going to think this is sci-fi and weird. We don't think we can

51:30

avoid that. So

51:31

leave into it. Many lovely journalists still call

51:33

this lab grown meat, as we've heard on this podcast

51:35

today already. Shame on them. So we don't

51:38

think we can avoid it, right? We know the meat lobby is

51:40

going to come after it hard for being frank in meat. So

51:42

we have to take something that people are going to think is sci-fi

51:45

and make it safe and fun. And we think we can do

51:47

that by kind of leaning in a bit of a wink and

51:49

just getting people excited that this is a cool new way to

51:51

make real meat. I remember years ago,

51:53

I went to what at the time seemed like it was

51:55

one of the hottest new restaurants in New York, which is called superiority

51:58

burger. Do you remember superiority burger? Yeah. So it was

52:00

a place where you could get plant-based burgers,

52:03

and everybody was like, you have to have this burger. It

52:05

is incredibly good. And so I was visiting

52:07

New York, and I went, and I had the plant-based burger. And

52:09

I really enjoyed the experience, but I think 90% of

52:12

it was just the taste of cheese and ketchup. But

52:15

I wonder if that presents an opportunity

52:17

to you. For the most part, people are not eating just

52:19

plain beef patties. They're smothering

52:22

it in things that taste very good. And

52:24

so maybe it doesn't matter as much if it isn't

52:26

an identical experience to eating a beef

52:29

burger. Yes, but one of the

52:31

challenges with a lot of the kind of plant-based burgers

52:34

is that you just have this kind of off taste of

52:36

plants and like serially, kind of few

52:38

in the industry, they've got serially off notes and planty

52:40

off notes. We hate serially off notes.

52:43

People do. It's kind of amazing. When they want

52:45

a burger. And that's actually one of the biggest

52:47

things that our product doesn't have. Well,

52:49

then maybe it's time to taste the darn product. What

52:51

do you say, Ken? Let's taste the darn product. You have brought

52:54

some of your burgers to

52:56

give us to try. Well, actually, we wanted to do Thanksgiving

52:59

themed. That's right. So instead of doing burgers

53:01

and smothering and ketchup, we have meatballs with

53:03

a cranberry glaze. Oh, that's great

53:05

to me. Sounds awesome. You can't talk about meat

53:07

for this long and not be a little hungry. Yeah. All

53:10

right, let's get some of this meat in here.

53:12

All right.

53:17

Amazing. All set? Yes,

53:19

thank you, Chef. Thanks, Chef.

53:21

Well, this is exciting.

53:23

Never done this format before. So this

53:25

is a special for you guys. Great.

53:27

Well, thank you for helping us get into the Thanksgiving spirit.

53:30

So we are here. And

53:32

in front of me are three meatballs

53:35

that are made out of your cultivated

53:37

beef plant mixture. And

53:40

there's a lovely looking cranberry

53:43

glaze on here. And I'm very

53:45

excited to dig in. But Casey, before

53:47

we dig in, we actually have to do something that I've never

53:49

had to do before I eat a meal before, which is to

53:51

sign a waiver. So Josh, what is this waiver

53:53

that you are having us sign? So this basically

53:55

just says this is a novel food. It's not yet

53:57

approved for commercial sale.

53:59

And we've done our own internal

54:02

safety assessments. We believe this

54:04

food to be completely safe. But because

54:06

this is not approved for commercial sale, essentially

54:09

you have to do this at your own risk.

54:11

And so that's basically what the way this

54:13

says. Now, I don't have time to read this whole thing.

54:16

Can you just tell me, am I going to die from this? Definitely

54:19

not. Okay. I've been

54:21

eating... I'm not prepared to die for the podcast,

54:24

but I would just like to know. I've been eating

54:26

this now for multiple years, regularly,

54:29

and we've done over 100 tastings and have zero

54:32

adverse effects. Okay. Wow. Okay.

54:36

So I have to... Okay. We're

54:38

going to sign the form here. All right. We

54:41

let Jesus take the wheel. Yeah. Let

54:44

Josh take the wheel. Okay. All

54:46

right. So we have our forks here. Yeah.

54:49

And Kevin, you actually didn't mention when you were describing this. Each of these

54:52

meatballs, which are quite appetizing looking, are sitting

54:54

in a cloud of what I believe is mashed potatoes.

54:57

Oh, that looks great. Very good. They're a little

54:59

sage leaf on top. All

55:01

right. Should we try it? Let's go.

55:04

Let's try it. Here goes nothing. Mm-hmm.

55:07

Hmm. It's good. Mm-hmm.

55:10

It is good.

55:13

And there is a beefiness to it.

55:16

Yes.

55:20

It didn't hit me right away, but I chewed it a little bit,

55:22

and then I got the... The texture is nice and

55:25

beef-like. Yeah. Which I appreciate.

55:27

Which you love it when people say that about you. Yeah.

55:33

So this might be... Well, I was going

55:35

to say this is the most science that has ever

55:37

gone into anything that I've eaten, but then I thought, I

55:39

eat Cheetos. And we all know

55:42

that nothing

55:44

in America has been engineered more than a Cheeto. Yeah.

55:48

Nothing in a Cheeto came out of the ground. They don't grow their Cheetos in fields. No.

55:51

I'm going to tell you. This is the sort of God's plan

55:53

for this world. Here's

55:56

my question. So, again, I

55:58

think there's good beefy flavor. in here.

56:00

I think that if I did a blind taste test and you

56:02

had a pure beef meatball and this

56:04

meatball, I think I would enjoy both, but I would

56:06

be able to tell the difference. What do you

56:08

think would be the difference in

56:11

taste if this were 100% cultivated beef versus

56:15

what I'm eating right now? Yeah, then you wouldn't be able to taste the difference

56:17

at all. Really? Okay. Yeah, and

56:19

we do... Because practically there is no real difference, right? No,

56:21

it's the same cells, the same fats, the

56:23

same things that create that flavor. So what

56:26

we find is that 10% makes a big difference.

56:28

You would taste the difference clearly between plant-based as

56:30

well. It is like I said, if you had to do

56:32

a blind taste without

56:34

beef, you would think it's beef. If you compare it directly

56:36

to 100% conventional beef, you're like, oh, I

56:38

can tell the difference a little bit. The more we add,

56:42

the smaller that difference becomes and

56:44

there's other things we can do to improve the flavor. So

56:46

if we're at 20% cultivated meat or 30%, it would

56:48

taste more like beef, but it would also be significantly more expensive.

56:51

But it's also much more expensive. Yeah, and so for us, it's just like, hey, this is

56:53

a journey that we're on. And we think we can be completely

56:56

transparent with consumers, just the 10% products.

56:58

And at some point when technology evolves,

57:00

we'll be able to go to the 20 and the 30 and

57:02

the 40 and eventually 100. Right.

57:05

And do you see a path to work on meats

57:07

other than beef? Like is there... Yeah. ... will

57:10

there be chicken and lamb and... Yeah, 100%. Beef

57:12

for us is really the holy grail. When

57:14

you think about people wanting to eat

57:16

different meat products, like mainly red meat,

57:19

and from a climate change perspective, certainly

57:21

beef is just like the biggest contributor by

57:23

a long way. And ground beef also

57:25

has the highest price per pound of

57:28

any ground meat products. And it's one of the biggest markets

57:30

from a revenue perspective of any meat products. So

57:33

beef is the really best place to start. But

57:36

our same platform approach can apply to any species

57:38

and we intend to keep going from here.

57:41

Well, Casey, do you feel like a socialist Democrat yet?

57:44

I don't know. I just ate a plate of meatballs

57:46

worth $90,000. No, no, no.

57:48

I think that makes me a capital... Yeah,

57:51

what would these cost today if you were

57:53

selling them in a store? Maybe

57:55

like $100. Okay. Wow. Yeah,

57:57

it's a nice lunch. That's actually about

57:59

the cheap. lunch you can get in the financial district of San Francisco.

58:03

That's true. I like these. They don't

58:05

taste exactly like meat to me, but they do taste

58:07

better than a lot of the alt meats that I've tried.

58:10

Casey, can you imagine eating this like in a year

58:12

or two as part of your normal diet? I

58:16

mean, yes, like I can. I think

58:18

if I'm thinking about, you know, I have friends

58:21

who are vegetarian, maybe we go to a vegetarian restaurant,

58:23

I see this is on the menu, like I can imagine

58:25

ordering it. Yeah. You're in the clean plate club over

58:27

there. I gotta, I gotta step it up. The other thing

58:29

I realized as I was eating was that I was actually quite hungry.

58:32

Podcasting takes a lot out of people. People don't know that. People

58:34

think, oh, it must be so easy to sit down, run your mouth.

58:36

No, you're burning a lot of calories. Well, Josh, it's

58:38

a very cool demo. I enjoyed

58:41

my meatballs and I think Casey did too, judging

58:43

by the fact that we all finished all three of them. Yeah,

58:46

we did. We were talking about two demos on the show. This was

58:48

the only one I could eat and so I did actually

58:50

prefer it. Josh, what

58:52

are you serving at Thanksgiving this year? You

58:55

know, I have a big group Thanksgiving with a load

58:57

of friends and I'll be allocated something.

59:00

Normally I do the green bean casserole. Well,

59:04

this would this would work in a pinch if the turkey, you

59:06

know, burns in the oven or something. Thank

59:09

you so much for coming. Really good to talk to you. Thanks,

59:11

Josh.

59:28

They have the Audi craftsmanship you're looking for,

59:30

along with the eye-catching Audi designs you desire.

59:33

They

59:34

even come equipped with the confidence of a century's

59:36

worth of Audi expertise. And

59:38

if you happen to get nostalgic, you

59:40

can still take your Audi e-tron to a gas station

59:43

for a drink or a snack. How

59:46

you get there matters. Of the five

59:48

fully electric Audi e-tron models

59:50

available now, getting there has never

59:52

been easier.

59:54

Audi.

59:54

Progress you can feel. Learn

59:57

more at AudiUSA.com. KC

1:00:02

has some feedback for you about our YouTube channel. Oh, great.

1:00:05

Glad I'm sitting down. I was watching one of our videos,

1:00:07

which are looking great, by the way. Love our channel. People

1:00:10

should go subscribe. Absolutely. It's

1:00:12

a good time. Good time. This swiveling in your

1:00:14

chair has to stop. I'm going to disable the swivel in your chair.

1:00:17

Listen, they've already disabled everything else. This chair

1:00:19

used to have wheels on it. It's true. And

1:00:21

so I used to get to kind of move around, and now I'm locked

1:00:23

into place. I don't get to have my little stand

1:00:25

for my laptop anymore. All I have is a little

1:00:27

swivel. I was watching our video, and I was watching

1:00:29

you, and I was watching you just rock gently back

1:00:31

and forth, and I was feeling like I was on

1:00:33

a ship. Just

1:00:37

the subtle swaying of your body back and

1:00:39

forth. What if I told you that it's a fundamental

1:00:42

element of my creative process? You know how

1:00:44

Cyclops has to wear the visor to prevent himself

1:00:46

from shooting people with lasers all the time? That is

1:00:48

what I'm doing while I'm swiveling. What are you preventing

1:00:50

via swiveling? I am preventing myself from interrupting

1:00:52

you even more than I already do. So

1:00:55

that is what I'm doing. Well, if that is the reason,

1:00:57

then I guess I'll accept it, and our YouTube audience will just

1:00:59

have to deal. Thanks, YouTube. Sorry for

1:01:01

the swivels. It's okay. Hard

1:01:03

Fork is produced by Rachel Cohn, Davis Land,

1:01:06

and Emily Lang. We're edited by Jen

1:01:08

Poyant. This episode was fact-checked by

1:01:10

Caitlin Love. Today's show was engineered

1:01:12

by Daniel Ramirez. Original

1:01:15

music by Diane Wong, Rowan Nemesto,

1:01:17

and Dan Powell. Our audience editor

1:01:20

is Nel Gilogli. Video production

1:01:22

by Ryan Manning and Dylan Bergeson. Special

1:01:25

thanks to Paula Schumann, Pooing Tam, Kaitla

1:01:27

Presti, and Jeffrey Miranda. You

1:01:30

can email us, as always, at hardfork

1:01:32

at nytimes.com. Anybody

1:01:34

got any lab-grown meat recipes? We're listening.

1:02:03

They have the Audi craftsmanship you're looking for,

1:02:06

along with the eye-catching Audi designs you desire.

1:02:09

They even come equipped with the confidence of a century's

1:02:12

worth of Audi expertise. And

1:02:14

if you happen to get nostalgic, you

1:02:16

can still take

1:02:17

your Audi e-tron to a gas station

1:02:19

for a drink or a snack. How

1:02:21

you get there matters.

1:02:23

And with five fully electric Audi e-tron

1:02:25

models available now, getting there has

1:02:27

never been easier.

1:02:29

Audi. Progress you can feel.

1:02:32

Learn more at AudiUSA.com

1:02:35

slash electric.