HomeVideos

Robinhood CEO Vlad Tenev on tokenizing stocks, expanding access to private shares, fintech's future

Now Playing

Robinhood CEO Vlad Tenev on tokenizing stocks, expanding access to private shares, fintech's future

Transcript

716 segments

0:00

I am Vlad Tennv, the founder of Robin

0:03

Hood.

0:07

We're talking about Robin Hood. The

0:08

stock's more than doubled since its last

0:10

report.

0:10

>> It stock surged 180% this year after

0:14

nearly doubling in 2024. The

0:16

>> shares of the trading platform now up

0:17

more than 400% in the last year.

0:20

>> Glad your presence there speaks volumes.

0:22

Robin Hood Gold, which hit a record 3.5

0:24

million subscribers. Most financial

0:27

services get worse the more money you

0:29

have, but we wanted to kind of invert

0:31

that.

0:32

>> Ladies and gentlemen, please welcome

0:36

Robin Hood CEO Vlad Tenner,

0:40

[Music]

0:46

>> my guy.

0:48

>> Good to see you, brother.

0:50

>> Good to see you.

0:50

>> Good to see you.

0:51

>> You're the reason Jake Health's here. I

0:54

mean

0:55

>> I mean it is a great story.

0:57

>> It's so true. Jake Cal bumbles into

1:00

eight shares of Robin Hood. Lad builds a

1:02

hundred billion dollar companies. It's

1:04

unbelievable.

1:04

>> I mean people know me for the Uber

1:06

investment which four or $5 million

1:09

valuation. But there's

1:10

>> When do I flip that?

1:12

>> Well, yeah, it's going to take a little

1:13

more. I think you got 20x left to go.

1:15

But

1:16

>> wait, was JL the third or fourth

1:18

investor in Robin Hood? It was $20

1:20

million valuation, but it's a good

1:22

story, I think, because you hadn't

1:24

launched. And we're at Antonio's

1:26

Nutouse. I went for a drink with my

1:29

friend Ado. He brought his college

1:30

roommate, Elon. We're hanging out at

1:32

Antonio's Nut House in Palo Alto. Rest

1:34

in peace, Antonio.

1:35

>> Keep dropping these names. One second.

1:36

>> So Vlad comes over. We we Vlad and I

1:38

knew each other a little bit. And Vlad

1:40

pitches me on this idea and he says,

1:43

"Um, I'm a quant." I said, "What's a

1:46

quant?" He said, "Quant analysis." I

1:48

said, "Yeah, I heard of it. Um, hit me

1:50

with the idea." And then he goes, "Is

1:52

that Elon Musk?" I said, "Yeah, just hit

1:54

me with the idea. I know you got a

1:55

startup." He says, "Well, I I want to

1:57

get

1:59

uh this generation, these millennials,

2:01

these Gen Z's. I want to get them to

2:03

trade stocks." I said, "Love it. They

2:06

don't care about like getting a driver's

2:08

license. They they don't they're still

2:10

on their mom and dad's Netflix. You're

2:12

going to try to get people who don't

2:13

care about the future to trade stocks?"

2:15

He says, "Yeah." I said, "What's the

2:16

business model?" I said, "This is the

2:18

best part. We're going to let him trade

2:19

for free." So I said, "Okay, let me

2:22

repeat this back to you, kid. You want

2:23

to get a group of people who don't have

2:25

any interest in the future to trade

2:28

stocks and then

2:30

>> we have 30 seconds.

2:30

>> You're going to make money. You're going

2:31

to make money.

2:32

>> Thank you, Vlad."

2:35

>> I said, "I'm in. I'm in."

2:37

>> Not only that, but he said, "This is

2:38

probably the best idea you'll ever

2:40

have."

2:41

>> I did say that to him, too. I was like,

2:42

"This is the best idea. What if it

2:44

works?" And here we are 10 years later.

2:46

what's worked.

2:47

>> And last week you were added to the S&P

2:48

500.

2:49

>> I was

2:50

>> last week or yesterday? Two days ago.

2:51

>> It was Friday.

2:52

>> I mean, what a huge accomplishment.

2:54

>> Thank you.

2:55

>> I mean, I think I think it was cuz I

2:57

rejected you for a job, right?

2:58

>> It you heard about that.

2:59

>> This is a series.

3:01

>> Thank you guys for upgrading me, by the

3:03

way. I guess that that's been the best

3:04

part of being added to the S&P going

3:06

from just a Jason interview to to the

3:08

whole squad.

3:09

>> We were doing the rehearsal yesterday

3:10

and everyone wanted to do the the

3:12

interview, so we said, "Let's all do it

3:13

together."

3:14

>> Why don't we all get in here? Can I uh

3:15

let's let's maybe start. So look, you've

3:17

built an incredible business. Um there's

3:19

a part of it that looks like the you

3:22

know what comes after the Erades of the

3:24

world etc. But there's an enormous other

3:26

part of your business and there's all

3:28

these emergent paths. I want to just

3:30

start by double clicking in something

3:31

that you announced a few months ago in

3:33

France and maybe you want to talk us

3:36

through it what the goal was. You got a

3:39

lot of support but you got a lot of

3:40

blowback as well. There was a lot of

3:42

people that were like, "Wow, this is a

3:43

little too disruptive tokenizing these

3:45

stocks, putting them on the blockchain."

3:47

Maybe talk us through the business and

3:49

then just double click on that narrow

3:50

thing so we can understand what you're

3:52

up to.

3:53

>> Yeah. So, we had an event in in the

3:55

south of France that we called to catch

3:57

a token. And the idea behind that event

4:00

was we wanted to show what Robin Hood,

4:04

the app, the platform would look like if

4:06

it was built from the ground up on

4:09

crypto technology. So what that looked

4:11

like was stocks on blockchains. Uh

4:14

obviously we added a bunch of

4:16

cryptonnative features like pers. Um we

4:19

launched in now 31 countries. And then

4:23

uh we also wanted to demonstrate to the

4:25

US what the power of putting traditional

4:28

financial services on blockchains was.

4:31

And to me, I think a lot of people talk

4:34

about 24/7 stock trading, instant

4:37

settlement, these things that do have

4:39

real value, but I think the most

4:41

powerful thing is taking inaccessible,

4:44

illlquid assets and making them

4:47

available. And and so we we were

4:49

actually, I think, the first to tokenize

4:51

OpenAI and SpaceX and we made that

4:54

available to our retail customers in

4:56

Europe in the form of a giveaway. and

4:58

and that was very very exciting, you

5:00

know, not not without its controversy,

5:02

but I felt like it was such a powerful

5:05

thing.

5:06

>> How do you do it? How did you enable

5:07

that?

5:08

>> It's actually very similar to uh

5:10

stablecoin in a way. So, if you're a

5:13

stable coin issuer, and this it's a

5:16

little bit oversimplifying, but you can

5:17

think of it as we keep some dollars or

5:20

treasuries in a bucket over here. uh we

5:24

mint and burn tokens against that bucket

5:26

but back at one to one and the tokens

5:29

can actually trade publicly on a variety

5:32

of of blockchains. So it's just

5:34

extending that token that that

5:36

tokenization concept from stable coins

5:39

to uh public and and private.

5:42

>> So you had to go and secure your own

5:44

block of SpaceX and OpenAI stock and

5:48

>> and then put it somewhere. Were the

5:49

companies okay with it? Um, it depends.

5:53

Uh, I think I think that, uh, I'll say a

5:57

couple of things. Um, I think a lot of

6:00

people are okay with it in principle,

6:02

but if you're a company and you're

6:05

focused on your mission like OpenAI is,

6:07

and you hear about some new thing, uh,

6:11

it's kind of a distraction. So, I I

6:13

don't really blame them for tokeniz

6:15

tokenization or private access not being

6:17

their their top priority, but I did want

6:20

to be the first to tokenize open AI. So,

6:22

uh

6:23

>> did Sam give you a call?

6:25

>> Uh I have had a couple conversations

6:27

with with Sam before and after.

6:29

>> Um cuz yeah, he's a spicy individual.

6:31

What was his take? Did he tell you to

6:32

stop?

6:34

>> Um I think that actually uh I I'd like

6:37

to think we get along quite well. I

6:39

think he understood why we were doing

6:40

it. Again, the distraction aspect when

6:44

they have so much going on uh is is a

6:47

real thing. But at the end of the day,

6:50

>> where do you take it from here? Are you

6:51

going to go and get 50, 100 of these

6:54

well-known private companies? Is that

6:55

the goal? Is it every private company?

6:57

How like what do you do from here to

6:59

build on top of it?

7:00

>> Yeah. So, we we've been hard at work

7:03

trying to figure out how to do it in the

7:05

US. I think that's that's uh everyone's

7:08

interested in that since the France

7:09

announcement. obviously expanding what

7:12

we do in Europe as well and there will

7:15

probably be different mechanisms in the

7:18

US and and in Europe at least for some

7:20

time but uh yeah you you should expect

7:23

that we go bigger and deeper into the

7:26

space and have plenty of things to do in

7:28

the future.

7:28

>> I'm curious what the relationship has

7:31

been with the new administration. the

7:33

last administration, you know, was not

7:35

very pro- innovation, crypto, and now

7:38

you got David uh running that

7:40

specifically. How has the change in the

7:43

administration changed how you look at

7:45

innovation at Robin Hood? And then Sax,

7:47

I'm sure you have some follow-up

7:48

questions here.

7:49

>> Yeah, it's been very positive. I mean,

7:51

just uh by nature of how many times I've

7:53

been to Washington, uh the last

7:56

administration

7:58

didn't invite me to the White House

7:59

once. Uh, I asked for meetings and they

8:02

wouldn't even meet in person. They were

8:04

all working remotely uh until I think

8:07

2023, early 2024. Um,

8:11

>> it was funny how remote work kind of

8:13

broke down along political lines. It's

8:16

sort of like the the Republicans wanted

8:18

to get back into the office.

8:18

>> I think you're referring to remote work

8:20

as not working.

8:22

>> I mean, that is uh I wasn't going to say

8:24

it, but uh yeah. Um,

8:27

yeah, it was funny how that works out,

8:28

but it's it's been very positive. I

8:30

mean, last administration,

8:32

we were playing a lot of defense. It was

8:34

just sort of like one enforcement action

8:38

and we had a wells notice and all

8:41

aspects of our business were sort of

8:45

under assault. So I think the most

8:47

direct thing was when that went away, we

8:50

had to think, okay, well there the

8:52

administration now wants to work with us

8:54

rather than just trying to attack us

8:56

from all these angles. And for a while

8:58

we didn't even know how to operate in

9:00

that environment because we were just

9:01

completely not used to it.

9:03

>> Yeah, you were on your

9:04

>> Have you met with Elizabeth Warren?

9:06

>> Uh I have not. No, I just receive

9:09

letters from time to time. But

9:12

>> she really But she hates you. Well, I

9:14

told me personally, she really can't

9:15

take it.

9:16

>> No, the reason the reason I asked is

9:19

what is the core motivation of the idea

9:23

of we need to enforce, we need to kind

9:24

of restrict, we need to prohibit. Is it

9:26

consumer protection that the belief is

9:29

that systems like yours that are more

9:30

open, more accessible, more usable, more

9:34

consumers will trade more and

9:35

potentially lose money and therefore

9:37

they have to try and play a role in

9:39

restricting consumer access to these

9:40

markets and these marketplaces. and

9:42

that's what they're mo ultimately kind

9:44

of driving towards or do you think that

9:45

there's something more vested

9:47

interestwise that's motivating?

9:49

>> I mean I think I think there's probably

9:50

both. Certainly consumer protection is

9:54

the stated reason but obviously these

9:57

folks have funders and backers and lots

10:00

of interests. I mean there's powerful

10:02

there's powerful financial services

10:04

companies in the state of Massachusetts.

10:06

So, I don't know what's happening behind

10:08

the scenes, but you know, I do think the

10:11

consumer protection angle is what

10:13

they're kind of

10:15

>> pulling out. How do you Yeah. How do you

10:17

look at this type of innovation and your

10:20

role at the White House to support it

10:23

and to foster it while still, you know,

10:26

having some rules on the field?

10:28

>> Well, I think Vlad's uh vision around

10:30

tokenization is very exciting. I like

10:32

you wrote a oped I think it's in the

10:34

Washington Post that I thought was very

10:35

good on this this topic. We now have a

10:38

regulatory framework in place the Genius

10:40

Act which the president signed in July

10:42

that creates um the the the the set of

10:45

rules for stable coins which are just

10:47

tokenized dollars. And like Vlad's

10:49

saying if you can tokenize a dollar you

10:50

can tokenize anything. You just

10:52

basically create a reserve of that asset

10:54

in a secure account at a bank somewhere

10:57

or a broker and then you mint tokens on

10:59

a onetoone basis. So I think it's very

11:02

exciting is there's no reason why we

11:03

can't tokenize. Let's start with public

11:06

securities. I think that's the easy case

11:09

because with public companies there's

11:11

already disclosure requirements. There's

11:13

an abundance of information and anybody

11:15

can buy a public security, right?

11:17

Because of those disclosure

11:18

requirements. Um and the companies don't

11:22

really care who their stockholders are,

11:24

right? I mean because they know that the

11:26

public owns these securities. So what we

11:28

could get right away with tokenized

11:31

securities, public securities is like

11:33

you were saying a 24/7 global

11:35

marketplace uh with instantaneous

11:38

blockchain based settlement and that

11:40

could be really exciting. There's no

11:42

reason why trading has to be on this

11:44

like 9-to-F5 exchange with all this, you

11:47

know, we could enable stocks to trade as

11:50

easily as you, you know, transfer a a

11:53

stable coin. Now the private security

11:55

part is interesting that that is I think

11:59

more complicated because first of all

12:01

the companies like you're saying they do

12:03

care who their shareholders are and they

12:06

do generally restrict those things and

12:07

that's why you probably got the phone

12:08

call from Sam and then the regulators

12:11

care also because there's not as much

12:13

public disclosure. So they there's more

12:16

uh of an impetus to protect the public.

12:18

By the way I'm not saying we can't get

12:19

there on private securities. Um about a

12:22

decade ago a ago I founded

12:24

a startup to tokenize uh real estate

12:27

called Harbor and we were just way too

12:29

far ahead of the curve but that that was

12:30

basically to to um tokenize uh private

12:34

real estate securities. So I think we

12:36

can get there but I think the place to

12:38

start that would be really exciting

12:39

would just be like let's do the public

12:41

securities first because it's easy or

12:43

easier from a regulatory standpoint and

12:46

then we can work our way into private.

12:47

>> Yeah, it's certainly easier

12:48

technologically. I mean, we we've we've

12:51

made both uh available to some extent in

12:53

the EU. Um I think private could be more

12:58

meaningful long term and I I'll I'll

13:00

tell you why I think so. If you look at

13:02

the technologies that are transforming

13:06

society right now and that we feel so

13:08

optimistic about over the next 5 years,

13:11

it's AI and to some extent, you know,

13:14

space exploration. And I think with AI

13:17

in particular, there's a lot of fear

13:19

right now. I mean, you talk to a random

13:21

person on the street, more than half the

13:23

time, they're a little bit nervous about

13:25

what AI is going to do to them. Now,

13:28

imagine the scenario if you know 20 to

13:32

30% of someone's net worth is in AI

13:35

companies. Now, suddenly they're not

13:38

fighting against this thing. They want

13:39

it to succeed because if AI succeeds

13:43

>> entrepreneurship as a way to let more

13:45

people participate in the boom.

13:47

>> Yeah. Because I worry about the status

13:49

quo. I mean these AI companies in

13:51

particular are getting into valuations

13:52

of hundreds and hundreds of billions

13:55

with zero retail ownership. And that

13:58

technology could completely disrupt you

14:02

know how normal people live their lives.

14:04

And we actually expect it to drive that

14:06

sort of disruption because if if you

14:08

look at Cathy's presentation, you know,

14:11

you're talking about negative inflation,

14:13

high GDP growth rates, giant

14:16

productivity improvements. I don't think

14:18

you're going to get there without some

14:19

significant labor force disruption.

14:22

>> Okay. So what do you need from let's say

14:24

the US government broadly whether it's

14:26

the SEC or whether it's maybe new

14:29

legislation on Capitol Hill what what

14:30

exactly do you need to bring about this

14:33

revolution?

14:34

>> I think I think uh relaxation of

14:36

accreditation uh standards towards more

14:40

self-certification. you you mentioned uh

14:43

a test. Um I think a test I consider to

14:47

be one form of self-certification,

14:50

but the simplest form is just someone

14:52

saying, "I understand the risks. I

14:54

understand I could lose 100% of what I

14:56

put in this investment." You could even

14:58

put like a skull and crossbones.

15:00

>> No crying in the casino.

15:02

>> Yeah, exactly.

15:02

>> You could put somebody crying in the

15:04

casino. Yes.

15:05

>> Yeah.

15:06

>> Literally in the app.

15:07

>> No crying in the casino. I think

15:10

>> I guess I mean the point you're making

15:11

is you can't on the one hand cry for

15:14

access and on the other hand cry in the

15:15

casino.

15:16

>> Exactly.

15:17

>> Can't do that.

15:17

>> But the executive order on 401k access I

15:21

think was a step in the right direction.

15:23

So we could also ease into it by

15:25

extending that to individual retirement

15:28

accounts which are great short-term

15:30

vehicles and then give it to

15:32

>> you need a secured interest. Why can't

15:34

you just create a synthetic or like a

15:36

fast contract saying if and when open AI

15:39

goes public cuz you can see how many

15:41

shares there are. You know what the

15:42

legal registration of the corporation

15:43

is. Can't you create like a synthetic

15:45

contract that just trades the value of

15:47

the stock and ultimately needs to settle

15:49

at some point after the company goes

15:50

public?

15:51

>> We can't do that currently. Uh and and

15:54

OpenAI in particular is a tricky one

15:56

because yeah, pick any other LLC or C or

16:00

something, right? Um but yeah, we're

16:03

we're continuing to look at all angles,

16:05

but I I think some clarity would be

16:07

helpful

16:08

>> because this is the whole value of of

16:09

like sorry Chimat, but like futures

16:11

markets and prediction markets is you

16:12

can effectively create a synthetic on

16:14

some underlying without actually having

16:16

ownership or a secured interest in the

16:18

underlying or delivery of the commodity.

16:20

You could basically just say when this

16:21

thing goes public, it's above 20 bucks a

16:23

share, below 20 bucks a share on some

16:25

number of days after something like

16:27

that. But I mean I guess one one

16:29

question for you is is that where

16:30

prediction markets can take us?

16:31

>> So um the the difference within with

16:34

prediction markets is you can create a

16:36

prediction market but it has to have an

16:38

expiry date for the contract. So for

16:40

example we have a prediction market live

16:43

on the platform now about which

16:45

companies are going to IPO. You could do

16:47

something like that. There have been

16:49

prediction markets uh in the past, not

16:52

on our platform, but some other

16:53

platforms that make a market around the

16:56

IPO price, but if you just want exposure

16:59

to underlying equity in a private, I I

17:02

don't think we can do that.

17:03

>> How do you think about, you know, the

17:05

one criticism people have had, which is,

17:08

hey, we have got a young generation.

17:10

They're frisky. They want to take all

17:12

this risk. They want to bet. and your

17:15

responsibility as a platform that is

17:17

giving that access if you're the on-ramp

17:20

the education you give and and I

17:23

remember with the options and people

17:24

being able to short you you came up with

17:27

an incredibly elegant solution when you

17:29

try to short something you give people a

17:31

test you give them education in that

17:33

moment before they do it so so how do

17:36

you think broadly about young people

17:39

getting into wagering they're playing

17:41

cards you know they're they're betting

17:43

on fantasy football and doing sports

17:45

betting, but they also want to, you

17:47

know, have their hand in crypto and in

17:49

puts and calls and and and pretty

17:51

sophisticated stuff. So, what's your

17:53

responsibility as a platform then in

17:55

introducing them to those sophisticated

17:58

>> um ways of betting and investing?

18:01

>> I have a lot of thoughts actually when

18:03

uh you interviewed me for a job in uh

18:06

2008, which by the way was the only it

18:08

was one of two like final job interviews

18:12

that I got. Most people just rejected

18:14

me. I never even got a call from from uh

18:16

from Google or any of the others, but uh

18:19

it was either Weather Bill, later

18:21

Climate Corp, uh or Optiver, where I

18:24

interviewed to be an options trader. Um

18:27

so I got very very close.

18:29

>> We had a math team.

18:31

>> Yeah, I got very close.

18:32

>> Was he like in the interview?

18:34

>> He you probably don't even remember.

18:36

>> You remember Alex Machulka? Yeah, I

18:39

think he was uh he

18:42

>> um

18:43

>> I forget my recruiter's name, but that

18:45

guy was great. Anyway, um and then I

18:48

became an entrepreneur. So uh it might

18:51

not surprise you to

18:54

know that me personally uh I'm I'm sort

18:57

of like adverse to control on the level

19:00

of risk that I would take because my

19:02

entire career path was sort of like

19:05

maximally levered bet on one company

19:07

which is the one that I started.

19:09

>> Um so I would be reluctant to discourage

19:12

people from being entrepreneurs to doing

19:14

what they want to do with their money or

19:15

time. Of course, I'm I'm in favor of

19:18

like reasonable things like you should

19:21

it should be clear to you what you're

19:22

investing in. Um but yeah, generally

19:26

speaking, I think if it's available to

19:28

wealthy people, high netw worth

19:30

individuals, it should be made available

19:32

to retail as well.

19:33

>> Let me broaden the conversation. Um it

19:36

used to be historically we would have

19:37

banks, we would have brokerages, we'd

19:39

have payment processors, we'd have

19:41

merchant acquirers. They were all

19:43

disagregated. They could all be public.

19:45

they could all build thriving companies.

19:47

Now with stable coins and everything

19:50

else, there's this creeping convergence.

19:53

You're issuing a credit card. Coinbase

19:55

has a credit card. SoFi has a federal

19:57

banking license. Stripe just launched a

20:00

new L1 called Tempo.

20:02

Everybody's competing with everybody.

20:04

Tell us the scope of where you think

20:06

Robin Hood goes in the next four or five

20:08

years and what the financial landscape

20:13

and infrastructure looks like. The

20:15

Visas, the Mastercards, the JP Morgans,

20:17

what roles do these companies play as

20:20

you guys just become more and more

20:22

ambitious and girthy and big and you

20:24

know market cap and all that stuff.

20:26

>> Yeah, I think the industry goes through

20:28

periods of consolidation and then

20:31

divergence. I think Robin Hood has a

20:34

unique advantage which is that our

20:38

customers put an increasing amount of

20:40

their dollars into Robin Hood. So what

20:42

we're thinking about and and it became

20:44

pretty clear to us as soon as we rolled

20:46

out our second product. We kind of saw

20:48

what happened. Customers spent more time

20:50

on Robin Hood. The two products help

20:53

each other. So for example with

20:55

retirement we noticed you know the big

20:58

question was well if we launch

20:59

retirement is it going to cannibalize

21:01

the core brokerage business but what we

21:03

saw was the opposite. If someone opens

21:05

up a retirement account they they tend

21:07

to actually increase the amount they put

21:10

in their individual account and we saw

21:12

that again with the credit card. If

21:14

they're a credit card primary user top

21:16

of wallet they actually put more money

21:19

into into Robin Hood. So then that gets

21:22

us to a future where we ask ourselves,

21:24

can we be your comprehensive financial

21:26

platform? Can you put your direct

21:28

deposit into Robin Hood? Can you put all

21:31

of your money into Robin Hood? Can you

21:33

get to gold subscriber premium status as

21:36

soon as possible? And then can we get

21:38

all of your family members onto Robin

21:40

Hood as well and your kids? So yeah, I

21:44

don't think anyone's really thinking

21:46

about it from that angle, but I think

21:48

that there's going to be over 130

21:51

trillion that changes hands from silent

21:54

generation and baby boomers to younger

21:58

people. And and I think Robin Hood's

22:00

actually very well positioned to be one

22:02

of the, if not the number one primary

22:06

institution that benefits from that

22:08

transfer. We've got over a quarter

22:10

trillion assets on the platform already,

22:12

which seems like a a big number, but

22:14

it's actually just a drop in the bucket

22:16

compared to what's going to happen.

22:18

>> Where do you see the JP Morgans and the

22:20

V Mastercards and Visas? How do they

22:21

compete with an elegant product with

22:24

hundreds of millions of users that just,

22:26

you know, the product velocity that you

22:28

have, the risk you're willing to take?

22:31

>> Yeah. I mean, I think that uh if if you

22:33

think about an incumbent, they have

22:35

certain benefits. They're like very

22:37

muscular from a regulatory standpoint.

22:40

Like they know how to deal with

22:41

regulators. They've got global scale.

22:43

They've got, you know, tens, hundreds of

22:45

millions of customers, lots of assets.

22:48

But the disadvantage is that they're

22:51

sometimes slow to adopt new

22:53

technologies. They don't have the best

22:54

engineering teams. They can't move very

22:56

fast. Um, and they they can't hire the

23:01

best talent. And and so I think that we

23:04

don't have those downsides. We have

23:06

great talent. We move really quickly. We

23:08

use the best technology. We haven't been

23:10

super acquisitive historically, even

23:13

though we're doing more. And that

23:14

prevents us from being bogged down by

23:16

these like massive integration things

23:19

that take multiple years. Um, and so

23:22

it's a question of can we get the

23:24

benefits of scale while also maintaining

23:27

the nimleness of a technology startup.

23:29

Do you want to give the audience before

23:31

we run out of time, maybe last question,

23:34

but tell them about the LLM you guys are

23:36

building? This is a different project

23:37

for you.

23:38

>> Oh yeah.

23:38

>> Um and the goal of that and why you

23:40

decided to fund that sort of outside the

23:42

scope of Robin Hood.

23:44

>> Yeah. So uh he's talking about Harmonic,

23:47

which is a company that I started two

23:49

years ago and I'm chairman of and uh

23:52

completely separate from Robin Hood. and

23:55

and basically the goal there is to build

23:58

what we call mathematical super intell

24:00

intelligence. So this is um mathematical

24:03

reasoning that is uh exceeding the

24:08

capability of any individual human

24:10

researcher. And we we had a pretty cool

24:12

result a couple weeks ago where we

24:14

announced gold medal level performance

24:17

at the international math olympiad which

24:19

is the biggest uh mathematics

24:21

competition in the world. And and I

24:24

think to my knowledge we were the only

24:25

formal model.

24:26

>> You're the only formal one that got IMO

24:28

gold. Yeah.

24:28

>> Yeah. So open eye and Gemini. Open eye

24:30

Gemini

24:32

>> uh did it with informal

24:34

>> uh Gemini's in formal model got a silver

24:37

last year.

24:38

>> Explain why it's going to be so critical

24:40

to have a a mathematical super

24:42

intelligence model.

24:43

>> Yeah. So two reasons. Um one that has to

24:46

do with how these models are trained and

24:48

the other which is more of a consumer

24:50

painoint. The thing that we've figured

24:52

out with formal is how to verify that a

24:55

statement is true very precisely. And

24:58

when you're doing reinforcement learning

25:01

of these models, having a strong reward

25:02

signal is very helpful because you can

25:05

just discard all the data that's not

25:07

helpful and train on the high quality

25:10

like correct data. And when you're a

25:12

user of these AI models, sometimes they

25:14

hallucinate. Uh and and this is actually

25:17

not just a consumer problem, but also an

25:19

enterprise problem because basically if

25:22

you're a software engineer using a

25:24

coding model, your job has has become

25:27

over the past couple of years, it it's

25:29

turned from writing a whole bunch of

25:31

high-quality code to reviewing LLM

25:34

generated code and making sure that

25:36

that's correct. So in a world where

25:38

you've got, you know, LLMs producing

25:41

thousands and thousands of pages of

25:43

code, human verification just doesn't

25:46

scale, particularly for backend. So we

25:48

want to solve that problem.

25:49

>> Okay, give it up for David Sax's second

25:53

favorite.

25:54

[Music]

25:56

>> Good to see you guys.

25:58

>> Thank you. Thanks, brother.

26:01

I'll see you in Vegas.

Interactive Summary

This video features an interview with Vlad Tenev, the CEO of Robinhood, where they discuss the company's growth, its inclusion in the S&P 500, and its future vision. Tenev talks about the company's mission to democratize finance, including experiments with tokenizing private stocks like OpenAI and SpaceX on the blockchain to provide retail access to assets previously reserved for the wealthy. The conversation also covers the regulatory landscape, the company's goal to become a comprehensive financial platform, and Tenev's separate venture, Harmonic, which focuses on developing mathematical super-intelligence.

Suggested questions

4 ready-made prompts