0:00

Hey upstream listeners, today on

0:02

upstream we're revisiting this classic

0:04

interview with Kevin Kelly, founding

0:06

executive editor of Wired Magazine,

0:08

Technologist, and author. This is a discussion

0:10

I turn to repeatedly. Not just for

0:12

Kelly's concise wisdom and optimism, but especially

0:14

now as we navigate how far technology

0:17

will take us in the age of

0:19

AI. Up ahead, you'll hear Kelly discuss

0:21

what makes us uniquely human and how

0:23

we've invented our humanity. Please

0:25

enjoy. Kevin,

0:41

welcome to the podcast. Excited to have

0:43

you on. Oh, it's always a pleasure

0:45

to chat with you, and I'm really

0:47

delighted to be invited back. Thank you.

0:49

We're excited to dive into your new

0:51

book, which is an advice book, different

0:54

kind of book. You read all advice

0:56

books that are written a long time

0:58

ago, and a lot of those principles

1:00

still make sense. Are there certain

1:02

genres of advice that you're inspired

1:05

by, or see yourself in the lineage

1:07

of, or how do you think about

1:09

advice as a category? you know, vice

1:11

columns. A lot of advice is wrapped

1:13

very, very smartly, I say,

1:15

in stories. We communicate

1:17

best in stories. And most

1:19

of vice books will have stories

1:21

about these things. And that is a

1:24

really great way to convey it

1:26

because you're kind of showing it

1:28

rather than telling it. And then

1:30

you will remember it. I'm not a

1:32

very good storyteller. So I

1:35

decided to do it my way, which is

1:37

these telegraph. aphorisms, these

1:40

adages, these lessons, these little

1:42

bits, these proverbs, which

1:44

is suits me and is how I

1:46

like to consume the advice. I like

1:49

to collect little quotes and

1:51

things and they work for me

1:54

as reminders. So that's what

1:56

I produced because I'm no good

1:58

in making a story. but

2:00

I can make a little bit

2:02

of viral tweeting and so that's

2:05

what I did. Let's go through

2:07

some of them. You mentioned gratitude.

2:09

You have one, gratitude will unlock

2:12

all other virtues and it's

2:14

something you can get better at.

2:16

Talk about how it unlocks other

2:18

virtues and talk about how you

2:20

get better at. I think in a weird

2:23

way, I think like gratitude

2:25

and trust and empathy are

2:27

connected. There almost maybe.

2:29

different faces of the

2:32

same virtue in some way,

2:34

where gratitude you kind

2:36

of are acknowledging your

2:39

unspecialness, your luck,

2:41

your acknowledging that you

2:44

did not earn in a certain

2:46

sense, that it was a gift.

2:48

And trust is in some

2:50

ways in my own mind connected

2:53

in that same way, or

2:55

gratitude is a

2:57

form of trust, maybe same.

2:59

or is this form of empathy,

3:02

where you're kind of being

3:04

able to put yourself in

3:06

someone else's position? So in

3:09

a way that I haven't

3:11

really articulated, I think those

3:13

three qualities are very

3:15

much joined in a way, and

3:18

they're kind of all

3:20

expressions of something similarly

3:22

deep in this kind of

3:24

connection that we have. not just

3:27

in a superficial way

3:29

between people, but that

3:31

we are actually of the

3:33

same life. I mean, the life that

3:35

we share is the same

3:37

exact life as we have

3:40

a common ancestor, no matter

3:42

who we are. And so I think

3:44

that's what I'm trying to

3:46

get at is that, you

3:48

know, being thankful for what

3:50

it is for our lives

3:52

or for what we have, which

3:54

are given to us, Very similar to

3:56

the kind of trust that we would have another person

3:59

that they're going to do well, which is kind

4:01

of a, the same kind of sense of

4:03

us having empathy for another, saying I'm

4:05

connected to you, we are in this

4:07

together, there is something in common that

4:09

we have, and when you work on

4:12

gratitude, it's kind of a way to

4:14

work on those other ones at the same

4:16

time. There's this old, my angelo quote,

4:18

of people will forget what you say,

4:20

but they'll remember how you made them

4:22

feel. And I used to think that was

4:25

kind of like a bug, like, isn't substance

4:27

really important, but People are

4:29

operated that way and you have an

4:31

aphroism here. You know, when choosing between

4:33

being right and being kind, choose being

4:35

kind. And it just kind of speaks

4:38

to what people really care about is,

4:40

are you with them? Are you part of

4:42

the tribe? Are you aligned with them? And,

4:44

you know, arguments are abstract. They're

4:46

kind of changing. They're evolving. I'm taking

4:48

a little bit of an abstract kind

4:50

of meta commentary on why that might

4:52

be true. Why people care more about

4:54

how you make them feel than what

4:56

you actually say, what you actually say. or

4:58

why they care more about, you know, being

5:00

kind of being right. It's sort of

5:03

like, why do we remember smell so

5:05

much, I think, because there is a

5:07

way in which they're kind of, they're

5:09

wires further down into our

5:12

brainstem, the experiences, it's fundamental.

5:14

So, so it's not like there's,

5:16

in one of the emotional components,

5:18

sort of like, not an overlay,

5:21

it's more like an underlay. Yeah,

5:23

the way that we have these

5:25

associations that they color everything, so

5:27

we can have. trigger responses that

5:29

are way below or way faster than

5:32

our intellectual intelligence.

5:34

I think we tend to

5:36

overrate intelligence. There was a whole

5:38

bunch of studies done about

5:40

people making decisions and then

5:42

how it reflected or didn't

5:44

their gut responses where they

5:47

would make a decision without thinking,

5:49

like on first impressions, and

5:51

then they would correspond that

5:54

with whether they were correct or

5:56

not correct or not correct or

5:58

not. Also particularly long-term things. but

6:00

character evaluations and

6:02

their first impression,

6:04

were usually the most accurate

6:07

impression before they even

6:09

applied logic, was faster than

6:11

logic. So I think there's something

6:13

going on about that, the

6:15

way they were constructed, where

6:18

the primeval circuits of first

6:20

impressions and responses and

6:23

feeling, they kind of work

6:25

faster. They work faster. They work

6:27

stronger. than the rational thing.

6:29

And a lot of what our humanity

6:32

is about is trying to

6:34

not just overcome that, but I

6:36

would say manage it or steer

6:38

it or the times elevate

6:40

it because a lot of times

6:43

it's absolutely right. And lots

6:45

of times it's not the best

6:47

thing. I mean, we have anger

6:49

and other emotions that are also

6:51

kind of have the same kind

6:53

of power. So yeah, so for

6:55

whatever reason we're governed by. by

6:58

these things that seem to trump

7:00

what the logic says and that's

7:02

another piece of advice about you

7:04

know arguing can't rationally argue

7:07

something somebody out of

7:09

an opinion that they didn't rationally

7:11

or you know get into through

7:14

rationality and so when you're

7:16

having disagreements the emotional

7:19

component is incredibly

7:21

important in changing by

7:23

someone's mind or convincing them

7:26

or persuading them or Yeah, you

7:28

can't reason someone out of a notion. They

7:30

didn't reason into it. I was gonna bring

7:32

that up. It's fascinating. And what you're usually

7:34

arguing about is not actually the thing that

7:37

you're arguing about. There's often a subtext to

7:39

the argument. And I'm curious, is your mental

7:41

model of emotions something along the lines of

7:43

almost feedback loops of what we're more likely

7:46

to help us thrive within our tribes like,

7:48

you know, a thousand years ago or whenever

7:50

it was? use some of those sort of

7:52

emotions are still adapted to modern environment some

7:55

of them are no longer adapted but is

7:57

that fundamentally what they were doing was kind

7:59

of intuitions of about what was more like

8:01

to raise our reputation in a hard

8:03

tribe or? Yeah, I think so. That

8:05

if you look at, there was some anthropological

8:08

work studying the last,

8:10

you mean the intact tribal people.

8:12

And it was really, some of this,

8:14

we went very deep and lived with

8:17

them for a long time and got

8:19

to know them very well and were

8:21

able to observe them for long

8:23

periods of time in all their dimensions.

8:25

It was really, really fascinating.

8:27

There was, I think Colin

8:29

turn. Turnbull, one of the

8:31

famous anthropologists, lived with, at the

8:34

time they were called Pygmy People, I

8:36

don't know the proper current, proper name

8:38

for them. And what was interesting

8:40

was, they lived in these grass leaf, leaf,

8:43

leaf, huts, basically had no

8:45

walls. Everybody heard everything in

8:47

the little clan in the settlement. They

8:49

would have these, like any humans,

8:51

like any humans, they would have

8:53

these, like any humans, they would have

8:55

these, like any humans, they would have

8:58

these, arguments and fights

9:00

over all kinds of you know things

9:02

that people argue and fight

9:04

about but they're really there

9:06

but but the and everybody could

9:08

hear everything and they could hear

9:11

those both sides and the whole

9:13

drama but the the interesting

9:15

was well while they could be

9:17

very very vocal and they almost

9:19

get to the point where they're

9:21

like fist fights and almost

9:23

hurting each other but they would

9:26

never go all that way

9:28

of really entering one another.

9:30

It was a lot of releasing.

9:33

It was kind of like

9:35

therapeutic fighting in a

9:38

certain sense, but they

9:40

always had certain lines

9:42

and the lines was they

9:44

would never do anything

9:46

that jeopardize the health

9:49

of the group. Okay. And

9:51

so there's something sociological

9:54

in us as a.

9:56

social animal

9:59

that as primates,

10:01

that even though we

10:03

have individuality, we've

10:05

involved to also pay

10:08

close attention to the health

10:10

of the group and the

10:12

status of the group. And

10:14

so maybe that's what makes

10:17

us kind of unique is

10:19

that we're social animals

10:21

as well as individuals. And

10:23

we have these competing instincts.

10:26

You have a frozen here.

10:28

A great way to understand

10:30

yourself is to seriously reflect

10:32

on everything you find irritating

10:34

in others. What do you impact that

10:36

one? Yeah, it has to do, and some of

10:39

the other ones, particularly these

10:41

days, with our new invention of AI, you

10:43

have to do with the fact that we

10:45

humans and individually are very,

10:48

very opaque to ourselves that

10:50

we... Collectively use the species.

10:52

We really don't know what makes us

10:54

human, how our minds work. That's

10:56

part of the thrill and excitement

10:59

and the worry about what's going on

11:01

with AI is that we're trying to

11:03

make something, replicate something that we

11:05

don't even know what it is,

11:07

our intelligence. But there's also

11:10

individually. This was one of the

11:12

lessons of the quantified self was

11:14

that we just don't have a very

11:16

good intellectual understanding

11:19

of. how our own minds work, how we

11:21

work, where we're coming from, why

11:23

we even make decisions that we

11:25

make. And looking at one of the

11:27

ways we can kind of look into

11:30

ourselves is find out what kind

11:32

of agitates us and what kind of,

11:34

where we care and where we're paying

11:37

attention to ourselves. Again, we're

11:39

kind of, that's another bit

11:41

like, pay attention to where you

11:43

pay attention to. And so our, you know,

11:45

consciousness is a very. Gossip

11:47

or a thing. It's a very slippery

11:50

thing. It's kind of like the

11:52

only tool we have to try

11:54

and probe us, but it's not

11:56

that dependable itself.

11:59

And so... So this is just another

12:01

way of paying attention to what

12:03

perks is another way to kind of

12:05

help us to get a view of us.

12:08

It's not necessarily saying you are

12:10

like that, just saying this is

12:12

a signal. This is another data

12:14

point. You can use to try and

12:16

dissect yourself because we are

12:18

really hard to self dissect.

12:21

Competing instincts, competing

12:23

emotions. People get anxious

12:25

about that, but that's kind

12:27

of the default state. you have

12:29

one, whenever you can't decide which

12:31

path to take, pick the one that

12:34

produces change. And I'm curious how

12:36

you balance that with sort of

12:38

like commitments and stick to whether

12:40

it's a partner or whether it's a

12:43

long-term job or. Well, yes, I think there

12:45

are, you know, it's a kind of elsewhere

12:47

that, you know, except for like marriage

12:49

always, it's always to think about

12:52

the exit first in some ways.

12:54

And whenever we're arguing a business

12:56

deal, it's exactly. where I start

12:58

with is, okay, what does the exit

13:01

of this look like? So we can

13:03

structure the beginning of

13:05

it. Yeah, so I think there

13:07

are some exceptions where you have

13:09

a commitment that's sort of

13:11

the nature of the commitment is like,

13:14

yeah, despite all the opportunities,

13:16

by all the other choices that

13:18

I'm not going to make that

13:20

choice. So I think there is, or

13:23

I think there is bounded in that

13:25

sense. But outside of that.

13:27

boundary, there's still many choices that

13:29

we have where we have, you know,

13:31

turns occupation where we live. We'll continue

13:33

our interview in a moment after a

13:36

word from our sponsors. Even if you

13:38

think it's a bit overhyped, AI

13:40

is suddenly everywhere. From self-driving

13:42

cars to molecular medicine to business efficiency.

13:44

If it's not in your industry

13:46

yet, it's coming fast. But AI

13:48

needs a lot of speed and

13:50

computing power. So how do you

13:53

compete without costs spiraling out of

13:55

control? Time to upgrade to the

13:57

next generation of the cloud. Oracle

13:59

cloud infrastructure. or OSI. OSI is

14:01

a blazing fast and secure platform

14:03

for your infrastructure, database, application development,

14:05

plus all your AI and machine

14:07

learning workloads. OSI cost 50% less

14:09

for compute and 80% less for

14:11

networking. So you're saving a pile

14:13

of money. Thousands of businesses have

14:15

already upgraded to OSI, including Vodafone,

14:17

Thompson Reuters, and Suno AI. Right

14:20

now, Oracle is offering to cut

14:22

your current cloud bill in half

14:24

if you move to OSI for

14:26

new US customers with minimum financial

14:28

commitment. Offer ends March 31st. See

14:30

if your company qualifies

14:33

for this special offer

14:35

at oracle.com/turpentine. That's oracle.com

14:37

slash turpentine. Every time I hop

14:39

on the 101 freeway in SF, I

14:42

look up at the various tech billboards

14:44

and think, damn, turpentine should

14:46

definitely be up there. And

14:48

I bet I'm not alone. In

14:50

a world of endless scrolling and

14:52

ad blockers, sometimes the most powerful

14:54

way to reach people is in

14:57

the real world. That's where AdQuick

14:59

comes in. The easiest way to

15:01

book out-of-home ads, like billboards, vehicle

15:03

wraps, and airport displays, the same

15:05

way you would order an Uber.

15:08

Started by Insticard Alums, Adquick

15:10

was born from their own

15:12

personal struggles, to book and measure ad

15:14

success. Adquick's audience targeting,

15:17

campaign management, and analytics,

15:19

and analytics. Bring the precision

15:21

and efficiency of digital to the

15:24

real world. Ready to get your

15:26

brand the attention it deserves. QUICK.com.

15:28

Make the case for why

15:30

people should consider having more

15:33

kids than they think. Oh my

15:35

gosh. Well, first of all, there

15:37

is no overpopulation on this planet.

15:40

There are certainly places that

15:42

have seen too many people

15:44

to live, but there are plenty

15:46

of places where there's nobody.

15:49

So there's, so not only is

15:51

there not an overpopulation,

15:53

there's actually going to be

15:55

quite the reverse. in the next

15:58

50 years and beyond. where there

16:00

is simply going to be

16:02

under population, a population

16:04

collapse. So that should be

16:06

removed. That's not, should not

16:08

be anybody's reason. We're not having

16:11

kids. And maybe even a reason,

16:13

if you care about the long-term

16:15

human species, because for every

16:17

woman who doesn't have any

16:19

children, another woman has four,

16:22

to just keep the current level,

16:24

or whatever level it is,

16:26

replacement level. But beyond that, I

16:28

think, I am trying to recall,

16:31

but I don't think if ever

16:33

anybody who had regretted having

16:35

more children. I would say modern

16:38

times, or in kind of the

16:40

developer. Certainly, there were people

16:42

who didn't have birth control, and

16:45

that might not be true, but I

16:47

would say in the modern developer,

16:49

I have not many people there.

16:51

And that's also been my own

16:53

experience. And so... I

16:56

would say that there

16:58

is a compounding

17:01

joy from it. So

17:03

the people who

17:06

benefit from it

17:08

is the kids

17:10

themselves, where we

17:13

did have her son

17:15

keep requesting that

17:18

my wife lay

17:20

another brother. They

17:23

lay him or brother. in

17:25

my group with five. And

17:27

so we know the joy of that. And that

17:29

was, I think, that to me is kind

17:31

of like one of the most

17:33

important ones. How do you think

17:36

we address the population collapse?

17:38

Because it feels like it's already

17:40

priced in. You can't just start

17:42

having 18-year-olds. Is it that robots

17:45

and AI just kind of stalled

17:47

for some of the labor stuff? Or

17:49

how do you think that that

17:51

plays out? I don't know how

17:53

it plays out. It's really perplexing

17:56

to me and I

17:58

do know from all

18:00

these solutions in terms of

18:02

having human replacement

18:05

to raise birth rates has not

18:07

worked, including just directly

18:09

paying people. Now, it may be

18:12

that they haven't been offered enough

18:14

sums that maybe if you do

18:17

calculations and you figure

18:19

that every baby born is

18:21

worth a million dollars to

18:23

the economy over its lifetime,

18:25

that it's worth to pay. a

18:27

million dollars, so maybe at

18:29

the right price, it would begin

18:32

to work. So far, I don't

18:34

know what that is, and if

18:36

there isn't a way to do

18:39

that, the question is, well, what

18:41

happens to the economy? And

18:43

there, you could imagine AIs

18:46

becoming, in some ways, the

18:48

audience in the market, in the

18:50

way that, you know, we build

18:53

houses for cars, we

18:55

call garages. Maybe we,

18:57

you know, make entertainment for

18:59

the AIs in some ways. Maybe we

19:01

have to build houses for robots,

19:03

pools, I don't know. So,

19:05

that's possible. That's one possible

19:07

way out. But it is something that

19:10

I don't have very many

19:12

good ideas about. Let's go

19:14

out more of the AI

19:16

human relationship because I've heard

19:18

you, you're famously an optimist.

19:20

And you're excited about where

19:22

things will take us. And

19:24

I'm curious. because some people

19:26

will say, hey, they're optimists, but

19:28

they also believe that, you know,

19:30

humans one day will be subservient

19:33

or to the AI in the

19:35

same way that, I don't know,

19:37

Neanderthals were kind of the previous

19:39

form of evolution. We're not the

19:41

last form of evolution. We're

19:43

not the last form of evolution.

19:46

We're not the last form of

19:48

evolution. We're not the last form

19:50

of evolution. And there will be, we're

19:53

not the last form of evolution.

19:55

I'm curious where you net out

19:57

of that. Well, yeah, I mean, maybe.

19:59

in between. So I first

20:01

I would concede or even

20:04

postulate that we've been our humanity

20:07

is malleable through something we've

20:09

invented and we're not done

20:11

yet. So we have been inventing

20:14

ourselves and we will

20:16

continue to change ourselves. And

20:19

so we will we want to become

20:21

better humans, but we don't

20:23

know what that is. And the

20:26

question that we also don't know

20:28

is is there multiple futures

20:30

for us? Okay, can we become

20:33

different kinds of humans? And

20:35

that prospect is really

20:37

very problematic in many ways.

20:39

We certainly could imagine there would

20:42

be varieties of us that are

20:44

not going to get anything

20:46

changed as naturals. They're

20:48

simply never going to doing

20:50

genetic engineering and not

20:52

permit very much alteration

20:55

in their bodies. And that's like

20:57

the Amish, right? And then we

20:59

can easily imagine people who are

21:01

going to be very quick to,

21:03

they want to eradicate the gene

21:05

for Alzheimer's and Parkinson's, from

21:08

their germline and all future

21:10

descendants tomorrow. You know, then we

21:12

could see a forking, at least one

21:14

fork, maybe more, going on and I

21:16

don't know what to think about that.

21:18

That's, is that, you know, is that

21:20

surrendering? Is it kind of like the

21:23

union of the United States? You can't

21:25

succeed. You can't succeed? There can

21:27

be no divisions, we must

21:29

remain one. So I can easily imagine

21:32

a bunch of people believing that,

21:34

that the worst fate would be to have

21:37

speciation among humans. And

21:39

others who would say, no, this is actually

21:41

natural, until recently there

21:43

were always other varieties of

21:46

humanoids on this planet, until

21:48

probably we got rid of

21:50

them, but then we don't know. And

21:52

so that was, so maybe there's kind

21:54

of like we're just returning to

21:56

the time. when we had multiple

21:59

sentient beings. which is still

22:01

a little different. I mean,

22:03

I call these AIs artificial

22:05

aliens and that's fine,

22:07

but that's not the same thing

22:09

as kind of allowing a forking

22:11

in your own species. The

22:14

danger there is, you know,

22:16

the most effective argument we

22:18

have against racism is

22:20

that there is no difference

22:22

between us. But what if it

22:24

really was? And that's like, what

22:27

do we do with that? And so,

22:29

there might be a very good reason

22:31

to not permit speciation. And there's

22:33

a few ways speciation

22:35

can happen, right? There's the genetic

22:38

engineering example, there's the, you know,

22:40

putting chips in people's brains, example,

22:42

there's the, you know, maybe the

22:44

chat GPT, you know, 50 or

22:46

whatever, just gets, gets, gets so

22:49

advanced that, you know, which speciation

22:51

versions do you feel are most

22:53

realistic or do you think all

22:55

of them will happen to something

22:57

greater? The cyborgian thing is going

23:00

to be very, very slow. You know,

23:02

I just recently went to see

23:04

the neural link stuff, where they have

23:06

the implant, but it's a science recorder,

23:09

but in the back of your head, and

23:11

it allows a monkey to control

23:13

the computer, but if they're

23:15

mine, and they're coming close

23:17

to actually doing human testing

23:20

on this for quadriplegics to

23:22

help them walk. And that was a

23:24

lot closer than I thought. So that's

23:26

how it starts. It starts with

23:28

a kind of a medical

23:31

therapeutic way to help

23:33

people. But that's, I'm not

23:35

sure, I mean, that will speed

23:37

up the evolution in certain

23:39

directions, but that's a very

23:42

very slow process. Whereas the

23:44

genetic germline thing

23:46

can happen, you know, at

23:48

the rate of human generations.

23:50

So I would, so I would

23:53

say in terms of like literally

23:55

speciating. and having different

23:57

varieties of humans. I

23:59

think the genetic side will be

24:01

the really the the moment the

24:04

one to pay attention to

24:06

and it can be informed

24:08

in terms of what we

24:10

discover from AI and other things in

24:12

terms of how to do it but

24:14

you know that's still a

24:16

slow process you know 25 years of

24:19

life for a human generation

24:21

to turn over it'll take a long

24:24

time before they're like you

24:26

can't breed kind of

24:28

speciation too. So that's kind of a

24:30

more technical biological thing is that

24:33

you can't breed. And I think

24:35

that's, I don't know if I, you

24:37

know, whatever be, consideration, we probably

24:39

can genetically engineer some way

24:41

to debris these hybrids. So,

24:43

speciation means only that more of

24:46

an identity, I guess. The LES, or

24:48

your calcium argument, as far as I

24:50

understand it, is something along the lines

24:52

of the idea that it's not that

24:54

AI needs to be conscious in order

24:56

to. get rid of us, it's

24:58

just, it needs to be a

25:01

different to us, which it is,

25:03

and then just say, or just

25:05

understand that our atoms are, you

25:08

know, something that they could use

25:10

to, or anything else, for it

25:12

to advance its own goals,

25:14

and then at some point

25:16

it will become in its

25:19

own interest to eliminate us.

25:21

Maybe I'm botching his

25:23

argument. How would you

25:25

comment on that? of the many

25:28

arguments out there

25:30

about existentialism, the

25:32

only one that sort of

25:34

makes sense to me as a

25:37

thing to worry about is that

25:39

basically that we over

25:41

time allow an engineer

25:43

AI to do more for us. And

25:45

at some point we give over

25:47

more and more of we voluntarily

25:51

engineer more and more

25:53

control to the A to the AIs.

25:55

on purpose to

25:58

run things. that

26:00

at that point of kind of giving

26:02

in power than we're at their mercy

26:04

should it awaken to to some kind

26:07

of sense of survival being

26:09

threatened. That's the closest I

26:11

can understand it. I

26:13

think it's a fantasy in many

26:16

many dimensions and one of them

26:18

is that's not going to happen

26:20

fast so the one fantasy is

26:22

that happens fast we can't back

26:25

out of it. Two, the fantasy is

26:27

that there's a single AI or

26:30

incredible collaboration among

26:32

several big AAs and there's

26:34

no evidence at all. The evidence

26:37

is going to be that there's

26:39

going to be millions, hundreds

26:42

of different AIs. So it's like

26:44

we have different engines and

26:46

tools and machines. We don't

26:48

have one big machine. And

26:51

then the third one, I think the

26:53

most serious fantasy part

26:55

is. this complete overestimation

26:58

of the role of intelligence

27:01

in achieving things, accomplishing.

27:03

There's this idea that

27:06

whatever is smartest will

27:08

dominate. But we know if you

27:10

put a human in a tiger in

27:12

a cage, we know which one's going

27:15

to live. It's not the

27:17

smartest one. There is just so

27:19

many other things that are

27:21

necessary to accomplish things in

27:24

the real world. Being the

27:26

smartest person in the room is

27:28

not necessarily mean that you are

27:30

the dominant person No matter how

27:32

smart you are you have other things to

27:35

you need other things that you've done

27:37

including Access to things and in

27:39

cooperation and collaboration many people

27:42

doing many different things or

27:44

many AIs or many things

27:46

and that middle-aged guys who

27:48

like to think a lot Really really

27:50

put a high emphasis on thinking

27:52

and they think that if they

27:54

could think faster and better

27:56

than they would be running

27:59

the world. And they can think of

28:01

all the ways in which they could do

28:03

usually figure out how they could

28:06

take over, but it's a fantasy

28:08

because that's not how the world

28:10

works. That's how reality works.

28:12

You have all these things and

28:14

they break down and they don't work

28:17

on the first attempt. And by

28:19

the way, the instinct for

28:21

survival usually will dominate

28:23

the, you know, the attempt to kill.

28:26

Survival is a more powerful motivator.

28:28

than trying to get rid of

28:30

somebody. I think it's a

28:33

fantasy kind of like Superman.

28:35

It's beautiful. It's miffic.

28:38

It makes it makes comic book

28:40

sense. But why isn't the

28:42

Neanderthals to humans kind

28:45

of like evolution? Why isn't

28:47

that a good analogy for what

28:49

could happen or will happen?

28:51

Well, well, yes. Okay. I think

28:53

it isn't analogy. And so

28:56

that took, I don't know. 10,000

28:58

years. Yes, if you would say that

29:00

over time that they could replace us

29:02

over time, I could buy that. But

29:04

to say that, you know, all at once

29:07

overnight, overnight, that's this comic

29:09

book, fantasy. Right. And we

29:11

do see, you know, in

29:13

time, there jumps in, you

29:15

know, sort of economic growth,

29:17

for example. So things aren't always

29:19

as linear, but even still, you

29:21

know, 10,000 years is a long

29:23

time. to expect this to happen

29:25

in 20 years. And the analogy

29:27

with Neanderthals, by the way, is

29:29

that much more likely had been

29:32

bred away rather than murdered. The

29:34

point is that there was a willing

29:36

merging going on. It wasn't that they were

29:38

killing us off. There was, and so in

29:40

that sense, it's like, yeah, we may merge with

29:42

them, but only if we decided it was

29:44

in our, you know, in our interest

29:47

to do so. And that's one

29:49

of the scenarios for Neandereros joined

29:51

us, rather than we murdered than we

29:53

murdered them. There's so, and it took

29:55

a long time. And if that's the

29:58

scenario that we're talking about. Okay.

30:00

We'll continue our interview in a moment

30:02

after a word from our sponsors. In an age of

30:04

AI and big data, your personal information has

30:06

become one of the most valuable

30:08

commodities. Data brokers are the silent

30:11

players in this new economy, building

30:13

detailed profiles of your digital life

30:15

and selling them to the highest

30:17

bidder. But data breaches happening frequently,

30:19

these profiles aren't just fueling spam.

30:21

They're enabling identity theft and putting

30:23

your financial security at risk. But

30:25

what if there's a way you

30:27

could opt out of that system

30:29

entirely. Incognite is a service that

30:31

ensures your personal data remains private

30:33

and secure. They handle everything from

30:35

sending removal requests to managing any

30:37

pushback from data brokers. What I

30:39

trust most about Incognite is how

30:41

it provides ongoing protection. Once you're

30:43

set up, it continuously monitors and

30:45

ensures your information stays off the

30:47

market. Want to extend that peace

30:49

of mind to your loved ones?

30:52

Incognomy's family and friends plan lets

30:54

you protect up to four additional

30:56

people under one subscription. Take control

30:58

of your digital privacy today. Use

31:00

code upstream at the link below

31:02

and get 60% of an annual

31:04

plan. Again, that's incognite.com/upstream. Hey everyone,

31:06

Eric here. In this environment,

31:08

founders need to become profitable

31:10

faster and do more with

31:13

smaller teams, especially when it

31:15

comes to engineering. That's why

31:17

Sean Lanahan started squad. A

31:19

specialized global talent firm for

31:21

top engineers that will seamlessly

31:24

integrate with your org. Squad

31:26

offers rigorously vetted top 1% talent that

31:29

will actually work hard for you every

31:31

day. Their engineers work in your time

31:33

zone, follow your processes, and use your

31:35

tools. Squad has front-end engineers excelling in

31:37

typescript and react, and NextGS, ready to

31:40

onboard to your team today. For back-end,

31:42

Squad engineers are experts at NojS, Python,

31:44

Python, Java, and a range of other

31:46

languages and frameworks. While it may cost

31:49

more than the freelancer on upwork billing

31:51

you for 40 hours, but working only

31:53

two, squad offers premium quality at a

31:55

fraction of the typical cost, without the

31:57

headache of assessing for skills and culture fit.

32:00

Squad takes care of sourcing, legal

32:02

compliance, and local HR for global

32:04

talent. Increase your velocity without amping

32:07

up burn. Visit choose Squad.com and

32:09

mention turpentine to skip the wait

32:11

list. You wrote a book a long time ago

32:13

called What Technology Wants? And you've written

32:16

a lot about the topic since then.

32:18

And obviously you've been thinking about technology

32:20

for, you know, many decades, both past

32:23

and future. Do you feel it's kind

32:25

of a precursor in some ways to...

32:27

market recent software is eating the

32:29

world thesis or how would you

32:32

kind of contextualize your your mental

32:34

model of of technology and and

32:36

where what it wants where it's going? I

32:38

mean does it completely take up do

32:40

we do we become kind of like

32:42

completely technical techno planet is that what

32:45

you mean by a takeover? We become

32:47

the tools our tools shape us literally?

32:49

Well yes so so in a certain

32:51

sense we are becoming as I influenced

32:53

we've invented our humanity we

32:56

invented ourselves we are in

32:58

invention and we will continue

33:01

to do that. But there are

33:03

many many attributes that we

33:05

have of living in these

33:07

kind of wet cell

33:09

biological self-reproducing

33:11

things that, you know, it's as

33:14

we're making these robots, we've

33:16

come to understand that the

33:18

kind of power density of

33:20

a human, we're quarter horsepower

33:23

in, you know, 60 watt brain. It's

33:26

going to be a long time before any

33:28

kind of a robot will be able to

33:30

operate at those low energy levels.

33:32

So yes, so we may continue diversals,

33:34

but we may not necessarily

33:37

diverge very much from this

33:39

miraculous machine that we have. I

33:41

think we're likely to populate

33:43

our surroundings in the environment

33:46

with all these artificial alien

33:48

beings of all different varieties

33:50

and stripes. But beyond that in

33:52

terms of the eye and just the

33:54

general drift of the techium which is

33:56

kind of to make it's basically just

33:59

to make organ and forms and

34:01

kinds of being that were possible

34:03

but not possible to be made

34:05

with living tissue. In the long trend,

34:07

you know, there's all these forms

34:10

that could be, that could be doing

34:12

things, but you're not gonna get there

34:14

if they have to make it in

34:16

our cells that are mostly made out

34:18

of water. You can get to those if

34:21

you can make them out of other

34:23

elements. And you can only get to

34:25

there through a mine. So we have. First

34:27

of human mind, then we're going to

34:29

make all these other AI minds that

34:32

will be inventing new ways to

34:34

make forms, new ways to make a

34:36

living, new ways to exist, to be,

34:38

that are technological, and

34:40

we can fill up that. I

34:42

don't think it's going to necessarily

34:45

replace biological forms,

34:47

because that's generally not

34:49

what we see. We see evolution

34:51

much more additive. There are things

34:53

that go extinct. However, in

34:55

technology we don't see extinction.

34:58

That's one of the differences,

35:00

because our idea base, and we can

35:02

carry ideas forward. So we have

35:04

shifted the evolutionary arc, because

35:06

now we don't have to have

35:09

as much extinction. And so we

35:11

can imagine going forward where we

35:13

retain as many of the

35:15

biological species as possible, while

35:18

adding on additional technological

35:20

species. And we make more and more.

35:22

So we have a world, like a planet

35:24

full of all the biological species that we

35:27

have today, and they continue, and

35:29

millions of more technological species.

35:31

And so that's what I think the

35:33

general pattern is. Yes, is the world filled

35:35

with all kinds of technologies that we don't

35:38

have today, but not at the cost

35:40

of the biological species. And

35:42

to make this a bit more concrete,

35:44

you said a few times, we invented

35:46

our community. unpack exactly what that means

35:48

in terms of like what were the

35:51

biggest inventions or just make that

35:53

a bit more concrete for us. One of

35:55

the our biggest inventions is the

35:57

invention of language, which we did

35:59

invent. I mean, it was

36:01

meaning that I think

36:03

that primitive, you know,

36:05

primates trying to communicate,

36:08

try things in their

36:10

mind, try to do

36:12

something, and they worked.

36:14

Those who were able

36:16

to do that survived

36:18

and try it again

36:20

to make something work,

36:22

and that kind of

36:25

that battery of things,

36:27

techniques that they discovered.

36:29

We're passed on and became the

36:32

basis of our language in that

36:34

language. What language really gives us,

36:36

there's two things. One is this

36:38

communication between members, which is very

36:41

powerful, and how we can hunt

36:43

better, so we can client could

36:45

survive. But there's something else really

36:47

important about language, which is that

36:50

it gave us access to our

36:52

own thinking. So the only way

36:54

we know what we're thinking is

36:56

through language. So the language was

36:59

a tool. two-prong thing that gave

37:01

us access to try and figure

37:03

out what we are. So it

37:05

was our origin or our consciousness,

37:08

basically. That was very, very powerful

37:10

because it kind of then could

37:12

give us to be purposes and

37:14

directions and intention that we didn't

37:17

have before. So that was something

37:19

we invented. And then from them

37:21

we invented things like domestication of

37:23

hurting animals, which we could milk,

37:26

and once we figured out that,

37:28

the human body. rapidly evolved in

37:30

certain populations, adult tolerance of lactose.

37:32

That happened within, I don't know,

37:35

7,000 years or something. It was

37:37

really fast. We invented cooking and

37:39

fire, which was an external stomach

37:41

that could digest stuff we could

37:44

not with our primate stomachs. We

37:46

could then access nutrition that we

37:48

couldn't get to, which changed our

37:50

teeth and jaws very, very, very

37:53

quickly. So we... So how we

37:55

look right now. is something that

37:57

we invented. And so we're continuing

37:59

to do that right now. The

38:02

actually biological evolution has not slowed

38:04

down with cultural evolution. It's actually

38:06

sped up. So that's what I

38:08

meant by inventing. And then all

38:11

the things that we think are

38:13

important to us, like fairness and

38:15

the moral judgment and all these

38:17

other things, we invented. Idea of

38:20

law and our idea of even

38:22

though there's some primitive traits. We

38:24

can see like fairness among primates.

38:26

We have invented elevated and much

38:29

more stricter levels of that. And

38:31

some of us communicated through culture,

38:33

which is something else we invented.

38:35

But when we think of a

38:38

elevated enlightened human, that's something that

38:40

we've invented. In one of our

38:42

first conversations almost a decade ago,

38:44

I remember you said something like,

38:47

we will need a new mythology

38:49

or new mythologies that helped us

38:51

kind of make sense of the

38:53

new world that we're entering. Yeah,

38:56

I mean, at the Low Now

38:58

Foundation, we're building this clock to

39:00

tick inside the mountain for 10,000

39:03

years, and that's, we hope that

39:05

to be a mythic thing, this

39:07

idea of the cock taking the

39:09

mountain for eons, for generations. That's

39:12

the kind of mythologies that I

39:14

think is helpful for us as

39:16

we. kind of reimagine ourselves. Yeah,

39:18

and our new purpose in this

39:21

world. It's interesting because, well, just

39:23

on the point of purpose, I

39:25

mean, some people say, hey, you

39:27

know, people will do art and

39:30

poetry and will find all these

39:32

new things once a lot of

39:34

knowledge work has been automated, but

39:36

the AI will be better at

39:39

that kind of stuff too. How

39:41

do you think about in a

39:43

world a decade from now or

39:45

two decades from now how humans

39:48

think about? purpose in a different

39:50

way once a lot of things

39:52

that they were the only to

39:54

use your words of you know

39:57

don't be the best be the

39:59

only could perhaps be be done

40:01

with a or feel free to

40:03

just be the premise. You know,

40:06

I just think we know so

40:08

little about what our own intelligence,

40:10

your own being is like. And

40:12

so AI is going to be

40:15

disruptive and instructive for decades because

40:17

it's going to help us experiment

40:19

on us in some ways. I

40:21

think we're going to learn more

40:24

about our own brains from AI

40:26

in making a than we have

40:28

from neuroscience or psychology together so

40:30

far. I think we're going to

40:33

be having this discussion about what

40:35

is we are about, where are

40:37

we going, what is it for

40:39

the next 50 years at least,

40:42

as all these things come along

40:44

and we kind of re-register them.

40:46

I think one of the things

40:48

that we're, one of the things

40:51

we've done just this year is

40:53

demoted creativity. Again, this has been

40:55

talked about as like, well, the

40:57

thing that humans do is creativity.

41:00

Computers are the opposite of that.

41:02

Now we know. wrong. Computers can

41:04

do creativity at a lower, minor

41:06

level very easily. And so now

41:09

we're saying, well, yeah, creativity is

41:11

not this high order supernatural thing.

41:13

It's actually very, very primitive. And

41:15

so we've kind of changed our

41:18

ideas about creativity very, very fast.

41:20

And I think that kind of

41:22

thing we're going to continue to

41:24

do, not always necessarily demoting and

41:27

just shifting and maybe having more

41:29

subtle new ones, because I think

41:31

there's what I call capital or

41:33

major creativity in the minor. So

41:36

we'll have to devise new language

41:38

for distinguishing between what this kind

41:40

of everyday novelty is versus a

41:42

kind of a breakthrough where you

41:45

are, we are trying to do

41:47

something that's outside the average. And

41:49

so we probably maybe have two

41:51

new words instead of just the

41:54

one word creativity. I'm fascinated by

41:56

the, you know, we'll learn more

41:58

about ourselves than we have with

42:00

neuroscience. and psychology, we still have

42:03

no idea what consciousness is, and

42:05

maybe we'll figure it out in

42:07

the process. One idea you've been

42:09

thinking a lot about for a

42:12

while is this idea of a

42:14

global government, or the need for

42:16

one, and also this idea of

42:19

co-balance, I think you call it,

42:21

there's the idea of kind of

42:23

two-way transparency. Well, yeah, it's not

42:25

surveillance, but co-violence. Yeah. And, you

42:28

know, it's interesting because when some

42:30

people talk about... existential risk, I

42:32

mean, one of the accidental risks

42:34

we've introduced in the last century

42:37

has been nuclear weapons, but the

42:39

competitive dynamic in terms of multiple

42:41

countries having nuclear weapons seems to

42:43

have staved that off. And so

42:46

I wonder if similar to, you

42:48

know, right now we're having conversation

42:50

about AI centralization and decentralization, and

42:52

if there should also be a

42:55

competitive dynamic, that might stave that

42:57

off. And one worry people could

42:59

have with the global government is,

43:01

you know, does that reduce power

43:04

a bit too much? and thus,

43:06

you know, why would they, you

43:08

know, be covalent, so to speak,

43:10

like, why would they allow, you

43:13

know, two-way transparency? How do you

43:15

think about sort of the, do

43:17

you see a need for decentralization

43:19

there, or do you agree that

43:22

competitive dynamic is what's preventing, you

43:24

know, perhaps this abusive power? Yeah,

43:26

I would say in general, what

43:28

we know about systems, again, me

43:31

taking the whole systems approach is

43:33

that there's a tremendous power in

43:35

the in the bottom. in a

43:37

very flat-ish, bottom-up, decentralized, distributed system,

43:40

which is a large part of

43:42

evolution, large part of ecology, a

43:44

large part of living systems, large

43:46

part of the mind. But it's

43:49

not the only part. And that's

43:51

the lesson is that most of

43:53

the systems we see are combinations

43:55

of... lots of decentralization in many

43:58

aspects and some centralization. What we

44:00

wait, we know that the... The

44:02

advantages of decentralized distribute systems

44:04

is agility, flexibility,

44:07

adaptability, supreme. They're

44:10

just the best ways to adapt

44:12

to changing environments,

44:14

changing circumstances, changing goals.

44:17

But we also know that

44:19

there's a cost to that. They're

44:22

incredibly inefficient. I

44:24

mean, just by the nature, you're

44:26

duplicating. thing. There's no

44:29

sense of efficiency whatsoever.

44:31

There's no mechanism for

44:33

efficiency. As soon as you

44:36

introduce efficiency, you begin

44:38

to centralized. So the question

44:40

always is there's tradeoffs

44:43

and all these systems being

44:45

hybrid is you're going to pay

44:48

the costs for certain aspects

44:50

of the decentralized

44:52

system, pay the cost

44:54

of inefficiencies and slowness

44:56

and other stuff. adaptability

44:58

or flexibility that it does

45:00

is so valuable that you're

45:02

willing to pay the costs and other

45:04

other aspects of it. It's not

45:07

worth paying the cost. You want to

45:09

have something more centralized. So

45:11

authentication could be

45:13

decentralized and there's reasons

45:15

to do that and there'll be some

45:17

cases where it's going to be

45:20

worth to pay the cost

45:22

of completely decentralizing it. But in

45:24

many other cases, it'll make sense

45:26

to have a more centralized version

45:28

of it, so there's no free lunch

45:31

in that way. The way I would say is,

45:33

you know, the bottom up is always

45:35

to take you further than you can

45:37

go, these centralized systems.

45:39

Take further you can go than

45:41

you ever thought, and they're usually

45:44

the best way to start, but

45:46

they don't take you all the way. And

45:48

so most of the systems that

45:51

are kind of highly boiled

45:53

and well working will be

45:55

some combination of... mostly decentralized

45:57

with little bits of

45:59

top-down centralized control. This has

46:01

been a fascinating conversation. I want

46:03

to close one more of your

46:05

book, which is you have this

46:08

one aphorism, which talks about instead

46:10

of thinking about can do or

46:12

can't do, think about I do

46:14

in terms of internalizing something identity.

46:16

This kind of meta concept, which

46:18

I'll say is a plug for

46:20

the book, which is there's this

46:22

basketball player, Giannes Antis Acublo, who

46:24

just lost in the first round

46:26

of the NBA playoffs. And before

46:28

the series started, if you were

46:30

to ask him, hey, you know,

46:32

if you guys lose this series,

46:34

is it a failure? He would

46:36

have said, yes, it's a failure.

46:39

We have to win at all

46:41

costs and got to have our,

46:43

you know, head of the game.

46:45

But then afterwards, when someone asked

46:47

him, you know, is this season

46:49

a failure, because they lost, he

46:51

said, no, it's a learning experience.

46:53

And he just talks about the

46:55

idea of like, certain proverbs might

46:57

serve you better in certain times.

46:59

And maybe life is about, you

47:01

know, you know, adopting the right

47:03

mindset. I think you agree. I'm

47:05

not sure if it's the same

47:07

guy, but there was somebody else

47:10

recently. He said, there's no failures

47:12

in sport. It's the same guy,

47:14

same guy. Yeah, I mean, that's

47:16

just very, very briefly, but that's

47:18

one of the base. One of

47:20

the newest, occasionally there are new

47:22

things under the sun, and there

47:24

are one new thing under the

47:26

sun that I think Silicon Valley

47:28

can take credit for is demoralizing

47:30

failure. To understand the failure is

47:32

seen now. as a necessary component

47:34

for science or innovation, for entrepreneur,

47:36

for the economy in general, and

47:38

that what you want to really

47:41

have are systems that manage failure.

47:43

Okay, the failure management systems where

47:45

you have your failures in small

47:47

doses and you manage them to

47:49

prevent the cataclysmic failures that you

47:51

want to avoid. And so I

47:53

think there's been a change, complete

47:55

sea change, where if you lost

47:57

money or if you had bad

47:59

grades or if a experiment failed

48:01

that was considered. a disgrace and

48:03

now is seen as you say

48:05

as a learning experience as a

48:07

way we go forward. So the

48:09

fail forward idea. So yes, fail

48:12

forward. That's a great place to

48:14

end. The book is excellent advice

48:16

for living. Wisdom, I wish I

48:18

had known earlier. Kevin, thanks so

48:20

much for writing this book and

48:22

for coming on the podcast. I

48:24

really appreciate your great questions, Eric.

48:26

Thank you. Upstream with Eric Tormburg

48:28

is a show from Turpentine. The

48:30

podcast network behind Moment of Zen

48:32

and Cognitive Revolution. Do you like

48:34

the episode, please leave a review

48:36

in the Apple Store. Hey listeners,

48:38

it's Eric. It's overwhelmingly clear that

48:40

health care needs fixing, but who's

48:43

actually doing it? The podcast's second

48:45

opinion brings you the problem solvers

48:47

making real progress across health tech,

48:49

biotech, AI, pharma, funding, and actual

48:51

medicine. Our host are subject matter

48:53

experts. Christina Farr is a leading

48:55

health tech journalist and investor. Dr.

48:57

Ashe Zanuz is a physician and

48:59

former tech CEO, and Luba Greenwood

49:01

is a biotech executive and lawyer.

49:03

Every week, they cut through the

49:05

hype to get at what actually

49:07

matters in health care's future. Get

49:09

the inside story on the second

49:11

opinion podcast. Search for it in

49:14

Apple Podcasts, Spotify, or your favorite

49:16

podcast app.