HomeVideos

Why The Laws of Startup Physics Have Changed | Ben Horowitz Interview

Now Playing

Why The Laws of Startup Physics Have Changed | Ben Horowitz Interview

Transcript

1611 segments

0:00

When Ben Horowitiz and his partner Mark

0:01

Andre came into the venture capital

0:03

industry, it was very different than it

0:05

is today. You can argue that it is them

0:07

more than almost anyone else that has

0:09

reshaped this industry and matured it so

0:11

much ever since. Andre Horowitz has

0:13

become one of the most important

0:14

institutions, not just investors, but

0:16

institutions in the private investing

0:18

landscape, having achieved a scale that

0:20

no one thought was possible in venture,

0:22

which was always supposed to be this

0:23

small, tiny niche corner of the world.

0:25

This conversation is a bit unique

0:27

relative to some of the other ones that

0:28

Ben has had more recently. I tried to

0:30

understand the shaping forces and

0:32

influences in his life and the ways that

0:34

he thinks America most needs to change.

0:36

He's taken this as his life's mission to

0:38

build a firm that affects outcomes in

0:40

the country, not just in a small niche

0:42

part of the market, but very broadly,

0:44

even having discussions in this

0:45

conversation about his work with, say,

0:47

the Las Vegas Police Department, which

0:48

he's tried to infuse with technology to

0:50

lower crime rates across the system. I

0:52

hope you enjoyed this great and

0:53

wide-ranging conversation with Ben

0:54

Horowitz.

1:03

I think a fun place to begin, Ben, would

1:05

be your take on the state of the

1:07

country. Like what does it feel like to

1:09

you in 2026? I know part of your mission

1:12

is to like directly impact the

1:13

trajectory of the country. We'll talk

1:14

about that a lot. Yeah.

1:15

>> Uh but but begin with what does the

1:18

landscape, the playing field like to you

1:20

today? I think the tech sector is very

1:22

very healthy. America's competitiveness

1:24

is very very good. The entrepreneurship

1:26

culture is outstanding. Uh which you

1:30

know and that's the main thing I look at

1:31

from my lens. If you look at and I go

1:34

kind of all over the world and everybody

1:36

wants Silicon Valley like how can we

1:38

have Silicon Valley in the UK? How can

1:40

we have it in France? the thing that's I

1:42

would say lacking so they have a lot of

1:45

the ingredients right they have um great

1:49

uh talent um they've got great

1:52

universities they have like a definitely

1:55

a worse regulatory environment um in EU

1:58

increasingly bad regulatory environment

2:00

for entrepreneurship but there there's a

2:03

cultural challenge where

2:06

you know succeeding doing something

2:08

larger than yourself making the world a

2:10

better place like those aren't things

2:11

that

2:13

uh young people feel like society

2:16

values. And so the likelihood if you're

2:20

building a company of getting people to

2:23

kind of work for you and dedicate their

2:25

life to a mission like that is just not

2:27

that great. Whereas in the US, it's

2:29

amazing. I think the economy is in much

2:31

better shape than

2:34

>> people realize uh and start to see that.

2:37

Well, we're getting, you know, we've

2:40

done a lot of kind of things to

2:42

stimulate it. You know, we've got lower

2:43

energy prices. We've got much less

2:45

regulation. We've got kind of a more

2:47

user-friendly tax code. Um, and that's

2:51

all starting to kick in now. And then,

2:54

you know, from our perspective, I think

2:55

the bigger thing is AI is, um,

3:00

you know, it's going to impact

3:01

everything. There's almost no problem

3:02

you can think of that you can't go,

3:05

well, we have a real shot at solving

3:06

that with AI. It's from, you know, what

3:09

were the big problems in the US? Auto

3:10

deaths. Well, we got an AI solution for

3:12

that. Cancer, we have an AI solution for

3:14

that. So, you know, the fact that we've

3:17

got a technology where we can address

3:20

everything uh is a real new phenomenon

3:23

>> and and all that's I think going to kick

3:26

in like in a fairly major way over the

3:29

next 12 to 24 months. Why do you think

3:31

12 24 months is is a time frame worth

3:34

mentioning that some of this stuff will

3:35

start to be felt more broadly?

3:37

>> It's all kind of starting to take effect

3:40

now and you know it's got to roll out

3:43

get deployed. Now you know deployments

3:44

of technology in particular in the past

3:46

have taken a long time. Um but you know

3:49

you had to build out the infrastructure

3:51

to do it. So like for cars you needed

3:52

things like roads and traffic lights and

3:54

all that kind of thing. And for the

3:57

internet, you needed, you know, fiber in

3:59

the ground and, you know, people to have

4:01

smartphones and you needed to do a lot

4:03

just to get going. Um, the internet is

4:06

here. So, if you want to use AI, if you

4:08

want to apply it to your business, you

4:09

just do it. Like there is no

4:10

infrastructure that needs to be built to

4:12

adopt the thing.

4:14

>> What could most interrupt this good

4:17

trajectory that America is on where we

4:19

are building solutions using technology?

4:21

Like what are the biggest risks?

4:22

>> I think policy. One of the things my

4:24

father said to me was a bad government,

4:27

no matter how many smart people you

4:28

have, no matter how great a culture you

4:30

have, no matter how great the country

4:32

is, can ruin the whole thing. Venezuela

4:35

was the fourth richest country in the

4:36

world. Crazy, you know, and

4:39

>> then like, you know, communism and and

4:41

that's that. And

4:44

you know if you look at

4:47

how little comes out of so many of these

4:49

countries in Europe that have so many

4:51

smart people and then you know and then

4:52

the ones that went into communism and

4:54

there's so many like genius Romanian

4:57

entrepreneurs John vonman and the number

5:00

of great genius scientists that came out

5:03

of Hungary like this little country and

5:05

then like it was just gone once the

5:07

communists took over is like completely

5:11

like nothing from from inventing

5:13

everything to nothing overnight. And I

5:17

think that that can absolutely happen

5:19

here. We could outlaw AI. Like I think

5:22

there there were like pretty aggressive

5:24

proposals. The last Biden administration

5:27

executive order said that you could not

5:30

sell a GPU without federal government

5:34

approval. Like that was a real executive

5:37

order and it got reversed. But like we

5:38

were that close to being basically out

5:42

of the uh global chip game. So it's it

5:45

is fragile. By the way, like technology

5:48

solutions

5:50

work much better than policy solutions.

5:52

That's the that's the other thing. Like

5:53

policy solutions is very hard to make

5:55

anything work. Uh so if you think about

5:59

um you know co we could tell everybody

6:02

to stay in their house. Well that's got

6:04

some like extremely bad side effects.

6:06

you know, turned out not to work that

6:07

well or like, you know, we could invent

6:11

a drug that cures it or like a vaccine

6:14

that works. It's just hard to have or a

6:16

policy solution like, you know, all the

6:17

policy stuff on climate change, you

6:20

know, and Europe actually, you know,

6:22

reduced emissions and all that, but it

6:24

didn't do anything because like China

6:25

didn't reduce emissions. But if you

6:27

build a technology a really safe nuclear

6:30

efficient or nuclear fusion facility

6:33

then like that that would have a big

6:35

effect and I think in general that's

6:37

true that uh you know and police like

6:41

defund the police did not make anybody

6:43

safer technology does and so you know if

6:47

you really want to change the world if

6:48

you really want to make it a better

6:49

place I think you can build a solution

6:51

for darn near anything. If you want to

6:53

change the world for the better, it's

6:56

never been a better time to be an

6:57

entrepreneur.

6:58

>> I was with a uh local restaurant tour

7:00

yesterday here in New York, one of the

7:02

best, for a couple hours having him

7:04

describe to us how he is planning on

7:06

using AI tooling to improve everything

7:08

about his restaurant business. How do

7:10

you think about the way all of this is

7:12

changing the sort of potentially large

7:15

attractive businesses that you want to

7:16

invest in? Because there's been stick

7:19

with the restaurant example. Toast is a

7:21

great there's many great companies that

7:22

have been built in and around restaurant

7:24

software businesses. It seems like this

7:27

restaurant owner is going to be able to

7:28

have his own spun up operating system

7:30

specific to him not going to need any of

7:32

that stuff. How how is this changing the

7:34

way in which you view investment

7:36

opportunities

7:37

>> on the positive? Uh one everything is up

7:39

for grabs, right? I I think people are

7:41

kind of over uh reacting to that in the

7:43

stock market and so forth and that if

7:45

you look at um existing software

7:47

companies like people think, oh, they're

7:49

all dead. Well, some of these guys are

7:52

extremely hard targets. Like it's not

7:55

that easy to take out Salesforce or SAP.

7:58

You you you would be surprised um even

8:00

with AI like how much uh heavy lifting

8:04

that is. Having said that, it is true

8:06

that you know a lot of these things,

8:09

yeah, you can just make your own, you

8:10

can do it yourself. Uh it's going to be

8:12

a lot easier. That's a just like the

8:15

number of possible interesting companies

8:18

I think went up a lot. I think the other

8:21

thing we're seeing is the products work

8:22

so much better than any technology

8:25

products we've seen in the past that

8:26

revenue growth is so much faster for

8:28

these AI companies and there's many such

8:31

cases of companies coming out you know

8:34

cursor which is ostensibly an IDE like

8:36

what's the biggest ID before cursor like

8:39

I don't know but it wasn't big uh and it

8:41

took probably 12 or 15 years to get to

8:44

that revenue level and you know they

8:46

went over a billion dollars in revenue

8:48

like in no time. So that's super

8:51

interesting. I would say though from an

8:53

investing standpoint,

8:56

the laws of physics of company building

8:59

changed which is going to in affect

9:02

investing in what's currently I would

9:05

say an unknown way. So if you look at

9:09

the one thing you knew if you'd ever

9:12

built a software company is you cannot

9:14

throw money at the problem.

9:15

>> Yeah. Yeah.

9:16

>> Like you know what's a man year? you

9:19

know, 700 IBMmers before lunch. Like

9:21

that uh you know, that phenomenon kind

9:25

of everything was built on because you

9:27

knew if somebody built a great product

9:29

and it took them three years and they

9:30

did it with a small team, Google's not

9:32

going to hire 2,000 engineers and catch

9:34

them. It's just not going to happen.

9:36

That was a law of physics. Now,

9:39

if you have uh the data and you have

9:42

enough GPUs, you can solve damn near

9:45

anything. and kind of we've seen that

9:48

with uh Elon catching the big models in

9:52

no time. I mean, he just took a lot of

9:55

money and a really good data center

9:57

design and some smart engineers. He's in

10:00

the game, you know, like he got in the

10:02

game very fast.

10:03

>> That would have never happened in the

10:05

past. The markets are also seem to be

10:08

much much much bigger than anything

10:09

we've ever seen. So it would cause you

10:12

to think about valuations and kind of

10:15

long-term value and other sorts of

10:17

things in a different way than we have

10:19

in the past. On the one hand, it's like,

10:21

well, when you calculate the long-term

10:23

value, what if this market wasn't, you

10:25

know, $50 billion? What if it was $5

10:28

trillion? And then on the other end,

10:32

well, what if somebody could catch you?

10:34

These are just concepts we've not dealt

10:36

with. So how would the conversations

10:38

feel different to me if I came in you've

10:40

got all these great investors working at

10:41

Andre and Horovitz the the the nature of

10:44

the conversation amongst your teammates

10:46

as they're debating this sort of stuff

10:47

versus four years ago or something where

10:50

where does it feel most materially

10:51

different internally

10:52

>> I would say one of the most different

10:54

things is when you look at AI

10:55

researchers it is really

11:00

a different kind of thing if you haven't

11:03

been at like Google or Facebook or open

11:05

AI or anthropic and like somebody gave

11:08

you hundreds of millions of dollars to

11:09

try and build a giant model and you

11:11

weren't like one of the main people,

11:12

then you probably don't know how to do

11:14

it because you can't learn it in school.

11:17

Um, and you can't learn it in school

11:20

because it's it's a little bit

11:22

alchemistic in nature. You know, you are

11:25

it's it's a little bit of an art. And so

11:28

if you've never done it before, the

11:29

chance of on your very first try of

11:31

building some kind of large model that

11:33

it's going to work well isn't that

11:36

great. Now that's,

11:38

you know, people are coming up to speed

11:39

more. There's more companies. Um, people

11:42

are learning it. But that's kind of why

11:44

you got to this, which from the outside

11:47

world probably looked absolutely bananas

11:50

that, well, why is somebody paying a

11:52

hundred million dollars for an AI

11:54

researcher or a billion dollars for an

11:56

AI researcher? Like, that's the craziest

11:58

thing I've ever heard.

12:00

Well, what if there were only 40 of them

12:02

like in the world?

12:03

>> And you have a$4 trillion dollar

12:04

company. Yeah, then it kind of changes

12:07

the math on it a little bit. And I think

12:09

that's sort of where we were because

12:11

it's kind of the first time we've had a

12:14

need for a technologist that academia

12:17

could produce. That um is kind of

12:22

probably one of the bigger things that

12:24

changed in the conversation is like who

12:25

are all these people? Like we track all

12:27

of them and you know know what they're

12:29

doing but uh it's very different.

12:30

>> Your finance team isn't losing money on

12:32

big mistakes. It's leaking through a

12:34

thousand tiny decisions nobody's

12:36

watching. Ramp puts guard rails on

12:38

spending before it happens. Real-time

12:39

limits, automatic rules, zero

12:41

firefighting. Try it at ramp.com/invest.

12:44

Every investment firm is unique and

12:46

generic AI doesn't understand your

12:48

process. Rogo does. It's an AI platform

12:50

built specifically for Wall Street,

12:51

connected to your data, understanding

12:53

your process, and producing real

12:54

outputs. Check them out at

12:56

rogo.ai/invest.

12:57

The best AI and software companies from

12:59

OpenAI to Cursor to Perplexity use work

13:02

OS to become enterprise ready overnight,

13:04

not in months. Visit works.com to skip

13:06

the unglamorous infrastructure work and

13:08

focus on your product. Everyone talks in

13:10

venture about the power law. The thing

13:12

underneath the power law is a sort of

13:14

inequality. It seems like so many of the

13:17

things that are happening are just

13:19

massive multipliers on the trend of

13:21

inequality in every way. the billion

13:23

dollar researcher, the size of the

13:24

biggest companies, the wealth of the

13:27

people creating those companies. I would

13:29

argue that inequality is a feature, not

13:30

a bug of the American system. But yeah,

13:32

>> I'm curious for you to riff on like the

13:35

nature of growing inequality and the the

13:38

good and the bad associated with that.

13:40

what's happening in AI is sort of, you

13:43

know, I I would just say an extension of

13:44

the Kobe Bryant effect, which is um, you

13:48

know, a basketball player in uh,

13:51

whenever James J. Nay Smith invented the

13:54

game, you know, like there was a limited

13:56

amount of money you could make because

13:58

you basically played the game in front

13:59

of the people who could show up for the

14:02

game and that was it. That's the whole

14:03

market. Um whereas once you add

14:06

television and the global audience and

14:08

these kinds of things, you can get to

14:10

you know a much bigger you can be LeBron

14:12

James, you can become a billionaire and

14:14

that that just was not at all possible

14:17

before. And I think that uh you know we

14:20

kind of first saw that with the internet

14:23

where okay now I can build a product and

14:26

I can get to global distribution very

14:27

fast then all of a sudden I can become

14:29

like extremely rich and then AI is

14:33

another layer on top of that and that

14:36

okay now take that same product and make

14:39

it just more a valuable thing and so

14:41

whoever invents that is whatever the

14:44

internet company was plus+ and so that's

14:47

going to make them even richer. That's

14:49

the kind of bad part of it. I think the

14:51

good part of it is it starting out day

14:53

one like completely democratized like

14:56

the AI

14:58

anybody gets access to like very

15:01

powerful AI uh you know anybody who has

15:03

a phone and now everybody

15:06

most people in the world uh at this

15:08

point have smartphones and now you've

15:10

got like super intelligence uh in your

15:13

phone. So that's a big it's an equalizer

15:17

of the opportunity in a lot of ways that

15:20

I don't think we've ever seen a bigger

15:23

opportunity equalizer than AI. Um in

15:26

that every child can have like a super

15:29

advanced amazing tutor teacher. Um so

15:33

like education

15:36

great education is accessible to all

15:38

now. So I think it's an equalizing

15:40

technology and uh you know there's some

15:44

drive in inequality and this is another

15:46

thing I I learned from my father. He

15:48

said look son life isn't fair and that's

15:51

extremely good advice because it's just

15:53

not going to be fair. Like no matter

15:55

what government or anything tries to do

15:58

it's not going to be fair. And the

16:00

problem is if you create a system that

16:02

tries to correct that it doesn't make

16:03

things more fair. It just transfers all

16:06

the power to the person running the

16:08

system. And that's what happened with

16:09

Stalin. That's what happened with

16:10

Chescu. That's what happened with Pulp

16:12

Pot. That's what happened with Mao. You

16:14

know, not an accident that every single

16:17

system like that went bad because it

16:19

really ends up just being a power

16:21

transfer. When you think about, well,

16:23

what do you want? You'd like everybody

16:26

to have a chance, you know, like don't

16:29

give me no chance. Give me some chance.

16:31

Now, it may not be as big a chance as

16:33

the other guy. It may not be, you know,

16:35

like a perfect chance. Um, but if I have

16:38

the desire, if I've got some capability,

16:40

give me a chance to be something like to

16:42

make my imprint on the world. And

16:46

a system like that is going to end up

16:49

with a lot of inequality. All, by the

16:52

way, all systems end up with a lot of

16:53

inequality.

16:55

But you can try systematically to give

16:59

everybody an opportunity. And I think AI

17:01

does a really good job of that. One of

17:03

the memes that's very popular today is

17:04

that you have a couple years to get some

17:06

capital or you're going to be a part of

17:07

the permanent underclass is like the is

17:09

the phrase that is used on Twitter. And

17:11

I certainly agree that um now everyone

17:13

has the best lawyer, accountant, you

17:15

know, adviser in their pocket and that's

17:17

amazing. Um but what do you think about

17:19

this notion that we're just going

17:21

because of that we need less labor. We

17:23

need it's going to be harder if you

17:25

don't have some capital to begin with to

17:27

accumulate capital and break in. I don't

17:29

I don't necessarily believe that. I'm

17:30

just curious what you think about

17:31

challenges we'll face because of AI

17:33

society.

17:35

>> Yeah, I don't really think that's right.

17:38

I think that I don't think like the the

17:40

door is going to close behind you. I

17:42

think like the opportunities tend to

17:43

multiply

17:44

>> um when you kind of open up a new door

17:47

and open up like a new way of doing

17:49

things. We saw that with crypto. So many

17:51

people who made money on crypto were

17:53

like people who,

17:55

you know, literally didn't have much to

17:58

start with. they just got into the

17:59

technology early um and then they kind

18:01

of parlayed it up. And so if you have

18:03

something that grows really fast, that's

18:06

actually the opportunity for somebody

18:07

with a little bit of capital to make a

18:08

lot of money because it doesn't take

18:10

much. You know, if you bought Bitcoin

18:12

for a nickel, you did really well. Uh

18:16

and all you needed was a nickel. And

18:18

that's uh you know, I think that's the

18:21

nature of these things that go

18:23

hyperbolic. And you know particularly if

18:25

you create something I also think uh

18:28

the the labor market stuff

18:32

I think people are acting as though it's

18:34

very predictable and when it's not at

18:36

all predictable. So

18:40

if you look at kind of the history of

18:42

the of the world um and automation and

18:44

this is what it is. It's a kind of like

18:46

an automation technology. We've been

18:49

automating things since the agricultural

18:51

days. And in in those days, I think 95

18:56

or 96% of all jobs in the US were

18:59

agriculture. Almost all those jobs have

19:01

been eliminated. Um and the jobs we have

19:04

now the people doing agriculture

19:06

wouldn't even consider jobs. And so like

19:08

the idea that we could imagine all the

19:11

jobs that are going to come, you know,

19:13

sitting here, you know, that AI is going

19:15

to enable, I think is low. I think the

19:18

need for like more creativity jobs um is

19:22

going to go way up and the kind of need

19:25

for kind of jobs to process work for the

19:30

creatives will probably uh go down in

19:32

some ways but um I'm not even sure about

19:35

that. You know, we've had AI going right

19:38

imageet was what 2012 and then natural

19:42

language stuff and Burton and all that

19:44

was like 2015 and then you know chatbt

19:46

was 2022 and like where's where's all

19:50

the job destruction?

19:52

You know why hasn't it happened yet? And

19:55

why are you so [ __ ] sure it's going

19:56

to happen next? And why are you so sure

19:58

no jobs are going to be created? I don't

20:00

think it's nearly as predictable as

20:02

people are are saying. How would you

20:04

describe the nature and scope of your

20:06

ambition over the next 10 20 years?

20:10

>> One of the things that I learned um so I

20:13

had a mentor who's a a great great CEO

20:16

by the name of Andy Grove. Um and he was

20:19

uh the CEO of Intel and he kind of

20:22

famously did the the major pivot of them

20:24

out of the memory business into the

20:26

microprocessor business. Maybe the

20:27

greatest tech CEO we've had. Uh and one

20:30

of the things that he he said that you

20:34

know in a way is very obvious but I

20:35

think is um also profound is if you're

20:41

the leader in the industry then the

20:43

growth of the industry is dependent on

20:45

you. Um like you it's up to you to

20:48

expand the market like nobody else is

20:49

going to do it. Uh and so when I think

20:53

about the firm I think of it a lot in

20:54

those terms. The reason America is

20:56

America and and there's many narratives

20:58

on this, but like I think the factual

21:01

one is like we won the industrial

21:04

revolution. We really did. We had Henry

21:07

Ford and we had Thomas Edison. We had

21:08

like great entrepreneurs. They built

21:10

great technology. The technology lead to

21:13

a military lead, led to an economic

21:15

lead, led to cultural dominance. None of

21:17

that was by accident. And had we not had

21:22

all those inventions, had all those

21:23

companies, um, which led to, you know,

21:26

everything from like winning World War

21:27

II, we just won't be, we'd be some other

21:29

thing. We won't be America. So, we're

21:32

there again. Like, this is the

21:35

equivalent change of the industrial

21:37

revolution in terms of how everything

21:39

works, governments, societies,

21:40

businesses. And you know, we're either

21:44

going to be uh the leader of that

21:47

technology, the provider of that

21:49

technology, or we're not. And if we're

21:52

not, we're not going to be um the

21:56

economic superpower, the military

21:58

superpower, the cultural influence, the

22:00

kind of standard of the world that we

22:02

are now. At least I think that would be

22:06

bad. I think you know uh America's been

22:08

kind of good for the world and good for

22:10

giving people a chance like we talked

22:12

about before and so our role in that you

22:14

know you know taking it try try and be

22:17

humble with the role but our role is

22:19

like from a policy standpoint from a

22:21

funding standpoint from a helping people

22:23

build standpoint to make sure that that

22:26

next set of great companies comes out of

22:29

uh America or allied nations a core

22:31

ambition is to do our part in kind of

22:33

helping that

22:34

>> I want to ask about some of the

22:35

ingredients to do that. Well, but just

22:36

as a quick sidebar on Andy Grove, uh his

22:40

book is incredible. Like everyone should

22:41

read High Output Management. Um what was

22:44

it about what what very specifically did

22:46

you learn from him? Like what did you

22:47

see him do that impacted the way that

22:49

you think or behave?

22:51

>> Well, like I'm so overly influenced by

22:53

him, it's hard to even pin it down. But

22:55

so high output management, you know, I I

22:57

actually wrote the new forward um for

22:59

it. Uh which I actually think that's the

23:02

best thing I ever wrote was a forward

23:04

high output management. Um, but the hard

23:06

thing about the reason I wrote that

23:08

forward was um I, you know, it was my

23:11

favorite book and I wrote uh the hard

23:14

thing about hard things was basically

23:16

intended to be um the updated version of

23:19

it. But the the thing in high output

23:22

management that um he did so well that I

23:27

I tried to you know kind of do my own

23:30

version of is

23:33

you know the the the concepts of

23:35

management are easy like

23:41

they you need an eighth grade education

23:43

maybe to kind of understand management.

23:46

It's not like physics. It's it's pretty

23:48

simple. Uh but the psychological part of

23:52

it is extremely difficult particularly

23:54

for a young person to be able to do. Uh

23:58

you know it's it's super

24:00

confrontational. You're having to kind

24:03

of look through the conversation you're

24:05

having to the entire organization. You

24:08

really have to be confusion at times.

24:09

The the good of the of the whole

24:12

supersedes the good of the individual.

24:14

uh and all these things are are really

24:17

complicated to do. His big influence on

24:19

me was me trying to not only absorb that

24:22

but then kind of tell it in a more

24:25

up-to-date kind of modern way. I went to

24:28

visit him. He had this award on the wall

24:30

which was it was literally like um

24:35

manager of the year from for the Santa

24:39

Clara facility of Intel and it was from

24:43

I don't know 1992. I'm like Andy we're

24:46

like the biggest CEO in the world like

24:48

why did they give you the manager of the

24:49

year award for the Santa Clara facility?

24:52

And he goes, "Oh man." He's like, "You

24:54

know, Santa Clara was like the always

24:57

scored like it was the lowest quality

24:59

scores, the lowest [ __ ] score on

25:02

everything at Intel." And so I was just

25:04

like, "I'm going over there and talk to

25:06

them." So I go over there

25:09

and he said, "I brought a roll of toilet

25:12

paper and I put it under my desk, under

25:15

my chair." And you know, I said like,

25:18

"When are you going to get this facility

25:21

up to code?"

25:23

And they just started in with all this

25:27

[ __ ] [ __ ] [ __ ] [ __ ]

25:30

[ __ ] And I [ __ ] reached under my

25:33

chair and put all the toilet paper up. I

25:35

said, "Clean up your [ __ ]

25:39

and tell me when the [ __ ] you're going

25:40

to be up to code." And in two months,

25:42

they were up to code. and they were

25:43

always the highest rated facility

25:45

thereafter you know just on that. Uh so

25:47

they gave them manager of the year for

25:49

that.

25:49

>> When did you first experience the

25:51

lessons that drove his success this

25:54

confrontational psychologically

25:56

difficult aspect of management yourself?

25:58

How would you encourage other people to

26:00

like get a get a taste of it? You can't

26:02

just read about it. Obviously

26:04

>> what happens uh to founders is you

26:07

invent something right now. I've got to

26:09

build a company. you don't know what

26:11

you're doing and you make mistakes and

26:14

then those mistakes really cost the

26:16

company and you lose confidence and that

26:20

leads you to hesitate and that

26:24

hesitation is what kind of causes the

26:27

failure mode. So then either like the

26:30

company's indecisive or they get very

26:32

open. All these guys got so open to

26:34

input from their team and their

26:36

executives and like but you know the

26:39

team doesn't have the full context. Only

26:41

the leaders got the context. So even if

26:43

they're smarter than you, you still

26:46

likely can have better judgment because

26:47

you have all the knowledge. Um, but you

26:51

know, they defer and then if you defer

26:53

to people who work for you, then that

26:56

kind of creates a weird political

26:58

situation because people jump into the

27:00

vacuum of like, you're not making the

27:02

decision, I'll make the decision. And

27:03

then that feels political to everybody

27:05

else. And so that's the pattern people

27:08

run into. And so, you know, you really

27:11

kind of have to build up enough

27:14

confidence in them to have that

27:17

confrontation. The hardest version of

27:19

this, by the way, is the reorg. Uh

27:21

because reorg is basically you're

27:23

redistributing power to make the company

27:25

work better, to like have communication

27:27

be better, to not have as much conflict.

27:29

But what's going to happen is somebody

27:31

who's really good, who you've had for a

27:32

long time, is going to lose power and

27:35

they're going to be [ __ ] pissed. Um

27:38

and so then if you compromise the

27:39

organization so they can maintain their

27:42

power then you've just kind of

27:45

redistributed power from the people

27:47

doing all the work to the executives and

27:49

that's a catastrophe. So it's always

27:52

that kind of thing where people don't

27:55

want to have that conver confrontation.

27:59

They don't want to tell that person look

28:01

the organization's here. you you helped

28:04

us tell here, but like you either have

28:06

to be happy in this new role or it's

28:08

going to be a rap. When you're young and

28:10

inexperienced, you know, it's going to

28:11

hurt to like tell him that, but I don't

28:16

know it's going to help me to do this

28:17

reorg because I don't I'm not

28:19

experienced enough to know that. I've

28:21

never done that before. And so I'm going

28:24

to go with the known avoid hurt to the

28:26

to the theoretical avoid hurt. Um, and

28:29

that's when you wreck your company. And

28:30

and and so that's the pattern. And I,

28:32

you know, I always do my best to like

28:34

lend them my experience on that.

28:36

>> You you were lucky that when you started

28:38

Andre and Horwitz, you and Mark had both

28:39

had tons of operating experience both

28:41

together.

28:42

>> Yeah. I still didn't know what I was

28:43

doing as CEO.

28:44

>> Fair enough.

28:45

>> And he didn't know what he was doing

28:46

either. Like his ideas now, like if you

28:49

ask Mark about management now, like he's

28:51

so different than how he actually did

28:52

it. Um, and it actually makes him mad if

28:55

you talk about it too much because he's

28:56

like, "I got such bad [ __ ] advice.

28:59

They told me to hire all these guys."

29:00

How do you think he's most different?

29:01

Like what would he say is or what do you

29:03

observe to him to be the most different?

29:06

>> I just think he's like way more um

29:10

in control of his own. Like Mark is

29:13

super emotional person. Um and he's just

29:17

way more in control of it than he was

29:19

then. Uh just in terms of just like the

29:21

personality. He used to be like zero or

29:23

100, right? Like so he would be like

29:25

full of emotion like what the [ __ ] are

29:27

we doing? or like I'm just not gonna say

29:29

anything like but nothing in between.

29:32

>> Something I know the least about about

29:34

your firm is like the first I don't know

29:35

what period of time three days, three

29:38

months, three years.

29:40

>> And I'd love to hear about how you

29:42

thought about the business right as it

29:44

was getting started. Of course, we're

29:46

I'm going to come back to what it is now

29:47

and and those ingredients you mentioned

29:49

for having the impact you want to have.

29:51

But uh lots of this is an incredible

29:54

part of the world. Silicon Valley, Wall

29:56

Street, you know, these are institutions

29:57

that make America great. Lots of people

29:59

listening have ambitions to do this sort

30:01

of thing. And I'd love to hear like the

30:04

very very ear early primordial case

30:06

study. Yeah.

30:07

>> Of what it was like and what kinds of

30:08

conversations you were having and what

30:10

your initial ideas were.

30:12

>> So venture capital, first of all, you

30:14

kind of have to understand the the

30:16

context of it was um there hadn't really

30:21

been new top tier venture capital firms.

30:24

So like the the last one before we

30:28

started that you would say is top tier

30:30

was probably Benchmark which ostensibly

30:33

started in 1995 but it didn't really

30:35

because all those guys came from another

30:37

firm called Merryill Pickard

30:39

>> and that firm was like from the 80s and

30:44

there there hadn't really been a new one

30:46

from the 80s and if you looked at why

30:48

every VC was kind of reputationbased and

30:51

so to be top tier you had to have

30:53

invested in Apple and Cisco and Google

30:56

and you know Yahoo and all the great

30:58

companies and you can't from a standing

31:00

start get to that and then if you're not

31:02

top tier in VC you're not going to last

31:05

because yeah in a super hot period

31:08

everybody makes money but the best

31:10

entrepreneurs will only work with the

31:11

top tier firms because that's how you're

31:14

going to recruit great engineers that's

31:16

how you're going to get follow-on money

31:18

like everything comes out of that so

31:20

you'd never take money from a tier two

31:22

if you could get it from tier one and so

31:24

That's why the tier ones always have

31:26

better returns. Um, so we knew we had to

31:28

be tier one, but we had that problem.

31:31

And the idea that we had was, well,

31:35

venture capital is a great product for

31:38

LPs, um, but it's not a great product

31:41

for entrepreneurs. And so, if we could

31:42

build a better product for

31:44

entrepreneurs, then we could win. And

31:47

that was like the the original kind of

31:50

framework. And the idea that we had for

31:52

the product for entrepreneurs was you

31:54

know because we had been entrepreneurs

31:55

was around what you and I had been

31:57

talking about which is well

32:00

if you're like a founder who wants to

32:04

run their own company you're not getting

32:06

much like you need so much you don't

32:09

have the confidence you don't have the

32:10

knowledge you don't have the knowhow you

32:11

don't have the network. Um what if we

32:13

built a firm that like was designed to

32:16

give you enough confidence, power,

32:20

network reach, advice that you could

32:24

actually be a CEO. And so that was the

32:26

whole idea behind the firm originally.

32:29

And then the second idea we had, which

32:31

was the other thing, like VCs didn't

32:34

ever market themselves at all because if

32:37

you're all based on your investing track

32:40

record, it's best that it's just magic.

32:42

Like why say anything? Like keep that a

32:44

secret. And so they weren't talking. And

32:47

so when we went out and talked, like

32:50

everybody covered it. So we instantly

32:52

everybody knew we had this product.

32:54

>> Where did that where did the germ of

32:56

that specific idea come from? like let's

32:58

be fairly loud relative to what others

33:00

do from the very beginning.

33:02

>> Well, it's funny because you know Mark

33:03

and I were talking about it. He said to

33:05

me, he's like why don't VCs market and

33:09

actually it the original thing went all

33:12

the way back to kind of the first uh

33:15

class of VCs which were the investor

33:18

revolution VCs were JP Morgan,

33:21

Rothschild, Goldman Sachs etc. right

33:23

like they were the ones financing these

33:25

things. Um, and it turned out that

33:30

these guys were financing both sides of

33:33

World War II. Um, and so they really

33:37

didn't want any publicity because that

33:38

would have been like an extremely

33:40

[ __ ] bad uh thing. To a large extent

33:43

that just carried over all the way

33:45

through Arthur Rock and

33:47

>> and all these things and then, you know,

33:48

the reputation thing clicked in and it

33:50

was working, so there was no need to do

33:52

it. Um, and we got a lot of criticism

33:54

when we did it. our LPs would say, you

33:57

know, like the other VCs say, you guys

33:58

are egoomaniacs. You name the firm after

34:00

yourself. You're marketing it like this.

34:02

And it was so funny because the reason

34:04

we named the firm after ourselves is

34:06

when we try we raised money in 2009,

34:08

which is right on the, you know, edge of

34:11

the financial crisis. And the big

34:13

objection from LPS was, well, like you

34:16

guys are like really good entrepreneurs.

34:19

You're just going to leave this thing

34:20

and go build another company and then

34:22

we're going to be stuck with the fund.

34:24

And we couldn't get them off of that.

34:26

And so then I had the idea. I was like,

34:28

"Well, why don't we just name it with

34:30

our names and then they know we're

34:31

safe."

34:32

>> Yeah.

34:32

>> And and that worked.

34:34

>> As your business grows, Vanta scales

34:36

with you, automating compliance and

34:37

giving you a single source of truth for

34:39

security and risk. Learn more at

34:41

vanta.com/invest.

34:43

Ridgeline is redefining asset management

34:45

technology as a true partner, not just a

34:47

software vendor. They've helped firms 5x

34:49

and scale, enabling faster growth,

34:51

smarter operations, and a competitive

34:52

edge. visit ridgelineapps.com

34:55

to see what they can unlock for your

34:56

firm.

34:57

>> Um, if you think about the the period of

35:00

takeoff of the firm in 2009 up until you

35:03

reach, let's call it like cruising

35:05

altitude, like when was cruising

35:07

altitude and and and what was the most

35:09

difficult part about getting it from

35:11

takeoff to that point?

35:12

>> Well, I mean, the first thing is we

35:13

really didn't know that much about

35:14

investing. Mark and I had done some

35:16

angel investing, but we not neither of

35:18

us had any venture capital experience.

35:20

And like you know, credit to Sequoia,

35:24

credit to, you know, Greylock and and

35:27

and Kleiner and all the guys who were

35:29

around at that time, you know, they just

35:31

had years and years of of doing it. We

35:34

made more than our fair share of

35:36

investing,

35:37

mistakes, you know, missing things we

35:40

should have done and and then, you know,

35:43

doing things that we shouldn't have

35:44

done. Um, but missing things that we

35:47

should have done was probably the bigger

35:48

one. And

35:50

you know and then the other thing is is

35:52

that kind of how we thought about the

35:54

profile of the investor was wrong. So we

35:57

so overindexed on our idea that we had

36:02

to help the founder uh become a CEO that

36:05

we made it a requirement that you

36:06

couldn't be an investor at Andre and

36:09

Horowitz if you hadn't like founded and

36:12

or run a company. And that, you know,

36:17

was a very good attitude and and set the

36:19

culture of the firm in a lot of ways and

36:21

had good uh things that came from it.

36:23

But I would just say that like most CEOs

36:29

aren't as interested in investing as

36:31

they think they are. Uh and then also

36:34

most CEOs aren't as good at helping

36:37

somebody else learn the job. Uh and so

36:39

those two things ended up being, you

36:42

know, not quite correct. Uh, so we made

36:45

some adjustments, you know, fund one

36:46

just went really well because we, you

36:48

know, we hit the scene hard. It was a

36:50

small fund. We did Skype, we did Slack,

36:52

we did Octa. I mean, like, there was

36:55

Stripe was in there. Like, so like there

36:58

were just too many good things in a $300

36:59

million fund for that thing not to blow

37:01

the doors off. Fund two wasn't as good

37:04

as one. Um, and then by the time we got

37:07

to three, that's when we had uh the

37:12

contention among like, oh, we really

37:15

don't have the right profile for GP

37:18

here. Uh, and there was a while where we

37:20

thought that was going to be a terrible

37:21

fund. It ended up being a great fund

37:22

because we had, you know, Coinbase and

37:24

Data Bricks and Lyft and GitHub, but

37:27

that that one was scary for a while, but

37:29

coming out of that, we kind of knew

37:32

what the firm needed to be. Uh, and so I

37:37

think it was settled down after that,

37:38

you know, it wasn't such a like startup.

37:40

It was like, okay, we got across that

37:43

chasm. But then the bigger thing was we

37:47

always had this idea about software is

37:49

eating the world. um and you know which

37:51

is Mark articulated really well in his

37:53

uh 2011 piece. And so we always felt

37:57

like venture capital firms

38:02

needed to be able to scale uh and that

38:05

the other firms would have trouble

38:06

scaling because of uh the way they

38:09

worked, the way they shared control. Um

38:11

so that could be an opportunity for us.

38:13

Um but we hadn't figured out how to do

38:14

it yet. And then I'd say um starting

38:19

with like the bio and the crypto fund I

38:22

started to get to the kind of

38:23

organizational picture of how we would

38:25

become um be able to address every

38:28

market of technology. Uh and but with

38:33

investing teams that weren't 20 people

38:36

like that doesn't work. So you need an

38:38

investing team of like four or five

38:39

people but you you have to you can't

38:42

address the whole technology market with

38:44

five people. So you have to have

38:45

multiple teams. Having multiple teams in

38:49

a venture capital firm, it was a little

38:51

bit of a novel idea particularly when

38:54

each team has like a platform that helps

38:57

the founder build the company. And so we

39:00

began it really in earnest with the

39:02

crypto fund I think like around 2018 and

39:05

then uh and you know now the whole firm

39:07

is kind of organized that way. If you

39:09

think if we zoom now to today and back

39:11

to what you said which is that the scope

39:13

of your ambition is big as the leader be

39:15

the one to help be the ones that are

39:17

expanding the market. What are the

39:19

components of doing that? Like what is

39:20

the system need that it doesn't

39:22

currently have that you might be able to

39:24

provide? One is um the the capital

39:29

markets are have changed you know

39:33

dramatically with

39:36

not much um

39:39

help. So I went public at 18 months old

39:42

with $2 million in trailing revenue.

39:44

That wasn't a good idea. But companies

39:46

used to go public routinely with $50

39:48

million in revenue. It was fine. Um you

39:51

know now nobody's going public

39:54

>> a billion, right? like you you get to go

39:56

public or something like that and if and

39:58

you're kind of small if you don't have

40:00

that and so you kind of need a lot more

40:03

out of the private markets than VCs are

40:06

built to do and so you know that's one

40:08

of the kind of things we have to think

40:10

about. Another one is the companies in

40:13

the portfolio

40:15

you know they'd leave you at 100 million

40:17

in revenue. They're going public.

40:19

They're out to the races. Well that's

40:20

not true anymore. And so what do you

40:23

need when you get to be 200 million 300

40:26

million in revenue? Well, you need to be

40:28

multi-product, you need to be multi-

40:30

channelannel, you need to be multi-

40:32

geography. Um, so as a venture firm, you

40:36

know, we need to help them and as a

40:38

venture industry, we need to help them

40:39

do that. Like how do I get to Japan? How

40:42

do I uh get to South America? Like

40:46

most venture firms don't provide much

40:49

along those lines. So we kind of have to

40:51

step up to those ideas if we're going to

40:53

have companies in the portfolio at that

40:56

stage.

40:56

>> Do you hope that over time your firm and

40:58

maybe some others like it that have

40:59

become these big institutions and

41:01

venture go on to be sort of like the

41:03

Blackstone you know Apollo type

41:05

companies that are big publicly traded

41:07

you know enduring businesses.

41:09

>> A big huge wave in among venture

41:12

capitalists is uh private equity AI

41:17

rollups. It's a good business idea like

41:18

a really good business idea which is

41:20

okay

41:22

you know just like the spreadsheet kind

41:24

of created the original private equity

41:26

business AI is kind of creating a new

41:29

private equity business where you can

41:30

buy any existing company optimize it

41:33

with AI and it'll be more valuable.

41:37

That's a good idea. It's a good thing to

41:40

invest in. It's not something we're

41:42

going to do for two reasons. Um, one,

41:45

it's like the cultural opposite of who

41:48

we are. So, we're about building new

41:51

things, um, growth,

41:55

believing in the entrepreneur, price

41:57

doesn't even matter. As long as the

42:00

thing succeeds, you're going to do well.

42:02

Private equity is like entry price is

42:05

key. like the I mean know I I had a

42:09

great dinner with Mark Rowan who's a

42:10

super genius uh runs Apollo and he was

42:13

like entry price entry price entry price

42:15

you know we never even think about that

42:17

we think about it but it's not like

42:19

first and foremost at all like thinking

42:22

about containing cost and this and that

42:24

and the other that's just not like what

42:26

a good venture capital frame of mind is

42:28

so like culturally I didn't want to mix

42:31

those two things but more than that like

42:33

I just didn't want to be in a business

42:35

where the way you make money is you

42:37

figure out how to optimize an existing

42:40

thing and you know lay off people and

42:42

that kind of thing. We're about like new

42:44

technology companies building the future

42:46

um taking things forward and I'll leave

42:49

that to the other smart guys in the

42:52

industry.

42:52

>> What if any trade-offs ex feel like they

42:54

might exist at this scale as you

42:56

continue to scale as you consider all

42:58

these different people you're trying to

43:00

serve? Well, the investors internally,

43:01

the LPs, the founders, so many people

43:03

need to nothing's perfect. like what

43:05

what are the trade-offs to the path that

43:08

you've chosen?

43:09

>> I think you know with any scale of

43:11

organization you really have to over pay

43:14

attention to culture um or the culture

43:17

will drift. We probably spend um more

43:21

work on that than than any venture

43:23

capital firm. I'm like you're not

43:24

allowed to join unless you sign the

43:27

culture document. I I spend an hour with

43:29

every single employee teaching them the

43:31

culture. Like it's like that level of

43:33

investment. Um, and then you know we

43:37

really try to enforce it uh hard when we

43:40

can and you know we have pretty good

43:42

consistency but like that's that is hard

43:44

to maintain as you grow.

43:46

>> Can you teach me more about culture? The

43:48

the you've written a book about it.

43:49

You've you've built them. You've studied

43:52

some very interesting cultures that you

43:53

wrote about in the book. if you had to

43:54

teach a seminar or something on like

43:56

what a culture is in the first place and

43:58

then how to design one given what you do

44:02

and who you are and then how to you know

44:04

make sure it it people live by it.

44:06

>> Let me give you kind of like the

44:09

a small but like the probably the most

44:12

important insight which is from Bashidto

44:15

the way of the warrior from the samurai.

44:17

Uh a culture is not a set of ideas. It's

44:21

a set of actions. Um, and so if you

44:24

define your culture as a kind of set of

44:27

ideas, integrity, do the right thing, we

44:31

have each other's backs or any kind of

44:33

like these ideas, they call them

44:34

corporate values, it's actually just a

44:37

bunch of [ __ ] platitudes, it doesn't

44:38

mean anything. The culture has to be

44:40

defined in terms of the exact behavior

44:43

that you want that support that idea.

44:46

What do you have to do to actually be

44:49

that thing that you want to be? And so,

44:52

and it's the little things, you know,

44:56

how responsive are you to your

44:59

colleagues? What's the SLA on returning

45:02

a Slack message or an email? Do you show

45:04

up to meetings on time? Um,

45:07

and and this is like not everybody has

45:09

those ideas, but if you want that idea,

45:11

it's got you've got to manifest it

45:13

through something else. So, like we have

45:15

an idea about like you have to respect

45:17

the entrepreneur. Well, what is that

45:20

behavior? Like one, you can't ever be

45:23

[ __ ] late to a meeting with an

45:24

entrepreneur. I used to find people $10

45:26

a minute in the beginning of the firm to

45:28

reinforce it. And then uh you know, you

45:31

have to get back to an entrepreneur. If

45:34

you say no, like you have to say no. You

45:36

have to explain why you're not

45:37

investing. Um and you know, it has to be

45:41

clear. And we're going to survey that

45:42

entrepreneur after you um say no to make

45:45

sure that you said no and that they had

45:47

a good experience. So like that that's a

45:49

behavior. If you uh try to make yourself

45:53

look good by making an entrepreneur look

45:57

bad, you're fired. So like you get on X

45:59

and say, "Oh, he's selling dollars for

46:01

85 cents." No, no, no, no, no, no, no.

46:04

We're dream builders. We're not dream

46:06

killers. [ __ ] that. We're Somebody wants

46:08

to do something larger than themselves.

46:11

Build a company, you know, make the

46:13

world a better place. We're for that. we

46:14

don't give a [ __ ] what the idea is, you

46:17

know, and or if Sequoia funded them or

46:19

whatever. We'd love that. That that's

46:21

who we are. And so the behavior

46:25

is the culture is the actual thing and

46:29

that gets you the idea as opposed to the

46:32

idea and then figure out how you're

46:33

going to behave. And so that's probably

46:35

the main thing on culture.

46:36

>> Can you say more about the influence

46:37

your dad had on you? You mentioned that

46:39

lesson of nothing's fair or life isn't

46:40

fair.

46:41

>> Yeah.

46:42

>> Tell me about your dad. He was what's

46:44

known as a red diaper baby. Uh he uh my

46:47

grandparents were communists. Like they

46:50

went to secret meetings. They had cards.

46:53

My grandfather was fired during the uh

46:55

McCarthy era from being a a junior high

47:00

school teacher um you know for being a

47:03

communist. And he grew up a communist.

47:05

And he started out um on the left. He

47:08

was uh editor of a there's very famous

47:10

new left magazine called Ramparts

47:12

magazine which he was editor of and he

47:14

uh

47:16

you know was involved in the the Black

47:19

Panthers with Huey Newton and um you

47:21

know the Oakland chapter Eldrich Clever

47:24

and he sort of dropped out of politics

47:25

and he reemerged

47:27

um I guess probably eight years later on

47:30

the right. He really understood kind of

47:32

the ills of communism and socialism

47:34

which which helped me a lot. Like one of

47:36

the things that he said to me that uh

47:38

always stuck with me. He's like, "Son,

47:40

go to the library,

47:44

pick any book on socialism. There's

47:45

hundreds of books. And in that book, I

47:49

guarantee you, you will find page upon

47:52

page, chapter upon chapter of how to

47:55

divide the wealth. You will not find a

47:57

single sentence on how to create how to

47:59

make it." And I was like, "Oh, wow.

48:02

That's not like a very good system, is

48:04

it?" I learned a lot about systems

48:07

thinking from that uh which I you know

48:09

ended up being I'd say very helpful to

48:12

me as uh CEO. He wasn't like uh you know

48:18

this this this new age father he wasn't

48:19

like that you know in the old days your

48:21

father like they wouldn't even talk to

48:22

you till he got to be like 12 and uh you

48:26

know and then you get these little

48:27

snippets of wisdom and like one of the

48:28

ones I actually put in the hard thing

48:30

about hard things but I had uh you know

48:32

I had three kids I was young um and I

48:36

remember there's like 102 degrees the

48:38

air condition was broken the kids were

48:40

going crazy like one of them poured a

48:41

whole bottle of apple juice like a

48:43

gallon of apple juice into the rug

48:45

Apple juice is steaming out of the car.

48:47

But I'm just sitting there looking like

48:48

I was going to die. And my father looks

48:51

at me and he goes,

48:53

"Son,

48:56

you know what's cheap?" I said, "What?"

48:59

He goes, "Flowers.

49:01

Flowers are cheap." I said, "Okay." He

49:03

said, "You know what's expensive?" I

49:05

said, "No, what?" He said, "Divorce."

49:07

And, you know, he had been uh married

49:10

four times, so he knew what he was

49:12

talking about.

49:13

>> Yeah. As you look out today in the

49:14

world, I'm curious what things are

49:16

captivating you most and maybe even like

49:18

most inspiring you. You get you have

49:20

such an interesting perch. You get to

49:22

see

49:23

>> so much at the frontier.

49:26

>> What's going on in coding now is like

49:27

quite phenomenal. You know, like we kind

49:30

of went through this period where like,

49:31

okay, AI can write code, cool. Okay, you

49:34

can vibe code stuff with a lot of

49:35

security holes, fine. Um, but I think

49:39

over the break, over the kind of winter

49:42

break,

49:44

it turned a corner where like really

49:47

really good programmers were going,

49:49

"Whoa,

49:50

>> oh god,

49:50

>> this is this helps me."

49:52

>> Like I just became a hundred times more

49:55

productive and I can't remember any kind

49:57

of

49:58

>> technology where like just all of a

50:01

sudden you wake up and everything the

50:02

whole world just changed like that. And

50:05

that's happening on a

50:08

pretty regular basis I would say. And

50:10

then you know you know we spent a bunch

50:12

of time with uh people in Hollywood who

50:13

are using AI. I think AI will help you

50:16

make movies both better and at much

50:19

lower cost because you can do you know

50:22

you can shoot a scene and then have the

50:25

AI do a variation of that scene. That's

50:27

very very good. Um, and so you don't

50:30

have to do, you know, the really like if

50:33

you're an actress, you have to shoot a

50:35

scene like 15 or 20 times or something.

50:38

Uh, wouldn't it be nice to shoot it

50:39

three times and then you just like take

50:41

the pieces you like and make it what you

50:42

want? Uh, so it's I I think it's a

50:46

little underestimated as a tool for

50:49

creatives. I think um, and I think

50:51

that's true in music, too. You know, I

50:53

was uh kind of a young person when

50:56

hip-hop started and

50:58

the the huge criticism like this is not

51:00

music. They're just taking music and

51:02

they're like remixing it um and they're

51:05

rapping over it and it's a bunch of

51:06

[ __ ] like it's a novelty. But it was

51:08

postmodern art and I think we're going

51:11

to get into uh kind of postmodern art

51:14

with like what people will be able to do

51:16

with AI and music. And that was like one

51:18

of the most exciting

51:21

times in music. Like the invention of

51:23

the new art form is when it gets really

51:26

exciting.

51:27

>> What people in hiphop specific people

51:31

have had the largest impact on you

51:32

personally and and how? Nas is a very

51:35

good friend of mine and um he uh

51:41

he's definitely had a big impact. Just

51:43

the the lens at which

51:46

he sees the world is so

51:51

different um and interesting for me. So

51:55

we're both like very big fans of Rock

51:57

Kim who is kind of like the John Cold

52:00

Train of rap. So, Rak Kim had one of his

52:03

first big song was a song called My

52:04

Melody. And uh Nas and I are listening

52:08

to My Melody and the the first line is

52:12

turn up the bass, pull up a chair, hand

52:14

out a cigar, I'm letting knowledge be

52:16

born. I'm my name. And so he puts it on,

52:19

hands out a cigar. Uh and he pauses it

52:22

and he goes, "Ben,

52:24

why is he handing out a cigar?" And I

52:27

go, "I don't know why." Then he plays an

52:29

line. I'm letting knowledge be born. and

52:30

he's like, "It's a birth bin. He's

52:32

passing out cigars at the birth of

52:33

knowledge." And I was like, "Oh [ __ ] I

52:35

listen that song a thousand times. I

52:36

never heard that." Um, and

52:41

I can't tell you how many times like he

52:44

sees or hears something uh that's there

52:47

that I don't see. So having, you know,

52:50

somebody that I can talk to who has just

52:52

like a completely different perspective

52:54

of all things in life. Um, and it was

52:57

interesting, you know, uh, we did the

52:59

Coinbase deal together, uh, and he had

53:03

called me like two weeks prior to us

53:06

really kind of, um, seeing that, uh,

53:10

because he wanted to learn about

53:11

Bitcoin. So, you know, I explained to

53:13

him how it worked and, you know, he was

53:14

very interested. And then, uh,

53:17

you know, when I was talking to Chris

53:19

Dixon, who was working on the deal, I

53:20

was like, "Tell me about the guys." And

53:21

he's like, "Well, you know, one of them,

53:22

Fred, is like really into hip-hop." I

53:24

was like, "Okay." And so, you know, I

53:26

brought Nas over over to my it's like

53:28

have him come over to my house. There's

53:29

a boxing match on Saturday. You know, I

53:31

had Nas come over and like that's how we

53:34

got that deal. Um,

53:35

>> wow.

53:36

>> But, uh, yeah, he he's just like a I

53:38

would say a big influence on me

53:40

personally. And then he's such a you

53:42

know I um you know I as a whatever as a

53:47

leader and so forth like storytelling is

53:50

and a writer um is important to me and

53:52

he I I think he's one of the great

53:54

storytellers of all times you know like

53:56

just a super genius on that.

53:58

>> Is there a CEO comparable to Nas where

54:01

you know there's this class of guys in

54:03

the 90s where Jay-Z you know I'm not

54:05

just a businessman I'm a businessman.

54:07

Um, and there were these just massive

54:10

franchises that got born. These guys all

54:12

became incredibly successful in the

54:14

business world. And it it felt more um

54:16

like industrialized almost like the

54:18

whole process whereas not like even just

54:20

his album that just came out is it it

54:22

feels just like Ilmatic feels like it

54:25

could have come out then or now. It's

54:27

like this weird timeless quality. He

54:28

still has that somehow. And like

54:30

Premiere, same thing.

54:31

>> Yeah.

54:32

>> Do you know anyone else like that in

54:34

another domain? He seems like such a

54:36

unique

54:38

>> person relative to his peers.

54:41

>> Maybe Jensen Jensen has like this like

54:44

very defined, you know, agree with it or

54:47

not, but it's like this view of

54:50

who he is, what the company is, and so

54:53

forth that's kind of gone

54:56

across

54:58

eras. Um,

55:01

but it's still the same thing, right?

55:03

Like it's not that like it played in

55:06

gaming, it played in Bitcoin, it plays

55:09

in AI, but it it's still Nvidia. Like

55:12

it's not he never thought he had to

55:15

change the name of the company. He's

55:16

gotten better over the years, but in a

55:18

weird sense, it it never felt like he's

55:20

trying to be current,

55:22

which like Nas never kind of feels like

55:24

he's trying to write a hit.

55:27

>> Can you tell the story of the work

55:29

you're doing with the Vegas Police

55:31

Department? And I'm asking about this

55:32

one because it's super interesting, but

55:34

also because it feels like uh an

55:37

interesting different kind of example of

55:39

what the application of this

55:41

constellation of new technologies might

55:42

allow for in terms of improvement

55:44

efficiencies. You know, it's just such

55:47

an interesting case study.

55:48

>> A couple things about the Las Vegas

55:50

Police Force were intriguing to me. The

55:52

biggest one was uh they they're kind of

55:55

they were different than other police

55:58

forces in the country because they're a

55:59

big metropolitan area that's not run by

56:01

the chief of police, but run by the

56:03

sheriff. And the reason that's important

56:05

is the sheriff uh is an elected official

56:09

and does not report to the mayor. So

56:11

they never got caught in the big

56:14

political movement and defund the police

56:15

and they were the one of the only cities

56:17

that didn't reduce the police budget or

56:20

anything like that. So they kind of

56:22

stayed intact and they're also

56:23

interestingly the one or the one that I

56:26

knew that never militarized and they do

56:28

community policing and you can see it in

56:30

the numbers. So the murder clearance

56:33

rate in Las Vegas is the highest murder

56:36

clearance rate meaning they they uh

56:38

solve the murder 94%.

56:41

>> And you know I think San Francisco is

56:43

like 75% and then Chicago's like in the

56:45

30s and the national average is below

56:47

60. And I asked him, I said, you know,

56:49

why why is your murder clearance rate so

56:52

high? And the sheriff, Kevin McMahill,

56:56

said, "Ben,

56:59

you know, when somebody is murdered,

57:01

there's always somebody who knows who

57:02

did it. They just don't talk to the

57:04

police." So, but they talk to us because

57:06

we're part of the community. Like, they

57:08

know us. And so, I was like, "Wow,

57:10

that's a great kind of environment to

57:13

see if this new technology worked." And

57:14

I knew about all the public safety

57:16

technology because we invested in

57:17

through American dynamism. So I I was

57:20

like, look, we're going to become the

57:22

highest tech police force in America,

57:24

hopefully the world, and I'm just going

57:26

to fund it. And so I bought, you know,

57:28

we've got a drone program and we've got

57:30

u, you know, prepared 911 and we've got

57:33

flag safety, you know, AI cameras. If a

57:35

an emergency call, if a 911 call comes

57:38

in or if a gunshot goes off, there will

57:40

be a drone deployed in there within 90

57:42

seconds. And then that drone uh video

57:46

feed will be in every police officer's

57:49

phone in the vicinity like instantly.

57:53

Since we started the program, I think

57:54

crime is down over 50%. Um and then uh

57:59

you know shooting of suspects by police

58:02

is down like close to 75%. But

58:05

everybody's safer. And I think this is

58:06

the thing that that was the most

58:08

surprising to me on the technology

58:11

deployment um is that so when you talk

58:14

to the police they go look the problem

58:16

is the descriptions cause like half the

58:20

violent confrontations. I'm like well

58:22

what do you mean? So somebody jacks a

58:25

car. There's a baby in the back seat. We

58:28

get a description of the car. It's a

58:31

2004 Hyundai that's blue. Well, it's

58:37

really a 2008 Hyundai that's green. But

58:39

we pull a guy over in a 2004 Hyundai

58:42

that's blue and you know that person has

58:45

had bad experiences with the police and

58:48

you know now he's got a gun in the car

58:50

and all of a sudden we've got an

58:52

incident and like an innocent citizen

58:54

gets harmed or police gets shot with AI

58:56

camera. We know that's the car. That's

58:59

it. Uh, and we know there's a baby in

59:01

the car. And so we're not sending one

59:03

guy with a gun to see if that's the guy.

59:05

We're sending a whole squad. Um, and

59:07

we're apprehending them safely. And so

59:11

everything um about like policing is

59:13

inherently dangerous, but intelligence

59:15

makes it dramatically safer. And so I'm

59:18

a huge believer in this technology for

59:21

making everybody safer. You know,

59:23

suspects, criminals, citizens, police,

59:26

everybody. The other kind of uh knock on

59:28

effect is it's kind of put the pride

59:32

back into policing. We used to have a

59:33

big problem in Vegas where uh you know

59:37

because nobody wanted to be a police

59:39

officer, we were lowering the standard,

59:41

but now the standard is really high. So

59:43

between the drone center, which is like

59:45

super state-of-the-art, and then you

59:47

have these um cyber trucks that look so

59:51

like amazingly futuristic and cool

59:53

driving around like everybody wants to

59:55

be a police now. And Las Vegas happens

59:57

to have the highest concentration of

60:00

veterans uh in the country. So plenty of

60:03

like super qualified people to choose

60:05

from. They all want to be police. Uh

60:07

it's so that's all gone really well.

60:09

>> The last question I ask everyone is the

60:10

same. What is the kindest thing that

60:12

anyone's ever done for you?

60:14

>> A mentor of mine, a fellow by the name

60:16

of Ken Coleman, who uh was a big

60:18

executive at Silicon Graphics. And when

60:20

I was um

60:23

uh I guess a sophomore in college, uh I

60:26

got an introduction into him and uh he

60:29

gave me a job as a summer intern. And

60:32

without that job,

60:34

I don't know that I ever get to Silicon

60:36

Valley or or that whole thing. So I

60:39

would say that's probably that that was

60:40

the highest impact. Just he didn't have

60:43

to do that thing that anybody did for

60:45

me.

60:46

>> It may interest you that that is the

60:48

most common form of answer. uh across

60:50

500 of these someone that like took a

60:52

bet when they didn't need to. Ben,

60:53

pleasure to finally do this with you

60:54

after a couple years of uh of watching

60:57

you and and learning from you. So, thank

60:58

you so much for your time.

60:59

>> Thank you, Patrick. It was fun.

61:05

>> You know how small advantages compound

61:07

over time. That's true in investing and

61:08

just as true in how you run your

61:10

company. Your spending system is your

61:12

capital allocation strategy. Ramp makes

61:14

it smarter by default. Better data,

61:15

better decisions, better economics over

61:17

time. See how at ramp.com/invest.

61:20

As your business grows, Vanta scales

61:22

with you, automating compliance and

61:23

giving you a single source of truth for

61:25

security and risk. Learn more at

61:27

vanta.com/invest.

61:29

Every investment firm is unique and

61:30

generic AI doesn't understand your

61:32

process. Rogo does. It's an AI platform

61:34

built specifically for Wall Street

61:36

connected to your data, understanding

61:37

your process, and producing real

61:39

outputs. Check them out at

61:40

rogo.ai/invest.

61:42

The best AI and software companies from

61:44

OpenAI to cursor to Perplexity use work

61:46

OS to become enterprise ready overnight,

61:48

not in months. Visit workos.com to skip

61:51

the unglamorous infrastructure work and

61:53

focus on your product. Ridgeline is

61:55

redefining asset management technology

61:57

as a true partner, not just a software

61:58

vendor. They've helped firms 5x and

62:01

scale, enabling faster growth, smarter

62:03

operations, and a competitive edge.

62:04

Visit ridgidelineapps.com

62:06

to see what they can unlock for your

62:08

firm.

Interactive Summary

In this wide-ranging conversation, Ben Horowitz, co-founder of Andreessen Horowitz, discusses the current state of technology, the transformative power of AI, and his mission to ensure American competitiveness. He reflects on the evolution of venture capital, the importance of culture as defined by actions rather than platitudes, and the management lessons he learned from mentors like Andy Grove. Horowitz also shares insights into how technology is being applied to solve real-world problems, such as crime reduction in Las Vegas, and discusses the influence of his personal background and hip-hop on his leadership style.

Suggested questions

6 ready-made prompts