HomeVideos

Trump AI Speech & Action Plan, DC Summit Recap, Hot GDP Print, Trade Deals, Altman Warns No Privacy

Now Playing

Trump AI Speech & Action Plan, DC Summit Recap, Hot GDP Print, Trade Deals, Altman Warns No Privacy

Transcript

2565 segments

0:00

How much founder mode did you do?

0:01

>> Are you saying that I popped an ALP? I

0:03

need an ALP right now. Hold on. You

0:04

don't need anything right now. Are you

0:06

chewing it? What are you doing?

0:07

>> No. You put this nicotine pouch, you

0:09

upper deck it, releases it, and then you

0:11

become a god.

0:12

>> Is that the app that Tucker sent you?

0:14

>> Yeah. Tucker and I are going to do a

0:16

crossover.

0:16

>> Wait, did you work out a side hustle

0:18

here?

0:18

>> I haven't presented it to the group for

0:20

a vote yet. You're pre-

0:21

>> Wait a second. Are you being paid for

0:22

this plug right now?

0:24

>> Yes.

0:24

>> I'm just saying if you use the promo

0:26

code JCL.

0:27

>> Wait a second. Promo code J 15.

0:32

>> Okay, he broke up, which is good.

0:33

>> Is he on drugs? Is he taking drugs?

0:35

>> He's on drugs.

0:36

>> No, I'm not on drugs.

0:37

>> And he's doing a deal with

0:38

>> This is like a PSA for not taking this

0:41

stuff. You're so out of control.

0:42

>> Did you take two of them? What are you

0:44

doing?

0:44

>> I thought this stuff relaxes you. What

0:46

the hell is going on? Your internet's on

0:49

the fritz, too.

0:50

>> I fixed it. I fixed it. I fixed it. That

0:52

was Putin. Putin's got my internet.

0:55

>> Putin's got my

0:56

>> Oh my god. What flavor are you eating or

0:58

using?

0:59

>> Oh, today's chilled mint. Today's

1:00

chilled mint.

1:01

>> You don't seem very chill. You seem

1:04

>> This is the first one.

1:05

>> Agitated and angry.

1:07

>> No, I'm trying to get us back to that

1:08

original all-in energy where we laughed

1:11

and we had fun and we enjoyed each

1:12

other's company.

1:14

>> No, but Jakeal, seriously, do you have a

1:16

side deal going on with Alp right now?

1:18

>> No, I don't have a deal yet. I don't

1:21

know if I have a deal. There's no deal.

1:23

I'm texting Tuck right now just to cut

1:25

you.

1:27

[Music]

1:29

Let your winners ride.

1:32

[Music]

1:36

>> We open sourced it to the fans and

1:38

they've just gone crazy with it. Love

1:40

you guys. Queen of

1:44

>> All right, everybody. Welcome back to

1:46

what Jensen from Nvidia has confirmed is

1:50

the number one podcast in the world.

1:52

Yes, the All-In podcast is here. We had

1:55

an amazing time in uh DC last week and

1:59

we'll get into that. But uh hey

2:02

Freeberg, you crushed it on all those

2:05

incredible speakers last week. 10 days

2:08

you had to pull off that event Freedberg

2:11

and you did it. Chimath and I just

2:13

parachuted in to DC last week for the AI

2:16

summit. Sax was busy working with pus to

2:18

get all those executive orders done.

2:20

Take us behind the scenes, Freeberg. all

2:23

of these incredible speakers. You got

2:24

Lisa from AMD. You had Lutnik, I liked

2:28

him. Bessent, I liked him. We had to say

2:31

no to a lot of tech company CEOs that

2:33

found out about the event and wanted to

2:35

speak on stage. So, there was a big kind

2:37

of cut off that we had to make around

2:39

making sure that we got our message

2:41

across. I think if you watch the

2:43

content, we talked briefly about it at

2:44

the beginning, but the focus was really

2:46

on trying to dispel the negative AI

2:50

narrative and myth that AI is just here

2:52

to destroy jobs because there's this big

2:54

economic boom that's happening both with

2:56

respect to new industries that are

2:58

emerging, which is why we showcased

2:59

Hadrian and others, but then also the

3:02

infrastructure needed to support the AI

3:05

race with data centers, chips, mining,

3:09

and energy. And so we highlighted each

3:10

of those four industries. And then the

3:12

cabinet people found out about it and

3:14

wanted to get involved. So we were

3:17

unfortunately squeezing people on and

3:18

off stage. It's kind of crazy to tell

3:20

the secretary of treasury he has to get

3:22

off the stage because he's passed his

3:24

20-minute allocation. But uh we had to

3:26

line everything up so that the president

3:28

could get his Secret Service detail to

3:30

clear the stage and get set up in time.

3:31

That's why we were rushing everyone. But

3:33

man, what a week. What a rush. It was

3:35

awesome. Thanks to David Saxs for the

3:37

leadership and pulling it all together,

3:39

bringing those folks to the table. And

3:41

Sax, congrats on getting your EO signed

3:43

and your action plan published. That was

3:45

pretty cool. Pretty awesome to meet the

3:47

president and meet all those cabinet

3:49

members and have all of this day come

3:51

together because of the work you've been

3:52

doing. How does it feel? Like, Sax, how

3:55

are you doing in the afterlow there? I

3:57

can see that you're in the afterlow. You

3:59

sent me a picture of the four besties

4:02

with our incredible 47th president. How

4:05

you feeling right now?

4:06

>> Are you going to put that on the screen?

4:08

>> I may have it. I don't I don't know if

4:10

that's allowed. Are we allowed to put

4:12

that on the screen? I don't know what

4:13

the protocol is.

4:14

>> Yeah, I think we can.

4:15

>> Yeah. I mean, I haven't gotten my

4:17

picture. Um I did notice that I was

4:19

unfortunately when they took the picture

4:20

of the four of us with the president,

4:22

somehow I got cropped out by accident. I

4:24

think maybe they weren't using a wide

4:26

lens.

4:26

>> Wait, Jason, what was it like for you to

4:28

meet the president? Cuz just for the

4:29

audience, we all stood in line and then

4:31

we took a photo with the president

4:32

backstage and then we did a photo with

4:34

the four of us. But Jason, when you had

4:36

your moment with the president, what did

4:37

you say? Did you ask him about

4:38

immigration?

4:40

>> Did you ask about I have your photo with

4:42

the president?

4:42

>> Oh, it's on my phone.

4:44

>> Did you Did you bring up solar panels

4:45

with him? Like what was your big moment

4:47

all about?

4:48

>> I didn't know we were taking a picture.

4:49

That was like sprung on me. So I

4:54

was like, "Oh, we're taking a picture."

4:55

So my brother Josh who runs security for

4:57

us was like they need you in the back to

5:00

take a picture with the president and I

5:01

was like yeah I'm good I I got to I got

5:03

to prepare for you know some

5:05

>> Oh you were going to pass.

5:06

>> Well I thought he was joking with me.

5:08

>> So I was like yeah I'm good. I'm good.

5:10

So he's like no no I'm serious.

5:12

>> They're they're taking pictures with the

5:13

president. I was like we are okay. So, I

5:15

ran back and uh they put us in line and

5:19

then I was like, I think I'm getting

5:20

punked here because they kept repeating

5:22

to me, okay, Jason, you're last. You're

5:26

last. And they, you know, and I know you

5:28

guys like to put in a joke or two. So,

5:29

I, you know, I just got in line last.

5:31

And it's obviously, you know, it's it's

5:33

a big deal to take a picture with the

5:34

president. So, I didn't want to um, you

5:37

know, use that time inappropriately or

5:39

anything. So, I just said it's a

5:40

pleasure to meet you.

5:41

>> Just say it already. You like it. Just

5:43

say it. Just let's get it over with.

5:44

Just get it. Get to the end. You like

5:46

him. You tried not to. You know, you're

5:48

all Mr. Big Shot, Mr. Big Talk, and then

5:52

you got in front of him and you like

5:54

him. Just say it.

5:55

>> Uh, what I will say is,

5:58

>> Jesus Christ.

5:59

>> Like him or dislike him.

6:00

>> What a joke. You're such a goon.

6:01

>> I I had a great time. I had a great

6:03

time. Predictable. You're a predictable

6:06

goon. You know,

6:07

>> you don't even know what goon is. Okay,

6:09

stop riz. Stop oraura farming. You don't

6:12

know what gooning is. Okay. I had a

6:14

great time meeting him. It was a great

6:15

event or farming. Obviously, he's trying

6:18

to get his RZ up to impress his kids.

6:20

But, um,

6:21

>> it was great. And I didn't know what to

6:23

do in the picture. So, he did.

6:25

>> We can move on.

6:26

>> I

6:29

What do you think of his speech, Jason?

6:30

>> After he gave you a shout out, Jason

6:32

after the president gave you a shout

6:34

out.

6:34

>> I don't know about love. He said, "Even

6:36

Jason." How many times have you listened

6:38

to that clip over and over? How many

6:40

times?

6:41

>> How many times? How many people have you

6:43

shared that with? How many? Play the

6:44

clip. Play the clip.

6:45

>> I want to also uh

6:47

>> Oh, no.

6:48

>> Say hello and thank to Jamath and his

6:51

wonderful wife Nat. Thank you very much

6:53

for being here. Thank you very much. It

6:55

was great seeing you again.

6:58

Great couple. David Friedberg and uh

7:03

even as we know Jason Gall.

7:08

>> I say even. Thank you, Jason.

7:10

[Applause]

7:12

appreciation. I appreciate that.

7:14

>> Yeah, he's a good person.

7:16

>> I mean, he's a good person.

7:16

>> He called you a good person.

7:17

>> He called you a good person. So, here we

7:19

are.

7:20

>> What president What president's ever

7:21

called you a good person this? Come on.

7:22

>> I mean, it's it's obviously like surreal

7:25

>> for all of us, I think, to be this close

7:28

to the administration and then for Sachs

7:30

to be part of it. What I will say is you

7:32

have to give a lot of credit to this

7:34

administration for the velocity they're

7:37

going, what they're accomplishing. even

7:39

if you disagree with certain items on

7:40

the margins and their ability to engage

7:44

with leaders doing important work. And

7:47

if we compare that to Biden and Kamla,

7:50

like they weren't even letting people

7:52

come to the White House.

7:53

>> Is it Is this like

7:54

>> I love this administration. I love the

7:56

administration. I I like Trump. This is

7:58

a cabinet of CEOs.

8:00

>> Let me just say this. I'm not in love

8:01

with Trump. I'm in like with Trump.

8:04

That's where I'm at. I'm not in love

8:06

with Trump. I'm in like with Trump. But

8:08

what better team has ever been put

8:10

together? It is a cabinet of CEOs. It is

8:12

a cabinet of managers. It is a cabinet

8:14

of people who know how to get done. And

8:15

every time I go there, I'm impressed by

8:17

this cabinet. I pull my hair out when I

8:19

meet.

8:20

>> You admit that you're proTrump finally

8:22

Friedberg. You've been splitting it.

8:23

You've been dancing around the issue.

8:25

Are you full 100% in support of Trump?

8:28

You want to sit here and put me on the

8:29

spot? I put you on the spot.

8:30

>> I support my president. I support the

8:32

president.

8:32

>> Okay. So, you voted for him and you love

8:34

Trump. You voted for him and you love

8:36

Trump.

8:37

I love what he's doing

8:38

>> and you voted for him

8:40

>> and I have issues with the spending and

8:42

that's not been resolved. So like I said

8:43

before,

8:44

>> okay, great. Here we are folks.

8:45

>> My fullthroated endorsement will come

8:47

around when Doge actions are taken

8:49

seriously andor the White House puts

8:51

pressure on Congress to take action on

8:53

spending the budget.

8:54

>> What is everybody's favorite moment?

8:56

Favorite other than Trump, you know,

8:59

being absolutely amazing, great speech.

9:02

He's he's hilarious. Whatever. We'll put

9:04

pus outside that because it's hard to

9:05

compete with the president of the United

9:07

States. Sax, did you have a couple of

9:08

favorite moments? Give us a couple

9:09

favorite moments.

9:10

>> First of all, I think we should talk

9:11

about the substance of the speech

9:14

>> because I think this was the first

9:15

speech that President Trump has given on

9:17

AI since the AI boom began. He's he's

9:20

spoken about it before, but this was a

9:21

full-length policy speech and he

9:24

declared that the United States was in

9:25

an AI race. It's a global competition.

9:29

I think the the language that he used

9:31

was reminiscent of how President John F.

9:33

Kennedy declared that America was in a

9:36

space race. And in a similar way,

9:38

President Trump declared that we have to

9:39

win the AI race. I think you can argue

9:41

that the AI race is more important than

9:42

the space race. It's going to reshape

9:45

the global economy. It's going to

9:46

determine who the superpowers are of the

9:48

21st century. And President Trump was

9:51

really clear that we had to win it and

9:53

that he was going to support a strategy

9:56

for winning it. And then he laid out

9:57

what some of those key pillars are.

10:00

Number one was was innovation. We have

10:02

to get the red tape out of the way and

10:03

let our geniuses cook and clearly was

10:05

very supportive to a lot of the CEOs and

10:08

entrepreneurs in the crowd. Number two

10:11

is infrastructure. He touted the

10:13

hundreds of billions of dollars of

10:14

investments in energy and power

10:17

generation and grid upgrades and data

10:19

centers that he's supporting. And then

10:21

he also supported AI exports. He said

10:24

that we have to make America's tech

10:25

stack the global standard. So I think

10:27

those were really important messages.

10:29

And then on top of that, I think there

10:31

was also some parts of the speech that

10:34

maybe have gotten less attention but are

10:36

also important where he said that it's

10:39

not only important that we win. He said

10:42

it's important how we win. And he sort

10:45

of mentioned three non-negotiables here.

10:47

Number one was that American workers

10:49

have to be at the center of the

10:50

prosperity that we create.

10:53

Number two is that the AI models that

10:57

the government procures and buys must be

11:00

free of ideological bias. So no woke AI.

11:03

And he also signed an executive order to

11:05

prohibit woke AI in the federal

11:07

government. We could talk about that in

11:08

a second. That probably was my favorite

11:10

moment.

11:10

>> That was your favorite moment.

11:11

>> That was my favorite moment.

11:13

>> The red meat moment. I thought that was

11:14

>> that was the red meat. Yeah.

11:15

>> That was the red meat for the base.

11:16

Yeah.

11:16

>> Yeah. The third thing is he he did say

11:18

that we do want to prevent our

11:19

technologies from being misused or

11:21

stolen by malicious actors. And look, we

11:23

we are going to monitor for emerging and

11:25

unforeseen risk. So, you know, we're not

11:27

going to disregard the risk. But he had

11:30

this really good line in the speech

11:31

about how even though AI, look, it's

11:33

it's a daunting technology because it's

11:35

so powerful and like any revolutionary

11:38

technology like that, it can be used for

11:41

bad as well as good. But the the

11:43

daunting nature of it is all the more

11:45

reason why we have to do it in the

11:47

United States. Why would the United

11:48

States has to be the pioneer and the

11:50

leader is cuz we don't want the power of

11:52

that technology being developed in other

11:54

parts of the world at least other parts

11:57

of the world are going to have it but we

11:58

want to be the ones on the cutting edge

11:59

who are defining it and leading it.

12:01

>> Fantastic. Okay.

12:02

>> So I think it was a it was a really

12:03

important speech. I think this idea of

12:05

an AI race that is similar to the space

12:08

race, I think is going to be the

12:10

dominant frame on AI policy for years to

12:14

come.

12:14

>> Well, it's pretty clear, you know, this

12:16

presidency, this term is going to be

12:18

earmarked, I think, by four key

12:21

initiatives. AI, crypto, immigration,

12:25

and tariffs. I think that feels like

12:27

what they're locking into as what's

12:29

important for the next three and a half

12:32

years. I think you would agree with

12:33

that. And it's just great that you're

12:35

spearheading and helping the president

12:37

with two of those four. And just I the

12:39

velocity to me is what's super

12:41

impressive. Any way you could take us

12:43

behind the the scenes of how this stuff

12:45

is getting done so quickly. It feels

12:48

like there's some operational

12:53

cadence here that we didn't see in his

12:55

first term. Certainly didn't see in the

12:56

Biden term, but there's a there's a

12:59

cadence here that's different. Yeah.

13:00

Startup speed. H how is that?

13:02

>> Well, yeah, we call it he's working at

13:04

tech speed. I just think that the

13:05

president's constantly working. I mean,

13:07

he's just so energetic. I mean, he

13:09

basically works like two full work days.

13:11

I think it's well known that he doesn't

13:13

need a lot of sleep and he's continues

13:16

to work late into the night. And I just

13:18

think his energy propels everything

13:20

forward. I also think that there's a

13:23

very cohesive team at the White House

13:27

>> under the chief of staff, Susie Wilds. I

13:29

think it's very important. I think she

13:31

runs a tight ship and then you've got

13:32

the deputy chief of staffs under her and

13:35

I think most of these people have been

13:36

working together for a long time and

13:38

it's a team that works well together and

13:42

I just feels very coherent and cohesive

13:45

to me.

13:46

>> So I think it's a very effective team.

13:48

>> It does feel like that. The pace is

13:50

great. It means you're going to get more

13:52

shots on goal and you'll be able to try

13:53

more things and and get more

13:54

accomplished just like we see in

13:55

startups. Chimamoth, outside of the

13:58

president's talk, we'll go around the

13:59

horn here. Top two or three moments from

14:01

the discussions, just lightning round

14:03

here, rapid fire. What do you got? Top

14:05

two or three moments for you, Chamath,

14:07

just in the discussions that were

14:08

enlightening to you, inspiring to you,

14:10

notable to you.

14:11

>> I came out of it very motivated. I think

14:14

that the combination of the speech, the

14:16

executive orders, and the clarity of the

14:19

big beautiful bill

14:22

now give those of us that are in these

14:24

markets a ton of runway to go and

14:27

execute. And so those things reinforced

14:32

by the various members of the cabinet I

14:34

think were very important. That was one.

14:37

And then the second thing were the

14:43

market commentary from both Lisa Sue and

14:45

Jensen I thought was really valuable.

14:48

>> And then the third was Chris Wright and

14:53

Doug Bergam talking about energy. And I

14:56

tweeted this yesterday, but we are sort

14:59

of back to basics almost in a sense

15:01

where in the absence of power, I think

15:04

AI is is not going to be the thing that

15:06

we think it can be. So that's going to

15:08

create an enormous amount of appetite by

15:10

the federal government to do deals and

15:12

get players on the field. That's to me

15:15

very exciting. So

15:17

>> yeah, I came away really

15:20

riskon, I guess, is best way to say it.

15:22

>> I love it. Freebridge, you have two or

15:24

three moments outside of the president's

15:26

speech. Obviously, that's the pinnacle

15:28

there. So, let's just go below the

15:30

pinnacle. What were the other two or

15:31

three moments for you that were salient,

15:33

inspiring, notable?

15:35

>> I thought Jensen did a great job. I

15:37

don't know what you guys thought, but he

15:39

is very compelling and has a incredible

15:42

uh vision and view on where AI is taking

15:45

us, where it's headed, and what the

15:46

challenges are. So, I really appreciated

15:49

him taking the time to come and join us.

15:51

last minute he rearranged his schedule

15:52

to come out for it and it was great. By

15:55

the way, on the point on energy, which I

15:57

still think is the biggest unsolved

16:00

issue right now in America besides the

16:02

uh the federal deficit and the debt

16:04

problem,

16:06

Chris Wright agreed to rearrange his

16:08

schedule to come and join us at the

16:09

all-in summit in September. Oh, great.

16:11

To continue the conversation. We didn't

16:12

get enough time to talk about it. So, we

16:14

are going to hear more from Chris

16:15

particularly with a particular focus,

16:17

which is what I wanted to spend time on.

16:18

didn't get a chance last week on nuclear

16:20

and where are we cuz he actually is very

16:22

passionate like he said at the thing

16:23

it's where he's spending most of his

16:25

time right now and I think it's very

16:26

good to hear the deep dive on where we

16:29

are in the cycle on trying to accelerate

16:32

nuclear energy deployment in the United

16:34

States same question to you after pus

16:37

you got two or three moments that stood

16:39

out

16:39

>> let's just talk about the executive

16:40

orders for a second because I think it's

16:42

pretty cool that the president of the

16:44

United States signed three executive

16:45

orders at the all-in summit that we just

16:47

hosted I And that was pretty amazing.

16:49

One of them was to promote AI exports

16:53

because we want the American tech stack

16:55

to become the global standard. The

16:56

second was around AI infrastructure to

16:58

make permitting easier so that we can

17:01

help solve those energy problems you're

17:03

talking about Freeberg. And then the

17:05

third one was on preventing woke AI in

17:07

the federal government. And that to me

17:09

is probably my personal favorite because

17:12

we spent a couple years on the show

17:14

talking about how when we're talking

17:16

about woke, you're really talking about

17:17

censorship, right? We were talking about

17:19

censoring people's views based on

17:22

ideological bias, ideological dogmas. We

17:24

saw what was happening

17:26

>> in social media before Elon bought X

17:28

that helped

17:30

>> bring things back. But we were on a

17:33

track, I think, before President Trump's

17:34

election to repeat that whole social

17:37

media censorship apparatus in the form

17:39

of AI bias or AI censorship. And we saw

17:43

this with the whole black George

17:45

Washington and where some AI models were

17:48

saying it was worse to misgender someone

17:50

than to have a global thermonuclear war.

17:53

>> Yeah. And this wasn't an accident

17:55

because if you go back to the Biden

17:56

executive order on AI, there was

17:58

something like 20 pages of language on

18:00

there encouraging DEI values to be

18:03

infused into AI models.

18:06

>> So again, we were on track to repeat all

18:09

the social media censorship, all the

18:11

trust and safety stuff in this new world

18:14

of AI, but it would have been even more

18:15

insidious because at least when someone

18:18

gets censored, you kind of find out

18:19

about it.

18:20

>> It's explicit. It's not.

18:21

>> It's explicit. But with AI, it would

18:23

have been worse because you wouldn't

18:25

have even known. It would just be there

18:27

rewriting history in real time to serve

18:29

a current political agenda. It would

18:31

have been brainwashing our kids.

18:32

>> Oh, and people trust these AIs more than

18:34

they should. I mean, these things are

18:36

making a prediction of the next word

18:38

coming. This is not like God-given truth

18:41

here. And so, Freeberg, you wanted to

18:44

interject about this one because this is

18:46

actually I'll be honest, Saxs, I'm

18:47

surprised you're saying this was the

18:48

most important one to you. I like that

18:50

you clarified it because it was the one

18:51

that was mocked or kind of like people

18:54

were like what why is this important? I

18:56

think you made a good case for why it's

18:57

important. Freeberg your response. Yeah,

18:59

but sex this is not about broadly making

19:03

quote AI non ideological. Private

19:06

companies should still have the right

19:08

through freedom of speech or freedom of

19:10

expression or freedom to operate to make

19:13

AI that does whatever they want it to

19:14

do. What the EO was was that the federal

19:16

government would not procure

19:17

ideologically biased AI. Is that

19:19

correct?

19:19

>> Yes. Exactly. No, we're we're aware

19:21

>> just to make sure that the the federal

19:23

government is not trying to instruct

19:24

private companies how to operate. It's

19:26

simply saying if you want to sell to us,

19:28

this these are the rules of the road.

19:30

>> Yes, that's true. So, we were very

19:32

careful about the First Amendment

19:34

issues. And you're right that if a

19:36

private company wants to put out a

19:37

biased AI product, we're not going to

19:40

tell people they can't use it.

19:41

>> And it could work. It could be

19:42

successful. People might like it. Ya.

19:44

>> Yeah. We're just saying that the federal

19:46

government is not going to spend

19:47

taxpayer money

19:49

>> buying AI models that have compromised

19:52

their accuracy and quality because

19:54

they're beholden to some, you know,

19:56

ideological agenda,

19:58

>> which is similar to the approach with

20:00

the universities, right? Hey, listen,

20:01

you could have a biased university.

20:03

We're just not going to fund it. We're

20:04

not participating. I think it's it's

20:06

quite reasonable in that way. And

20:08

>> yeah, and I would just say that, you

20:09

know, we were a lot more careful about

20:11

this than the Biden administration was

20:12

when they required that DEI be inserted

20:15

into all these models. They didn't

20:16

distinguish between public and private

20:19

money or government procurement versus

20:21

private models. So they just they were

20:23

trying to suffuse DEI into everything.

20:25

And what we're looking for here is just

20:28

neutrality, right? We're looking for a

20:30

lack of ideological bias. The first step

20:32

was to get rid of that Biden EO, which

20:34

the president did his first week in

20:35

office. This goes a little bit further

20:37

and it's a little bit of a shot across

20:38

the bow of these Silicon Valley

20:41

companies saying, "Look, you need to

20:42

play it straight. You need to be

20:44

ideologically unbiased."

20:45

>> As the default,

20:46

>> as the default, when you sell to the

20:48

government, you can't insert your values

20:50

at the expense of accuracy. Look, at the

20:52

end of the day, accuracy and truth

20:54

seeeking is the standard, right? You can

20:57

measure. That's the goal.

20:59

>> So, we don't want the quality, accuracy,

21:02

and truth seeeking to be sacrificed

21:04

because of these any are you still are

21:07

you still seeing that are like when you

21:08

say these Silicon Valley companies I

21:10

mean is this still kind of a widespread

21:13

concern or widespread deployment from

21:16

your point of view where you're sitting

21:17

like are you still seeing a lot of the

21:18

models being trained on ideological

21:20

systems that you know are preferential

21:23

to one group and not to another

21:24

>> I think it was a much bigger concern 6

21:26

months ago and I think there's been such

21:27

a huge vibe shift since President

21:29

Trump's election and taking office that

21:31

like the woke stuff is sort of going

21:33

away on its own but

21:34

>> I and I I think that's the trajectory

21:36

we're headed. But it was

21:37

>> But you still think it's important

21:38

enough to make sure that there's any

21:39

error?

21:40

>> Yeah. It's like, look, this is make sure

21:41

this thing doesn't come back from the

21:43

dead. I think that there's been a huge

21:44

vibe shift since President Trump's

21:46

election and woke has definitely fallen

21:49

out of favor and it seems to be going

21:51

away on its own. But we could still get,

21:54

you know, Orwellian outcomes with AI.

21:57

And I do think it's very important to

21:58

just keep underscoring that what AI

22:01

models should be focused on is the truth

22:04

is on accuracy. And we don't want

22:06

ideological agendas to sacrifice that.

22:09

And um and I think I think that even

22:11

though this is a less salient issue now

22:14

than 6 months ago, precisely because of

22:16

the vibe shift, I still think it's

22:17

important to underscore this point that

22:19

we don't want

22:20

>> would you go so far

22:21

>> we don't want AI taking an Orwellian

22:23

direction.

22:24

>> Yeah. Would you go so far as to limit

22:26

free speech and and make it

22:29

non ideologically biased? Like would you

22:31

make that law if you could

22:33

>> because again the the decision about the

22:35

federal government procuring versus what

22:37

these private companies can choose to

22:39

reflect as their quote values in their

22:41

systems.

22:42

>> No, he just you already answered that he

22:43

would not.

22:44

>> Yeah. No, look, I we understand the

22:45

difference between public procurement

22:47

and private speech and again in a way

22:49

that the Biden administration did not

22:51

because they were saying that all AI

22:52

models

22:53

>> Yes. had to be adhere to a specific

22:55

ideology

22:56

>> to the CI stuff. So

22:57

>> yes, it was an ideology they wanted

22:59

embedded in it. You're saying don't put

23:00

an ideology in. But just to be clear

23:02

here, I want to make one point. This is

23:04

the defaults. Anybody who wants to could

23:07

when they start their prompt or they set

23:08

up their preferred language model could

23:10

say, "I'm an atheist. Here's what I

23:12

believe. Please speak to me with this in

23:14

mind." Or, "I'm a Catholic. Uh, you

23:17

know, I'm a Protestant." Whatever you

23:18

want. Here's my belief system. Please

23:22

never reference, you know, these three

23:24

subject matters in this way. So this is

23:26

the default. I think it's a great thing

23:28

to

23:29

>> I think that's a great example, JCAL. I

23:30

do think we'll end up seeing religious

23:32

AI. I think we'll see AI that's tuned to

23:34

people's religious

23:36

ideological, but I think yeah,

23:37

>> I have one of the startups we did was

23:39

doing a learning app and they were

23:41

struggling and they just made a prayer

23:42

app and their prayer app went parabolic

23:46

and now they're just like printing

23:47

money. So there is definitely a a huge

23:50

market here.

23:51

>> Check out what what were your

23:52

highlights?

23:53

>> It was great to be, you know, included

23:56

in everything. So I appreciate that we

23:57

had uh No, I'm being dead serious.

23:59

>> Your invitation finally didn't get lost

24:01

in the mail.

24:02

>> No, but here's the thing. It I think

24:04

this could have been a nonall-in thing.

24:05

It could have just been, you know, you

24:08

could have done it and just invited who

24:09

you wanted to. So I like that it was

24:10

under the all-in umbrella and that we

24:12

didn't censor anything and we went right

24:13

at hard topics. I'm a moderate. I know

24:15

people want to make me into like a

24:17

stupid lib, but I am an independent

24:18

moderate. And there were moments in time

24:20

when we had great debate, too. This

24:22

wasn't just a love letter to the

24:24

administration. One of the great moments

24:26

was J. D. Vance. It was just great that

24:28

he wanted to come chop it up and just

24:30

hang with the besties. And he came out

24:33

and he went right at me. He was like,

24:34

"Hey, you treated me like a beep at the

24:37

thing. We had a big debate and you know,

24:39

he went right at me." And then I was

24:41

like, "Okay, it's on. Want to talk about

24:43

stuff?" And he said, "Yeah, let's get

24:44

into it." And that's what I love about

24:46

JD. JD to me seems like the politician

24:49

of the future. I know this is like the

24:51

Trump's administration, but

24:52

>> So, you like him?

24:53

>> No. No. I I'm in like with Trump. I'm in

24:56

love with JD because he's young. He's

24:58

opinionated and he likes to mix it up.

25:00

He's on Twitter all day long. He engages

25:03

people on Twitter. He engages people in

25:05

other groups. I'll leave it at that. And

25:07

uh we had a really, I think, honest

25:10

discussion about immigration. And we got

25:12

back to the highskilled immigration

25:14

question. That's the third rail for MAGA

25:16

and and for the country right now.

25:18

Immigration recruiting.

25:19

>> You mean you brought it up right off the

25:20

bat?

25:21

>> No. No. He said he wanted your hobby.

25:24

You brought it up. Your hobby horse very

25:27

I want to continue the debate and I

25:28

said, "Okay, let's continue the debate."

25:30

So, here we go. He was super spicy and

25:32

he made a great super spicy point that I

25:34

want to point out here on Amplify. if

25:36

companies are going to be laying people

25:38

off. And there was an incredible uh

25:40

chart that came out. It was in the

25:41

Financial Times and they showed male

25:43

college graduates versus

25:47

non-olgraduate males. And there was

25:49

usually a huge gap in unemployment

25:50

between those two. In other words, if

25:52

you had the college degree, you you had

25:54

a much better chance than the non-ol

25:56

male. And now those two things have

25:59

flipped or they're like neck and neck.

26:01

If you have a college degree, you have

26:02

no advantage as a man coming out in

26:05

this, you know, 20 to 27 year old range.

26:07

This is men. Women are actually doing

26:09

better. More women in college than men,

26:12

yada yada. But he's very attuned to

26:15

this. And he said he's got big concerns

26:17

right now. So this is again why I love

26:19

JD because JD is very tuned into the

26:22

fact that people are asking for more

26:24

H-1B visas and that typically is to save

26:27

money and supposed to be very skilled

26:28

people. But why is Microsoft laying off

26:30

9,000 people then asking for more, you

26:32

know, H-1B visas? This is a really

26:34

honest, truth-seeking question. And I

26:37

it's hard for this administration to

26:39

talk about this issue because I know you

26:40

got Steve Miller, Bannon, whatever

26:42

people all the way on one side who want

26:44

to deport 20, 30 million people, Tucker,

26:46

and you know, and then you have other

26:48

people who are more moderate. And I

26:49

thought that was like a really great

26:50

moment in time for America and for us as

26:53

a podcast to challenge and have a really

26:55

important discussion. And he made some

26:56

great points there. Number two, we had a

26:58

great debate, I think, about energy. Uh,

27:02

disagree.

27:03

>> You disagree. Okay.

27:05

>> I disagree with the because I think you

27:06

challenged him with I I think you

27:08

challenged him with things that were not

27:09

facts and not true. And I'm happy to

27:11

debate that with you offline. I think he

27:13

was caught off guard, but I think it was

27:14

pretty like

27:16

>> rough and inappropriate.

27:17

>> 100 If you think it's inappropriate,

27:18

that's fine. I

27:19

>> Jake's favorite moments were all the

27:21

ones where he got the ones where we had

27:22

debates.

27:23

>> That's what you're describing.

27:24

>> No, no. Where there were debates. got

27:27

with the vice president. You got into it

27:29

with the secretary of energy. Those your

27:31

favorite moments when you got to

27:33

>> Okay, great. So, okay, fine. I like when

27:35

there's a little conflict, a little

27:37

debate about an important issue. And

27:38

when I walk the audience, which was, you

27:41

know, 90%

27:43

Republican, GOP, MAGA, etc. People said

27:46

that was a great moment. I really like

27:48

that debate because he kept saying like

27:50

nonreliable energy and whatever. And I

27:51

said, are you talking about solar? And I

27:53

think there was a little misinformation.

27:54

>> Reliable. No, it's not.

27:56

>> You put it with a battery. Right now,

27:58

Texas is 30% some days wind and energy.

28:01

You know, like I can tell you I live in

28:02

the great state of Texas.

28:03

>> Texas is Texas roughly 5% solar. Just so

28:06

you know.

28:06

>> What's that?

28:08

>> Texas is roughly 5% solar,

28:10

>> right? And wind puts it up to 25 to 30%

28:12

on the top days is coming from that. My

28:14

point about that is and it's cheaper to

28:16

put in a solar uh and battery farm than

28:18

a new coal plant. It is 100%. We can

28:21

pull up the stats.

28:22

>> It is twice the cost to do solar than it

28:24

is to do nat gas. It takes 4,000 acres

28:26

whereas gas takes 20 acres.

28:28

>> I said coal.

28:30

>> Yeah, but these the big advocacy with

28:32

these guys is to use nat gas to use

28:34

methane. He was saying coal clean coal

28:36

clean coal. He said it 50 times.

28:38

>> These methane plants are half the cost

28:39

of solar. They can get stood up in less

28:41

than two years to generate a gigawatt.

28:44

And instead of being 4,000 acres of

28:46

solar, you can get it done for, you

28:48

know, call it 20 acres. Now talk a

28:50

little bit about pollution. And that's a

28:52

big part of why they're doing this.

28:53

Well, a big part of methane is that it's

28:55

actually cleaner than coal, which is why

28:57

they're using it. Cleaner than oilar.

28:59

>> Cleaner than oil and the two ways of

29:02

getting energy. Now, science guy, now do

29:07

>> I'm trying to give you the fact about

29:08

why about why it is cheaper and faster,

29:10

which is what he was making an advocacy

29:12

for, right? It's not about like solar.

29:14

Yes, you're right. It has a lower carbon

29:15

footprint when you're running it. But at

29:17

the end of the day, what these guys are

29:19

focused on and a big challenge for

29:20

America is how do we scale energy

29:22

production in the states and scaling

29:24

energy production, I personally think we

29:26

need to fix the regulatory roadblocks in

29:28

nuclear. And Chris Wright's been very

29:29

vocal on this.

29:29

>> We all agree on that.

29:31

>> But the fact is this NA gas supply that

29:33

we have in the United States and the

29:36

fact that we can deploy nat gas energy

29:38

production very quickly is what makes it

29:40

such a reliable source right now. if the

29:41

US wants to have a chance at scaling

29:43

from 1 terowatt to two faster than

29:46

projected today and that's the reason

29:48

you know it's not it's not about like

29:49

solar is being bad solar is bad like

29:51

that's not the the argument it's just

29:52

like dude we we got to get moving fast

29:54

and we got to have reliable energy

29:57

in our debates I just want to point in

29:59

our debates when there's bad faith

30:00

moments I think it's a bad faith moment

30:02

for when I say coal versus solar and

30:04

then you say no you're wrong it's solar

30:07

versus n gas and that's what he was

30:09

doing this is what politicians do here

30:11

all in. We like to do, you know, uh,

30:14

fact-based, truth first stuff, not

30:16

biased stuff. And so solar, you're

30:18

comparing. So, you know, solar and how

30:22

fast it is versus how fast it is to go

30:23

to net gas. Of course, it's faster to go

30:25

to net gas if we have those available.

30:27

Let's put that aside. It's an important

30:28

debate. The fact that you and I are

30:29

debating it is important. And I also

30:32

thought Lisa from AMD was fantastic. I

30:34

haven't heard from her. By the way, I

30:36

just want to point out that when I got

30:38

back to the conference, so I I left for

30:40

a time to go back to the White House and

30:42

then I came back.

30:43

>> The first thing everyone said to me

30:45

>> when I got back was,

30:47

>> "Did you see Jal being a jerk to Chris

30:50

Wright?" They were everyone was like,

30:52

"Oh, about the Yeah, a jerk.

30:54

>> He's a civil servant. He has to answer

30:57

hard questions. You didn't talk to him

30:58

in a in a in a in the way that you

31:00

would." Basically, everyone thought you

31:01

were a jerk to Chris, right? And you

31:03

were kind of a jerk to JD. And what are

31:05

your favorite moments?

31:06

>> Call me an

31:07

>> What are your favorite moments from the

31:09

conference? You're reminiscing about

31:10

>> that. You were an to me.

31:15

>> Anyway, the point is one thing you're

31:17

going to get here at the Allin.

31:20

This is where everyone%

31:22

two out of three were

31:23

>> everyone was saying this. You almost

31:24

derailed the whole thing.

31:26

>> Nobody derailed. You're a civil servant,

31:29

Mr. Sachs. You're a civil servant.

31:31

You're all civil servants.

31:32

>> I've been putting up with you for 5

31:33

years on this podcast.

31:34

>> The hard questions.

31:35

>> It was perfect training for government

31:37

services beyond the podcast being

31:39

interrupted by you for 5 years.

31:41

>> Yes.

31:42

>> That's why I'm so ready to dwell.

31:44

>> You learn about you work for us, all of

31:46

you. And you're all going to take hard

31:48

questions. And you're all going to take

31:49

hard questions on September 7th, 8th,

31:51

and 9th when we have the allin summit in

31:55

>> Los Angeles. By the way, by the way, one

31:57

thing I'll say is Chris Wright's chief

31:59

of staff came out to me afterwards and I

32:01

said, "Oh, I'm sorry. I heard Jake was a

32:02

jerk to to Secretary Wright and he's

32:05

like, "Oh, no. Chris loved it. He loves

32:07

mixing it up." Okay, of course.

32:08

>> And he's coming to he's coming to Olan

32:11

Summit in on September 8th. So,

32:12

>> can't wait to debate him more. Can't

32:14

wait to mix it up more.

32:15

>> So, he he likes mixing it up. So, kudos

32:17

to him.

32:18

>> Okay.

32:18

>> And so did so did JD Vance, the vice

32:22

president to you. Stop calling him JD,

32:24

by the way. Well, I mean, listen, I just

32:26

want to say Vice President J. D. Vance

32:29

and I have been directly communicating.

32:31

We had a We Yes.

32:33

>> No, you haven't.

32:34

>> David Sachs, your worst nightmare.

32:36

>> Oh my god. Your worst nightmare.

32:38

>> The nation is ruined. What the

32:40

>> We let Jake Allen to Washington and now

32:42

look what's happening.

32:43

>> Yeah.

32:44

>> And listen, I want to level set with

32:46

everybody.

32:47

>> We I am going to ask whatever

32:50

question I want to whatever guest we

32:51

have and nobody's stopping me. The only

32:54

way you're going to stop me is by

32:55

writing me a huge check to buy

32:57

me out of this podcast and replacing me

32:59

with some mid until

33:00

>> or if the Secret Service keeps you off

33:02

stage, which might be an option.

33:03

>> Or Secret Service keeps up stage. But

33:04

the truth is, this is one of the great

33:06

things about this administration sex is

33:08

that they love to mix it up. They like

33:10

great debate. You know who didn't like

33:13

great debate and ran from it? Kalama

33:15

Ding-Dong. She wouldn't even come on

33:17

this podcast. You know who doesn't like

33:19

debate? Tik Tok.

33:20

>> We get that at Bernie's Biden who didn't

33:23

even know what a podcast is.

33:25

>> Tim Waltz who doesn't own

33:27

>> you. You definitely you definitely have

33:29

your moments, bro. You definitely have.

33:30

>> But Tim Waltz doesn't own an equity. He

33:33

literally doesn't own one share of any

33:34

company, doesn't his home

33:36

>> and Tim Waltz is on there giving a hard

33:38

time about the uh Trump savings

33:40

accounts.

33:42

>> I mean, I don't even know if that's a

33:44

hat though, which you loved. You thought

33:46

that was going to win the election. I

33:47

thought he might be able to speak to

33:49

like the middle of America and then I

33:51

find out like

33:52

>> when they do the the Deep Oppo research

33:54

that the guy doesn't own one stock.

33:57

>> The guy doesn't own his home. He's

33:59

financially illiterate and we're making

34:01

him

34:01

>> employed by the government. He's been

34:03

employed by the government his whole

34:04

life.

34:05

>> I mean, have you

34:07

There it is.

34:08

>> That's what Jake thought would win the

34:10

election.

34:11

>> You're never going to live that down. I

34:12

remember when you tweeted you thought

34:13

that was it. You thought that was the

34:15

master stroke. I thought it might

34:17

>> master stroke that was going to win them

34:19

the election.

34:19

>> Hey, listen. No straanis does not bat a

34:21

thousand. No, no, even no straanis

34:25

cannot bat a thousand. But it did come

34:26

out, by the way, that Nancy Pelosi

34:29

wanted to do the speedrun primary. I

34:30

don't know if you saw that, just not to

34:31

rehash too much stuff. Sax, I want to um

34:34

say there was one point of difference if

34:35

you want to get into it around the the

34:37

the content part of part of it where and

34:40

this is something that

34:42

>> the press was having a field day with

34:44

and they really keyed on which was hey

34:47

respecting IP, respecting copyright.

34:50

What's the feedback been so far on that

34:52

which was a pretty spicy part of

34:54

President Trump's speech?

34:57

>> Well, I think what the president said

34:59

was just very pragmatic. He said we had

35:01

to have a common sense approach towards

35:03

intellectual property. And he said if

35:05

you have to make a deal with every

35:07

single article on the internet, every

35:10

single website, every single book, every

35:12

piece of IP in order to train an AI

35:15

model, it wasn't feasible. He said,

35:17

"Look, I appreciate the work that went

35:20

into people creating these works, but

35:22

you're not going to be able to negotiate

35:24

a deal for every single one of them. And

35:26

if we require our AI models to do that

35:29

and China doesn't and they won't.

35:31

They're just training on everything

35:32

whether it's you know pirated or not

35:34

then we're going to lose AI race. So I

35:36

think he took the side of a fair use

35:38

definition. I don't know if he used the

35:39

term fair use but effectively he was

35:41

taking the side of a reasonable fair

35:44

use.

35:46

>> What did you think of that part Dave

35:47

Freeberg? You have any thoughts for

35:48

Jimoth on that part?

35:49

>> I think he's absolutely right. I've said

35:50

this before. If something's in the

35:52

internet, if something's in the open

35:53

domain, and I strongly disagree with the

35:57

idea that AI getting trained is the same

36:00

as AI replicating copyright material. If

36:03

AI outputs text or outputs audio or

36:07

outputs video that contains copyright

36:10

material, it is 100% in violation of

36:12

copyright.

36:13

>> And he said that, by the way,

36:14

>> yes. And if the AI is learning, it is

36:16

understanding patterns, it is

36:18

understanding reasoning, it is

36:19

understanding concepts by reading

36:21

copyright material just like humans do.

36:24

A writer, an author reads a bunch of

36:26

fiction, learns good techniques, learns

36:29

good concepts, learns good theory from

36:32

reading all those books and then goes

36:33

and writes his or her own book. They are

36:36

not violating copyright material. In the

36:37

same way,

36:39

Freeberg, what if it's all the New York

36:40

Times on the open internet?

36:43

>> 100%. You're you're 100% correct that

36:45

should be paid for or licensed. I'm

36:47

talking about the open internet. I'm

36:49

talking about open material. I'm talking

36:51

about stuff that's in the open domain

36:52

which

36:52

>> common crawl.

36:53

>> There's a thing called common crawl.

36:55

>> If there was if somebody stole a hundred

36:59

books, let's say, and put them on their

37:01

website and it was a pirated Russian

37:03

website with a thousand books on it and

37:04

you accidentally crawled it, you would

37:06

be obligated to take that out then. I

37:08

think we all agree.

37:09

>> Correct.

37:09

>> Okay. Correct.

37:10

>> Cuz that's what a lot of the lawsuits

37:11

around. So, I think we're reaching

37:12

something. I just want to say, you know,

37:13

this is such an important point,

37:15

especially to me as a content creator

37:16

and somebody who spent his career in

37:17

this. I've been thinking about the

37:19

endgame and um I was I'm here in Park

37:22

City. I was just giving a a keynote and

37:24

I wanted to show you something I made

37:27

saxs because I think we have to get to

37:28

the the endgame here. So, in my talk, I

37:33

talked a little bit about how can we get

37:35

through this fight and then maybe

37:37

getting to a solution. So, I had my team

37:39

mock up

37:41

the New York Times website here and chat

37:44

GPT doing a deal with them. So, here you

37:46

see you're on the New York Times website

37:48

and you ask it a question powered by

37:49

GPT. You ask it, hey, you might ask this

37:53

question. In fact, you log in with your

37:54

chat GPT credentials. You're and it

37:56

could be Gro, it could be Gemini. Give

37:58

me the earliest mentions of Putin, you

37:59

know, if you were a fan of Putin or

38:01

something. And it would then go through

38:02

that and give you your your Putin

38:04

references. And then I made another one.

38:06

And then obviously this would be an

38:07

exclusive to Chat GPT. It would be one

38:09

of those things where you know they get

38:11

an exclusive. And then here on the uh

38:13

Disney Plus channel, imagine you could

38:15

make yourself into a Jedi Knight and you

38:17

could then upload your photo. You know,

38:19

kids might really get into this. You

38:21

upload your photo, you can make You

38:22

talked about this free a couple of times

38:24

of the future of narrative storytelling.

38:25

You up your photo and then it makes you

38:27

into a Jedi Knight. There's there's

38:29

Darth Calacanis. So

38:32

>> that looks to me like you're infringing

38:33

on their trademark. What's that?

38:36

>> Are you infringing on their their

38:37

copyright?

38:38

>> This is fair use. This is fair use. This

38:40

is a perfect example of fair use for

38:41

editorial.

38:42

>> You're also infringing on some Osmpic.

38:43

That's absolutely infringing.

38:46

>> Trust me, I am definitely infringing on

38:48

some OMIC here, guys. I'm I'm past

38:50

those. I'm on to peptides now, man. I'm

38:52

on the Wolverine protocol. So, look at

38:54

>> Are you

38:54

>> Yeah, I started doing the I mean I I

38:57

don't What could go wrong?

38:59

>> Don't take a podcaster's advice. Please

39:03

don't take a podcaster's advice on your

39:06

healthcare rule number one. Take

39:08

Chimat's advice because he's got 6% body

39:10

fat, which I think attributes to much of

39:13

your pomp and circumstance around your

39:15

privates. I think it has to do with the

39:17

lack of fat. But I'm going to leave it

39:18

at that.

39:19

>> First of all, it's 11 and a half, but

39:21

you know, it that's that's like right

39:23

that's like right before I go on summer

39:25

vacation. Then it then it ends up at 12

39:27

or 13.

39:27

>> Did you go get that? Did you go get that

39:29

gelato? What was that place we went that

39:30

we love? Lulu me. I've gone there every

39:33

day. Every day so far.

39:34

>> Did you do two or one? Be honest. Two or

39:36

one?

39:36

>> Bro, I've had I've had I've been doing

39:38

>> No, no. Per session. Do you do two or

39:40

one? Be honest.

39:41

>> Per session, too. I start with the

39:42

medium and then and I finish with a

39:43

small.

39:44

>> Yeah, exactly. You This stuff is so

39:46

good.

39:47

>> I've never tasted any gelato like this.

39:49

It's incredible.

39:49

>> I mean, it's unbelievable. We have to

39:50

license it for the United States and the

39:52

all-in brand. We have to license it from

39:53

them.

39:54

>> It's really incredible. But Chamath just

39:55

generally speaking or anybody who wants

39:56

to have at it. Friedberg Sachs, what do

39:59

we think about the endgame here? Because

40:01

there's some major lawsuits here.

40:02

They're going to get settled in the next

40:04

year or two. What What do we think about

40:06

sort of the future I've shown here

40:08

today?

40:09

>> I think what Sax just highlighted is

40:11

exactly right.

40:12

>> Look, we got to have a common sense

40:13

approach here or we're going to lose the

40:15

AI race. I mean, one of the out one of

40:18

the key determinants

40:20

>> of AI quality is the amount of data that

40:22

you have. It's very simple, right? It's

40:24

there's a few building blocks. There's

40:25

energy, there's chips, and there's data,

40:28

and there's algorithms. And if you lose

40:30

on any one of those dimensions, then

40:32

you're in trouble, right?

40:33

>> So, look, you just can't have a

40:35

situation where China can train on the

40:37

entire internet. And RAI models are

40:40

hamstrung by needing to contract,

40:43

>> negotiate contracts with every single

40:45

website.

40:45

>> But right now, Europe, Elon owns X,

40:47

right? He owns Twitter for now X. Does

40:50

Sam Alman have the right to use X in his

40:53

corpus?

40:54

>> It's public.

40:56

>> No, it's not. No, it is not public

40:58

endpoint.

40:58

>> It's not a public.

40:59

>> I just honestly I don't

41:01

>> there's I don't know the answer to that.

41:02

There's some edge cases here. We're

41:03

going to have to come up with

41:04

>> It's not about whether it's behind a

41:06

payw wall or not. It's whether these

41:07

APIs exist and whether you're you're

41:09

actually contractually allowed to use

41:11

them or not.

41:12

>> The terms of service. Correct.

41:13

>> The terms of service. It's published on

41:15

every website what the terms of service

41:16

are with respect to the content. I think

41:17

it would be okay to let people opt out,

41:21

you know. So, we already have this with

41:22

Common Crawl. You can put in the footer

41:24

of the website, you put in robots.txt

41:27

and you opt out of Common Crawl.

41:29

>> Common Crawl is like this nonprofit

41:31

organization that basically archives the

41:33

entire web every few months.

41:35

>> Funded by Gil Elbaz,

41:37

>> former Yeah.

41:39

>> formerly of Google. Great fan of the

41:40

pod. Shows up to our summits. Great guy.

41:43

>> And all of Open AI was built off of

41:45

Common Crawl originally.

41:47

And he but they're very clear by the

41:48

way. They say you have to clear

41:50

copyrights. You don't get to just use

41:51

open Chrome.

41:52

>> Can I go out on a limb? I don't know if

41:54

you guys saw this Amazon deal with the

41:57

New York Times for $25 million. Did you

41:59

see that today?

42:00

>> No,

42:00

>> I didn't see today. Explain it, please.

42:02

>> I think that the New York Times licensed

42:05

Amazon all of their content, including

42:07

The Athletic and a bunch of other things

42:09

for training. 20 million. Sorry, 20

42:12

million a year.

42:13

>> Okay, here we I read that and I thought

42:16

this is the peak of these deals. These

42:18

deals will only go down in terms of

42:22

dollar value from here.

42:24

And it it actually brought me to this

42:27

point where I was thinking to myself, is

42:29

it even realistic

42:32

to believe that patents and copyrights

42:36

actually exist in 5 years? And I went

42:39

through this exercise of like if a

42:41

computer studies the

42:45

periodic table and also understands the

42:48

laws of physics, the laws of biology,

42:49

the laws of chemistry and then

42:50

independently deres

42:53

some material that is otherwise

42:54

patented,

42:56

what will happen?

42:59

And then separately, if two competing

43:01

AIs invent a new material from scratch,

43:05

how will the international courts deal

43:08

with this? And if you take all of these

43:10

examples to the limit, at the limit, the

43:13

idea that there are copyrights,

43:16

enforcable copyrights,

43:18

I think is a very fragile assumption. So

43:22

I'm actually thinking more that we have

43:24

to spend some time understanding the

43:27

landscape of a world that doesn't have

43:30

copyrights and patent protections and

43:32

instead what is the surface area in

43:35

which you compete what is trade secret

43:37

what does that mean in a world of AI and

43:39

I think it's quite an interesting thing

43:41

to think about

43:43

patents are a totally different piece I

43:45

think that's a fascinating string to

43:46

pull on I will tell you I will take the

43:49

other side of the bet if we want to make

43:50

a poly market on this. I will guarantee

43:52

that this will be the beginning of the

43:54

deals and the deals will go up from

43:55

here. I'll tell you why. The reason the

43:57

New York Times made that deal is to make

44:00

it apparent that what OpenAI has done

44:02

has damaged their business because now

44:04

they have a customer and their customer

44:06

is Jeff Bezos at Amazon and Jasse and

44:09

now they can show damages because and

44:11

now those damages could give them an

44:13

injunction against OpenAI and OpenAI's

44:16

got to take it out of their crawl of

44:18

their you know construct and that's

44:20

going to be really expensive for them.

44:21

It's not not doable, but it's going to

44:24

be expensive. And let's think on a

44:27

societal basis of what we want as a

44:29

society. Do we want a society in which

44:31

journalists, writers, artists,

44:33

musicians, filmmakers, actors cannot

44:35

make a living, podcasters, or do we want

44:37

a world in which they can?

44:38

>> And I think technologists,

44:40

>> hold on, let me hold on. As a

44:43

technologist, as a technologist, we

44:45

typically think if we can crawl it, it's

44:47

ours.

44:47

>> What I can tell you as an artist is if I

44:49

make it, it's mine. and you need my

44:51

permission because it's my art and I

44:53

think it the industry will do better if

44:55

they respect them because now the New

44:57

York Times can hire more fact checkers.

44:59

>> But can I just ask you a question?

45:00

>> Yeah, go ahead. Sure.

45:01

>> But why do you have to connect the two

45:04

as immutable things? meaning why can't

45:06

somebody make something still you know

45:10

let's just say it's a song but that song

45:12

can now be made by multiple AI models

45:16

but if they make the song

45:19

there's a reasonable claim that even if

45:21

they don't have the copyright more

45:23

people will want them to perform the

45:25

song than some random AI

45:27

so can't you make a living without

45:29

having the copyright

45:31

>> which is the choice of the artist some

45:32

artists are were very well known for not

45:35

wanting their art to exist in some

45:36

mediums. As a perfect example, the

45:38

Rolling Stones for a long time thought

45:40

they would be sellouts if they had their

45:41

music used in commercials. And when they

45:43

did Start Me Up with Windows, that was a

45:45

really big concession from them. And

45:47

that's up to the artist to make that

45:48

decision. You make a a valid claim. Hey,

45:50

yeah, you go on tour and make more

45:51

money. But that's the artist decision,

45:52

not the technologists or the people

45:54

stealing their content. And by the way,

45:55

$20 million a year is a hundred $200,000

45:59

highly paid journalists, fact checkers

46:01

at the New York Times. They're going to

46:03

get 10 of those deals. And it's going to

46:06

create a golden era age of journalism

46:09

and content. And we should be happy.

46:11

>> I told you this example, Jason, but at

46:13

at Beast, we did a a licensing deal of

46:16

our content to allow OpenAI to learn

46:19

>> Yeah.

46:20

>> to run training runs on our videos. And

46:23

at the board, the thing that we kept

46:25

talking about was I was I was really

46:27

concerned like, let's just do a a couple

46:28

year deal max.

46:30

>> And the reason is we have no idea what

46:32

this looks like. in

46:34

>> five or 10 years. And there's just as

46:37

much chance to your point that we get it

46:39

wrong as right. Now that was about 6

46:40

months ago. And so the intuition that I

46:42

had back then was maybe we should keep

46:44

the deal term as short as possible. But

46:47

now when I see how important AI is in

46:50

the global landscape and what China is

46:51

doing, I think on the margins that this

46:54

idea that these copyrights will mean

46:56

something

46:58

>> in my mind, I am underwriting the value

47:03

of these things going to zero and I'm

47:04

asking myself instead for my businesses,

47:07

how are we actually building a real

47:09

defensible moat and not a piece of paper

47:12

that we can use to sue somebody?

47:14

>> Okay, Freeberg, you want the last word

47:15

here? We got to move on to some other

47:17

topics. I

47:17

>> I just want to be the last word.

47:19

>> I just want to be clear that nobody is

47:22

losing their copyright.

47:24

>> Copyright is the right not to have your

47:26

work copied. And if an AI model produces

47:29

outputs that copy or plagiarize your

47:32

work, then that's a violation of the

47:34

law.

47:35

>> And I think the president specifically

47:37

said that we're not allowing copying or

47:39

plagiarizing. The question is whether AI

47:42

models are allowed to do math on the

47:45

internet. You know,

47:46

>> pattern recognition pattern recognition.

47:47

>> Basically, that's what it is. And it's

47:49

and Jal, I think you're conflating the

47:51

two. And I I don't want to be

47:52

interrupted. I just want to say this.

47:54

>> I understand the distinction.

47:55

>> And and I think that this idea that like

47:57

I can't, for example, go to the library,

47:59

rent a book, read it, and then learn

48:01

some of the good techniques on how to

48:03

write a good book should be restricted

48:05

to humans in this AI context. Like this

48:07

is exactly what they're doing. They're

48:09

identifying patterns and then they're

48:11

building predictive algorithms that

48:12

allow them to output stuff that starts

48:14

to fit within different kind of, you

48:16

know, variable settings. Do you guys

48:18

think it's possible that if you

48:20

allocated enough compute

48:22

at the problem, you could write Michael

48:26

Kryton's Jurassic Park Denovo without

48:29

ever having read it?

48:31

>> Yeah, me too. Me too. I I think

48:33

>> I don't know what that would mean. Like,

48:35

>> well, this is my point. I know who

48:37

Michael is and I know what Jurassic Park

48:39

is. I don't know what it means

48:40

>> this issue.

48:41

>> I don't know what it means to say can AI

48:42

write that like

48:43

>> but you guys remember the Ed Sheeran

48:45

lawsuit. Do you remember the lawsuit?

48:47

>> I I did. But let me just make one point

48:48

here on this cuz you're saying I don't

48:50

understand it. I spent my career in it.

48:51

I understand it much better than you do

48:53

>> and I understand it from lawsuits and

48:54

being in the weeds on it. Like I

48:56

understand it from first principles

48:57

which you do not. And I will say this is

49:00

what we're talking about here is the

49:02

definition. It's the definition of a

49:04

derivative work. and the output matters.

49:07

So if you were to take my knowledge and

49:09

then create a derivative work from it

49:11

and you used a percentage of my work and

49:13

that's where this will get into the

49:15

nuance is what percentage of the

49:17

original work is used in the derivative

49:19

work and under what context a commercial

49:21

context or a non-commercial this is

49:23

clearly a commercial one if it's a if

49:25

openai was a nonprofit right now we'd be

49:27

having a distinctly different discussion

49:29

because it would there would be you

49:30

wouldn't be competing with me as the

49:32

copyright holder to use this new medium

49:34

and create the derivative works and it

49:36

has to change substantially. So if it's

49:38

a if it's a cliffnotes

49:40

>> when China has the only models that are

49:42

able to meet your stringent definitions

49:44

of copyright.

49:44

>> Well, no. Here's the thing. I think the

49:46

China fear the China fear is

49:48

I'll be totally honest here.

49:49

Just because China steals IP does not

49:52

mean you get to steal from Americans. In

49:55

America, we have rules. And when you go

49:57

to China, and by the way, we spent the

49:58

last 30 years. The major issue with

50:00

China is not Taiwan. It has been

50:03

the technology industry itself. Let me

50:05

finish. The technology industry itself

50:07

has leaned on our government for 30, 40

50:10

years, including Microsoft, including

50:11

Google, to make sure our trade secrets

50:13

are not stolen, our IP is not stolen,

50:15

our movies are not stolen. That is the

50:16

key issue with China. So just because

50:18

China's a thief does not mean American

50:21

companies get to

50:21

>> have you seen Have you seen the latest

50:23

batch of Chinese open source models or

50:26

open models?

50:27

>> They they steal everything. Does that

50:28

mean you should be able to steal

50:29

Windows? Should you be able to steal

50:32

Jason? We don't think it's stealing.

50:35

Elon has said this pretty clearly, but

50:37

Grock 5 and for sure Grock 6 will not

50:40

use Common Crawl. It will not use the

50:42

internet. Okay? It'll just be an

50:44

enormous amount of synthetic data. And

50:46

back to what Freeberg and I just agreed

50:49

upon, if you synthetically go and try to

50:51

generate all this content to learn

50:53

across, you're invariably going to

50:54

produce something that's already been

50:56

created.

50:58

>> That's like some sci-fi level.

51:00

>> I understand. That's what's happening

51:02

now. It's happening now.

51:04

>> If somebody

51:05

>> What do you think happens to Grock 5 or

51:07

Grock 6? Is that violating copyright? It

51:09

didn't even know that it existed

51:11

>> on the output. Yeah, that's fine. If it

51:13

on the output created a similar work,

51:17

they would need to then take it down.

51:19

And so that that would be a really

51:22

interesting new that's a new space we're

51:24

going to have to contend with. So you

51:25

can if it does happen is a new concept

51:29

that we would have to address in a new

51:30

way. I'll give you I'll give you a

51:32

science corner example. There's this

51:33

EVO2 model that they published at the

51:35

Arc Institute which Patrick Collison you

51:37

know is the name Sherman.

51:39

>> So that EVO2 model they just ingested

51:41

all the DNA data they could find in the

51:43

world. Trillions and trillions of base

51:45

pair of data that they ingested

51:47

>> and then they looked at patterns in DNA

51:50

and that's it. They had no context for

51:53

what the DNA represented. They had no

51:54

context for the concept of genes. none

51:56

of the structured understanding of what

51:58

that DNA does, what it is. And you know

52:00

what it did? It fed in the BA gene

52:03

variant and the thing output a warning

52:06

saying I think that this is a pathogenic

52:08

variant to DNA without having any

52:10

context. This is the the breast cancer

52:12

alil and it didn't have any knowledge

52:14

and it did it wasn't trained on that at

52:16

all. It had no knowledge that there are

52:18

pathogenic variants for cancer and it

52:20

identified that this was a genetic

52:21

variant that can cause some sort of

52:23

pathogenic outcome in the organism. So

52:25

that was that's a great example where

52:27

there's a lack of understanding at the

52:29

human level on what really drives some

52:32

of the patterns in nature, the patterns

52:33

in society, the patterns in behavior

52:35

that are kind of emergent phenomena

52:36

perhaps that these AI models are

52:38

starting to identify. And I think to

52:40

Jama's point, we may end up seeing this

52:41

in things like entertainment as well.

52:43

All right, this has been an amazing

52:44

debate. We got to move on. And you know

52:45

what? We're going to have more amazing

52:47

debates September 7th through 9th in Los

52:48

Angeles at the All-In Summit. The lineup

52:51

is stacked. Alibaba's co-founder Josai

52:54

Tomma Bravo co-founder Arc Invest Kathy

52:56

Wood Uber CEO Dra Sequoa's Rolaf Botha

53:00

YouTuber Cleo Abram and many many more

53:02

coming you get the last word here go

53:06

>> I was just highlighting this tweet that

53:08

I saw

53:10

>> where talking about Chinese openweight

53:12

models are basically open source models

53:14

so

53:15

>> basically all the leading American

53:17

models are closed source and all the

53:18

leading Chinese models are open source

53:19

this is kind of the way things played

53:21

out.

53:22

>> It's a pretty good technique for

53:24

catching up because then you got the

53:26

larger open source developer community

53:28

helping you out.

53:29

>> But the point is just that these open

53:31

source models are catching up pretty

53:33

fast. We're ahead in many other aspects.

53:35

Our chips are a lot better. Our data

53:37

centers are better and so on. And I'd

53:38

say our closed source models are better.

53:40

But they have this one area of open

53:41

source models. So again, if you

53:43

hamstring our AI models access to data

53:46

by creating a whole bunch of new

53:48

requirements for contract negotiations,

53:50

like we could really lose the AI race.

53:52

This is a really big deal. It's not a

53:54

madeup concern. I don't know why you

53:55

think it's made up.

53:56

>> I never said that it's made up. I think

53:59

it's an opportunity for America to

54:00

actually have a distinct advantage,

54:01

which is that $20 million from Amazon

54:04

alone is 1% of the New York Times

54:07

revenue. And that's going to go directly

54:08

to the bottom line. It's going to allow

54:09

them to hire more journalists. Then that

54:11

protected site will have be giving in

54:14

real time something. These language

54:15

models are going to have to go hack and

54:17

steal that real-time data is going to be

54:19

a distinct advantage for Gemini, OpenAI,

54:21

Amazon, whoever chooses to do it. And we

54:23

can create

54:24

>> you have this like nostalgic sort of

54:26

quasi romantic notions about like

54:29

journalism and the need to save the New

54:31

York Times.

54:32

>> It's also art. It's I mean you can say

54:34

all the derogatory things you want about

54:36

me personally, Saxs. That argument

54:38

doesn't work. No, no. You just said I

54:39

have this whole nostalgia whatever when

54:41

you

54:41

>> Yeah, you do. You're nostalgic for

54:42

journalism as it used to exist.

54:44

>> When I know I've beat you in the debate

54:45

is when you make it personal like that.

54:47

It's not personal. I'm not being

54:48

nostalgic. I'm trying to create a

54:50

sustainable a sustainable advantage for

54:52

America. And you are our public servant

54:55

and you're learning AI. You will take my

54:57

feedback.

54:59

>> You will take my feedback.

55:00

>> We're going to ignore your feedback.

55:01

Take your feedback. Public service.

55:03

>> Throwing in the trash. No, you take it

55:06

and I will be showing up at the White

55:08

House for my tour.

55:10

>> You have this crazy idea that we're

55:11

going to win the AI race by tying one

55:13

hand behind our back so that you can

55:14

subsidize journalists so you can

55:17

subsidize movies. You'll get more

55:20

content. You said before you want more

55:21

training data, pay for it. Pay for more

55:24

training data. You're the zar. Take it

55:26

back to pus. All right, let's keep

55:28

moving here.

55:29

>> We have to keep moving. We have a great

55:31

This is great debate. Great debate here

55:33

on the All-In podcast. It's not going to

55:35

stop, folks.

55:35

>> It's just you yelling. It's just you

55:37

yelling saying things that don't make

55:38

sense.

55:39

>> Okay, you can say that.

55:40

>> You only have like three topics going

55:43

on. You can you can personally attack.

55:45

>> You know what it is? It's like we got to

55:46

let in more immigrants. Number one.

55:48

Number two, high skilled immigrants.

55:50

>> AI is going to put everyone out of work.

55:52

By the way, no sense of perceived

55:53

contradiction between those two things.

55:55

Number three, we need to like subsize

55:57

here. You know, the audience says the

56:00

same. When the three of you guys attack

56:02

me, Jason,

56:05

when the audience

56:06

gang up on me like this, the three of

56:08

you gang up on this and you personally

56:09

attack me, the audience comes up to me

56:11

and they say, "Wow, you really nailed

56:13

and beat."

56:13

>> Have you done that today?

56:14

>> No, not yet. Not yet. But a little bit

56:16

of the buting.

56:19

>> Yeah, that's true. Let him eat the

56:22

He's emaciated. He's 11% body fat. Let

56:25

him eat. Let him cook. All right.

56:27

Listen, you and I, Saxs, will do more

56:29

debate and it's going to be amazing.

56:30

allin.com/ yada yada yada for tickets.

56:33

Get in there, folks. Uh, we have to get

56:35

to the docket. We're an hour in and we

56:37

still have all the news.

56:38

>> We should talk about this this um AI

56:40

privacy issue that Sam Alman mentioned.

56:41

>> All right, that's a great segue cuz I

56:43

saw that as well, David Saxs, and as our

56:46

civil servant working on AI, this is

56:48

something where you can have an

56:49

additional contribution. There's more

56:51

work we can give you. All right, listen.

56:53

Here it is. AI user privacy is becoming

56:54

an issue because friend of the pod Sam

56:58

Alman says there is no legal

57:00

confidentiality when using his product

57:03

chat GPT. Here's a 30 secondond clip

57:06

again. Friend of the pod fop Sam Alman

57:10

on the vaugh people talk about the most

57:13

personal in their lives to chatt

57:15

young people especially like use it as a

57:16

therapist a life coach uh having these

57:18

relationship problems. What should I do?

57:20

And right now, if you talk to a

57:23

therapist or a lawyer or a doctor about

57:25

those problems, there's like legal

57:27

privilege for it. We haven't figured

57:28

that out yet for when you talk to

57:29

ChachiBT. So, if you go talk to CHP

57:31

about your most sensitive stuff and then

57:33

there's like a lawsuit or whatever, like

57:35

we could be required to produce that.

57:36

And I think that's very screwed up. I

57:38

think we should have like the same

57:39

concept of privacy for your

57:41

conversations with AI that we do with a

57:44

therapist or whatever.

57:45

>> Okay, Saxs, this is bringing up

57:47

something super important.

57:49

What's your take on it?

57:50

>> Okay. Well, I I think this is an

57:52

interesting topic because like

57:53

copyright, this is an area where we have

57:55

existing law, but it does make you

57:59

rethink whether those laws are truly

58:02

applicable or make as much sense in this

58:04

new world. So, the existing law, the

58:08

existing example is search history. You

58:10

know, the government can get a copy of

58:12

your search history. They can subpoena

58:14

it.

58:14

>> Yeah. Every true crime story starts with

58:17

the person's search for how do I kill my

58:19

husband slowly with you poison and then

58:21

they Yeah, that's

58:22

>> right. Exactly. The point is though that

58:26

I think Sam is right about the legal

58:28

treatment right now, which is that your

58:31

chat history isn't any different than

58:33

the search history in the eyes of the

58:34

law. But it is much more personal. It's

58:37

much more interactive than your search

58:38

history. You are using it like you said,

58:41

you could use it as a as your doctor.

58:43

You could use it as your therapist. You

58:45

could use as your lawyer. And so the

58:48

ability for the federal government to be

58:50

intrusive is so much greater than with

58:53

your search history. So I don't know

58:56

what like the right policy should be

58:59

yet, but I I will say it does make me

59:01

uncomfortable.

59:02

>> Yeah, there's a market.

59:04

>> Can I make a recommendation to my AIR?

59:06

>> Yes, please. He's our

59:07

>> Why don't we Why don't we let AI models

59:11

get bar certified and get medically

59:15

certified? So, if the AI models, it

59:18

turns out, are actually proving to be

59:19

more accurate, more thoughtful,

59:23

more responsive, more reasonable,

59:25

whatever it is, whatever metric we're

59:26

using, and they pass the same criteria

59:29

as one would need to pass to qualify for

59:31

the bar or to qualify for a doctor

59:34

certificate. Why don't we do that for

59:35

the AI? If that then happens, then the

59:37

the same privilege occurs to the AI as

59:40

it does to the individual human that

59:42

does it. And now if you extrapolate from

59:43

where that takes us, if we're suddenly

59:45

giving AI the same sort of privileged

59:47

rights that we give to privileged

59:48

humans, where's that going to take us

59:50

ultimately with respect to the overall

59:52

rights for AI?

59:52

>> Well, and they have responsibility. Hold

59:54

on a second. I just want to point out

59:56

here once again, you have a mind-blowing

59:58

concept here. I've never heard anybody

60:00

vocalize that. Could they actually be

60:01

certified in that knowledge? And if they

60:04

pass the test, makes sense they would.

60:05

But then you also get responsibility.

60:07

>> So with great power comes great

60:08

responsibility. I will tell you this.

60:10

>> You can turn this stuff off. But this is

60:12

an opportunity. I'm going to send a note

60:13

to you.

60:14

>> And it sounds crazy today, but I

60:15

guarantee if you put it on Poly Market,

60:16

there will be a date when this happens.

60:18

>> Poly market. Shout out to Shane. Let's

60:19

get that up there. I just want to point

60:21

out I'm going to email Elon about this

60:22

when I get off the pod.

60:24

>> This is an opportunity to create the

60:25

signal of the signal equivalent of an

60:28

LLM. The all of your chat should be

60:30

encrypted. All of it should be by by

60:33

default. Encrypt it by default on Grock.

60:35

Make it so that Grock can't even see it.

60:38

They don't have it. So when you try to

60:39

subpoena it, you can do what Tim Cook

60:41

does, which he says like, I don't have

60:42

it. You, if you want to try to back door

60:45

it, you can. That's a market

60:46

opportunity. I can tell you I only use

60:48

the Brave browser and Brave search for

60:50

this reason. I don't want my search

60:52

history like saved somewhere or

60:54

whatever. that. You can take

60:56

control of this as an individual, but

60:57

the defaults matter and you have to then

60:59

do the work. It's a great market

61:00

opportunity. Chimoth, I don't even want

61:02

to know what you're talking to chat GPT

61:04

about. What are you What's in your chat

61:05

GBT logs?

61:07

What's in there, Chimoth? How to extend?

61:09

How to get the extra centimeter? What's

61:10

in there? You trying to

61:13

What's in there?

61:14

>> I keep asking it to find me a moderator.

61:16

>> Oh, great. I keep asking it to find me a

61:19

participant who's not a douche. Um,

61:25

>> my god, you are so deep in your villain

61:27

era and you're leaning into it and I'm

61:29

so here for it, Chimat. I love your

61:30

villain.

61:31

>> You know why? I am so

61:33

>> Why are you going into your villain?

61:34

>> I am so riskon right now. It's like

61:37

>> it's liberating actually. It's amazing.

61:41

>> It's really amazing.

61:42

>> Is there any blowback to how outlandish

61:45

you've become this year? Any blowback at

61:47

all? Has it had any negative consequence

61:49

on business or hiring or anything?

61:51

>> No. But but outlandish? How how have I

61:53

been outlandish?

61:54

>> You're you're you're just filter off.

61:56

You're filter off. And I think it's

61:58

great. I think you're over two windows

61:59

back. It's absolutely fantastic we're

62:00

seeing here. I asked chat GPT about my

62:03

future and uh my IQ. It's very

62:05

interesting when you ask chat GPT to

62:06

analyze you.

62:07

>> I suggest everyone do it.

62:08

>> Well, actually, yeah, when you just ask

62:10

chat GPT or whatever, what do you know

62:12

about me?

62:14

>> And it's scary how much it already does.

62:16

>> It's scary. There's this great

62:17

personality test. You can put this

62:18

personality test into Grock and this guy

62:20

like made this prompt and it goes and it

62:23

tells you all your personality based on

62:25

your Twitter ex history. It is wild how

62:28

accurate it is. What does it say about

62:29

you, J? I'm actually curious.

62:31

>> It says the same thing about all of us.

62:33

We're all likeworked

62:35

narcissist, ENTJ. You can literally run

62:38

the MyersBriggs against your

62:40

>> Yeah, your chat history. It's actually

62:43

But I I like your mind-blowing concept

62:44

there, by the way, of like them becoming

62:46

certified in some way. Okay, fresh

62:49

economic news. It's time for the

62:51

administration to take their victory

62:53

lap. GDP growth was 50% higher than

62:56

expectations in Q2 as the Fed held rates

62:59

at 4.25%. In Q1, GDP declined 50 basis

63:03

points. That's probably due to the

63:04

imports. People were stockpiling goods.

63:06

>> That's the most pointless chart ever.

63:09

>> Okay. And then Yeah, it is. I agree.

63:10

It's a little bit Yeah,

63:11

>> it's distorted by I wanted to have both.

63:14

I wanted to have both as bar charts.

63:15

This one

63:16

>> totally on drugs. Just say it. It's

63:18

okay.

63:18

>> What drugs are you on? I'm not I had

63:21

coffee and now I'm out. I'm out.

63:23

>> We're all friends. You can tell us. Is

63:24

it really just out?

63:25

>> All right, that's it. I'm taking it out.

63:27

>> Oh my god.

63:28

>> I took it out. And now let's get back to

63:30

the here. Okay. The Fed kept rates

63:32

unchanged for the fifth straight

63:34

meeting. This time, two out of 11 Fed

63:36

governors desented from Pal's decision.

63:38

Two of the dissenters were both

63:39

Republicans nominated by Trump. So, it

63:41

seems like the Fed is becoming a little

63:42

polarized now, too. First time in 32

63:44

years that more than one governor

63:47

dissented. And um yeah, even one person

63:50

dissenting is rare. Here's a 25-second

63:53

clip of Pal explaining how GDP factored

63:57

into the cut decision. Nick, please play

63:59

the clip. Recent indicators suggest that

64:02

growth of economic activity has

64:04

moderated. GDP rose at a 1.2% pace in

64:08

the first half of this year, down from

64:10

2.5% last year. Although the increase in

64:14

the second quarter was stronger at 3%,

64:16

focusing on the first half of the year

64:18

helps smooth through the volatility in

64:21

the quarterly figures related to the

64:23

unusual swings in net exports. The PCE

64:26

index and then I'll throw this over to

64:27

you Sax for for the official position

64:29

here for June dropped on Thursday. PC is

64:32

the Fed's preferred gauge of inflation

64:33

over CPI. PCE rose 30 bips in June in

64:37

line with estimates and um if you

64:40

remember we talked about in a previous

64:41

episode CPI rose a bit 13% or 30 bips

64:45

from May to June. So, we're not any

64:47

we're not close to the 2% uh target and

64:50

that's what the Fed keeps saying. We're

64:53

not there yet. And the economy is al

64:55

fuego saxs. You note I don't know if you

64:58

noticed this, but people are talking

64:59

about the QDP the second quarter print

65:02

which was amazing for GDP. You were

65:04

talking about it a bunch on the socials.

65:06

He keeps referencing the first half. So,

65:08

he's trying to blend those two together,

65:10

I think, because of the the tariff

65:11

differences or, you know, maybe to to

65:13

smooth it out as he said. What's your

65:15

take on this? The GDP boomed in, you

65:17

know, 3% which is pretty great.

65:20

>> The problem No, the problem that Jerome

65:22

Powell has is that he's trying to smooth

65:24

it because it allows him to justify his

65:29

political decision.

65:30

>> Okay.

65:31

>> But the reason why you have to segregate

65:33

Q1 and Q2, Q1 was before tariffs and Q2

65:37

was after tariffs. So I think you have

65:38

to segregate these two things. And if

65:40

you look at the run rate from Q2, what

65:42

you're probably going to see in Q3 and

65:44

beyond is more similar to Q2, which is

65:48

to say a large surplus, good GDP

65:52

expansion,

65:53

and moderating inflation. So why does

65:57

the Fed not cut? Because at this point,

66:00

not cutting is the only thing that you

66:02

can do to slow the Trump administration

66:04

down going into the midterms if you

66:06

wanted to politicize the job.

66:11

If, however, on the other hand, you just

66:12

take the data as is and you ignore Q1

66:15

because it was pre-tariff and you start

66:17

to look at Q2 and you project forward,

66:20

if you inject a 100 basis point cut into

66:22

the economy, this thing is going to go

66:23

gang busters and Trump is going to look

66:25

like an economic genius going into 2026.

66:28

So I think that again in the absence of

66:30

politics you cut

66:32

>> Okay. Sachs, what's the take from inside

66:34

the administration and around it? I know

66:36

you're you're not speaking for the

66:37

president on this issue, but you're in

66:39

the administration, so I'm assuming

66:40

you're

66:40

>> Look, I'm not speaking for anyone, but

66:42

obviously the 3% number is way ahead of

66:44

expectations. It's a fantastic number.

66:47

It just feels like, you know,

66:48

everything's humming on all cylinders

66:50

here. One thing you didn't mention, but

66:52

I think is relevant, is the new trade

66:54

deal with the EU.

66:55

>> We're about to get to that, by the way.

66:57

That's the next story.

66:58

>> Oh, okay. Well, I mean, I would include

67:00

that because

67:00

>> Okay, include it. Yeah.

67:02

>> I mean, I think it was a deal that just

67:04

got announced where the EU is going to

67:06

open its markets to US products. No

67:08

tariff on US products, but they will pay

67:10

a 15% tariff coming into the US. They're

67:13

going to be investing 600 billion in the

67:16

US. They're going to be buying 750

67:18

billion of US energy. and then some very

67:22

large number I guess they didn't specify

67:24

number on defense products basically

67:26

American military products hundreds of

67:28

billions which is the followup to their

67:31

commitment to raise their contribution

67:33

to NATO to 5% of GDP up from I guess it

67:36

was sort of like 2% before

67:38

>> so I mean this is a huge deal for the

67:42

United States I think it's a huge win

67:44

for the Trump administration and the

67:46

deal is so good that what I'm seeing

67:48

from European sources on X European

67:51

publications just commenters is that

67:55

they they were like outraged. They felt

67:56

like they got taken to the cleaners

67:58

here.

67:59

>> Good.

67:59

>> And

68:00

>> okay,

68:01

>> you see you see a lot of that on X by

68:03

European side. A lot of the European

68:05

leaders are saying that Ursula chickened

68:07

out. So, you know, all those stupid taco

68:10

memes are going away now because people

68:11

are realizing that

68:14

Trump's willingness to raise tariffs on

68:18

these countries as a threat to

68:21

renegotiate better trade deals is

68:23

working. It's working extraordinarily

68:25

well. Just this EU deal, one way to

68:27

think about it is you add it all up,

68:28

it's about $2 trillion. It's effectively

68:31

$2 trillion of stimulus into the US, but

68:34

without money printing.

68:36

>> Yeah. Over the next three years. So it's

68:37

not inflationary.

68:38

>> It's not insignificant. Freedberg, your

68:40

thoughts on the Fed, the GDP print, and

68:44

then maybe you could get into the

68:45

granular

68:47

details of that print. If you pull up

68:51

the schedule of data, so this is the

68:53

national income and product accounts

68:55

data from the Bureau of Economic

68:57

Analysis. So this is where the inflation

69:01

print comes from. I think there are two

69:03

lines worth taking significant note of.

69:06

The first is the furnishings and durable

69:08

household equipment line. So in June,

69:11

the cost for furnishings and household

69:13

stuff jumped 1.3% month over month on an

69:17

annualized basis, right? That's almost

69:19

15% year-over-year if it were to

69:22

continue at that level. And then the

69:24

second one is this recreational goods

69:27

and vehicles that jumped 9% month over

69:30

month. Neither of those categories have

69:31

jumped that much in in kind of recent

69:33

history. So part of the argument that's

69:36

being made is that what we are seeing in

69:39

these jumps is actually some of the

69:41

first effects of the tariffs and the

69:44

cost of goods that are being imported

69:45

because these are largely imports having

69:48

an adverse effect on the consumer. And

69:51

so I think this is kind of a wait andsee

69:52

moment on some of these categories that

69:55

are predicted to have a tariff price

69:57

effect starting to show through. So I

69:59

think this is where a lot of folks are

70:00

keeping a close eye on and it kind of

70:02

provides a little bit of the support for

70:04

the economists that are saying we should

70:05

keep rates steady because if we are

70:07

seeing a significant inflationary effect

70:09

here it's worth noting that there's

70:10

something that we need to be thoughtful

70:11

about rate policy.

70:13

>> I think this is um a really good point.

70:15

If you look in this debate, which is

70:17

obviously highly political,

70:20

we're at inflation 2.567%.

70:24

Spending is increasing. Obviously, stock

70:27

market at an all-time high. Unemployment

70:29

trending down again. So, we're at like

70:31

4.1%.

70:33

And people are just yoloing into crypto

70:35

and they're doing sports betting.

70:36

Bitcoin at an all-time high. I think the

70:39

Fed now is in a position where cutting

70:41

rates seems like putting kerosene on the

70:43

fire. If Trump

70:45

tanked the economy in Q2, he probably

70:47

would have gotten the rates. But now I

70:49

don't think it's reasonable, as you're

70:51

saying, Dave, the the reasons to not cut

70:54

are building because the economy is on

70:56

fire. So maybe the shock and bore

70:58

approach to tariffs, which is now

71:02

becoming a playbook. I had a nice talk

71:03

with Lutnik about this, who I love, by

71:05

the way. He really described to me how

71:08

they're doing these. And the shock and

71:09

bore playbook is basically Trump says

71:12

something completely outrageous,

71:14

shocking, everybody goes crazy, the

71:15

media loses their mind, business leaders

71:17

lose their mind. Lutick told me that

71:19

what he does is he sets the table and

71:21

proposes something reasonable because,

71:23

you know, now I'm a big, you know,

71:25

direct contact with all the

71:26

administration. Sax, thank you for that.

71:28

Um,

71:30

and he described it. Trump comes in,

71:31

sees all the stuff, and then he starts

71:33

making these micro tweaks. So it's on

71:36

the finish line. It's in the red zone, 5

71:38

yard line. Trump comes in and then he

71:40

sticks it to them again with three or

71:41

four extra asks and then they wrap it up

71:44

and that this is becoming really

71:46

effective. So it was chaotic at first.

71:48

It seemed nonsensical, but they've put

71:50

the Fed in a really bad position because

71:53

they never seen this before. They've

71:56

never seen this before. So now they're

71:57

going to be in this defensive position

71:59

of what if we cut it and the market

72:02

rips. To your point, Chimoff, you just

72:04

said the market will rip the second they

72:05

cut that. And the cynical view of this

72:08

is the market rips as we go into the

72:10

midterms, which is the same claim the

72:12

Republicans made about the cuts that

72:14

Biden did in September right before the

72:17

election. So, there is some level of

72:18

politics and gamesmanship going on here.

72:20

But you have to hand it to the Trump

72:22

administration for what they're doing

72:25

with this sort of 2.0 playbook. If this

72:28

was Sachs premeditated and we all just

72:30

didn't understand it, fine.

72:32

The outcome here is

72:35

this administration has to live or die

72:37

by the results of these 600 billion from

72:40

the EU, 550 billion in investment from

72:45

Japan. You put those two together, I

72:46

asked Lutnik, is that at the event, is

72:49

that going into the sovereign wealth

72:50

fund? And how does that get, you know,

72:53

spent? And he said at the discretion of

72:55

the president and he's advising him to

72:57

spend it on putting more nukes in. So

72:59

that's fascinating. We have a trillion

73:01

dollars now that we can put into nuclear

73:05

power plants and these small module

73:08

reactors. And that's what Lutnik said he

73:10

wanted to spend it on. He's going to

73:11

advise the president to spend it on. But

73:12

now we've got them investing in our

73:14

country. It's absolutely brilliant. If

73:16

it works out, we'll see if it works out.

73:18

>> April 2nd was liberation day and the

73:22

media went crazy. They were predicting a

73:24

black Monday. The market crash. They

73:26

basically tried to spook the markets and

73:28

create fear. They said that we're going

73:30

to go into a recession or depression.

73:32

And now look at where we are. It's just

73:33

a few months later. All the markets are

73:35

at all-time highs. Trump has extracted

73:38

trillions of dollars in these trade

73:40

deals that people know.

73:41

>> Premeditated. Tell us the truth.

73:43

>> Premeditated.

73:44

>> Hold on. President Trump has extracted

73:47

trillions of dollars from other

73:49

countries using powers that other

73:51

presidents didn't even know they had.

73:53

>> 100% 100%. Was it premeditated? Cuz it

73:56

was chaotic. the market did the market%

73:59

moves because of the media by the way

74:01

those moves because they were scared

74:02

>> and we just had a 3% GDP growth

74:05

>> print

74:07

how things could be

74:08

>> what I think happened is that President

74:11

Trump saw an opportunity here that other

74:13

people ignored it's like when a CEO

74:15

comes in to a company a new CEO comes in

74:18

and that company's been mismanaged for a

74:20

decade but it's got wonderful assets on

74:22

its balance sheet it's got a market

74:24

position that's still very strong has

74:26

been underutilized. And he came in and

74:29

understood that the United States had

74:30

tremendous leverage in all these trade

74:32

negotiations. Actually, they weren't

74:34

even trade negotiations then, in all

74:35

these trade relationships. And he was

74:38

able to essentially renegotiate all of

74:40

them. And look at the results. I mean,

74:42

they're just staggering. And you know,

74:44

everyone said that, "Oh, Trump's going

74:46

to chicken out. He's not going to hang

74:47

tough." It's all these other countries

74:49

that have folded like, I don't know,

74:51

lawn chairs. I mean, they have all

74:54

capitulated.

74:55

>> Yeah. Yeah, they follow.

74:56

>> It's really remarkable.

74:58

>> But you're not answering my question.

74:59

Was this premeditated? Give us some

75:00

insight here.

75:01

>> I don't know what this is. What are you

75:02

talking about?

75:03

>> When they came out and they was like,

75:04

"Oh, 100% tariffs, 200% tariff." The

75:07

market was not making that reaction

75:09

based upon the media. They were making

75:11

it based on Trump was saying. So, was it

75:14

premeditated this shock and bore shock

75:16

and reasonable negotiating strategy or

75:19

do you not know? Well, you're not privy

75:20

to Look, I'm not speaking as an insider

75:22

here, but we said at the time that all

75:25

of that was happening and Larry Summers

75:26

was on the pod preaching doom is that

75:29

all of that was an opening bid. It was

75:31

all a start to a negotiation and we had

75:33

to see where it ended up and that the

75:36

administration still had to stick the

75:38

landing.

75:38

>> Okay,

75:39

>> but I got to say based on

75:41

>> EU, Japan, and South Korea, I mean, this

75:43

is looking really good right now.

75:45

>> Well, listen, it's the top five that are

75:47

like 90% of the negotiation. As Trump

75:49

said, there was another little note he

75:50

did in the keynote when he kind of

75:52

drifted into his, you know, different

75:53

things he wanted to talk about where he

75:54

said, "I don't even need to know about

75:56

the bottom countries. I've never even

75:57

heard that names of some of these

75:59

countries." He's just got to nail the

76:00

what, the top five, the top 10, and

76:02

we're done. And this administration has

76:04

to stick the landing as well because

76:07

these are handshake deals right now.

76:08

They have to be inked. They have to be

76:10

approved. So, there's there's a lot more

76:12

work left to be done. But I said as

76:14

well, there's one other piece of it. We

76:16

talked about the the um

76:18

>> you know the fact that Europe has 0%

76:22

tariffs on American products but but

76:23

even after this deal

76:26

>> that the European products into coming

76:28

into the US will have a 15% tariff and

76:30

we're not including the $600 billion of

76:32

European investment in the US. We're not

76:33

including the 750 billion of sales of

76:36

American energy to Europe. Okay, just

76:39

talk about the tariff that 15% and what

76:41

we're seeing now across the board is

76:42

generating about 300 billion a year of

76:45

additional tariff revenue that goes to

76:47

help balancing the budget.

76:48

>> Yeah.

76:49

>> So 300 billion a year over 10 years is 3

76:51

trillion. That is a big number.

76:52

>> It's incredible. Yeah. It's got to

76:54

>> I don't know if that that completely

76:55

satisfies Freeberg, but that's a big

76:58

help.

76:58

>> Freeberg, do you think that there is a

77:01

chance that inflation is going to tick

77:04

up because of all this? like this is a

77:06

lot of money being pushed into the

77:07

system again. So, could we see a

77:09

three-handle on inflation in the next 6

77:11

months or what's the probability of that

77:12

in your mind?

77:13

>> That's the big concern everybody has.

77:15

>> I don't I I don't know. I I don't know.

77:16

I think the the big question if you look

77:18

at each of these categories, one way to

77:20

think about it is how much margin is the

77:23

seller making?

77:24

>> If they're making 30% margin and we

77:27

charge a 15% tariff, does their margin

77:30

go down to 15%. or do they take their

77:33

margin down to 20% and raise the price

77:35

by 5%. What's the right balance? And

77:38

what will happen is that now with this

77:41

effective you know tariff which is a

77:43

sort of tax on the system a tax on the

77:45

market will find its kind of new

77:48

equilibrium where the buyers are willing

77:50

to pay X and the sellers are willing to

77:52

sell it Y and I think every market's

77:54

going to be a bit different. So I think

77:55

in some of these categories we will see

77:57

significant inflation where there is a

77:59

very thin margin that the seller has in

78:01

selling and in some of the categories

78:03

where there's a monopoly and they have a

78:04

big margin they're going to eat it

78:06

because they don't want to have

78:07

competition and they don't want to see

78:09

pricing competition emerge. So I think

78:10

we'll see it vary by category and you

78:12

know we'll see how it goes. All right

78:14

listen this has been another amazing

78:16

amazing episode of the number one

78:17

podcast in the world according to Jensen

78:19

Wong from Nvidia and me and uh great job

78:23

everybody. Great job to everybody. It's

78:25

a class.

78:25

>> Great job everyone. Even JCAL. Even J.

78:28

Even Jason. Great job.

78:31

>> And actually, I want to thank Freeberg

78:33

cuz Freeberg did most of the work to

78:34

organize the AI.

78:36

>> Let's give him a big shout out. There's

78:37

me.

78:38

>> Great job.

78:39

>> I mean, guys, can we just make a note

78:41

here?

78:42

>> One of us can run for Manurian candidate

78:45

president in eight years. And look at me

78:47

and the president. I put on the red tie

78:49

out of respect. I put my blue suit on

78:51

out of respect for the president. Does

78:53

it not look like I'm running president

78:56

Jason.com?

78:59

All right, listen.

79:00

>> That photo could be like, you know, that

79:02

famous photo of uh Bill Clinton meeting

79:04

JFK. You know, that could be the thing

79:06

that that could be the thing that

79:08

>> I'm in like that famous image

79:10

>> that propels you to the presidency.

79:12

>> I'm in like Thank you for giving me that

79:14

and and for putting me in touch with

79:15

each member of the administration

79:17

directly. Thank you for that. And we had

79:19

a wonderful tour of the White House the

79:21

next day. What a wonderful tour. Some of

79:24

us had at the White House the next day.

79:25

But in all honesty, no. I was

79:27

>> Did you?

79:27

>> No. I was taking the pictures. That was

79:29

my joke cuz it was all of you guys were

79:31

giving you a tour. We could have gotten

79:33

you a tour.

79:33

>> I mean, listen, I love J.

79:34

>> Did you ask for a tour?

79:36

>> I'm I'm not the kind of guy to ask. I'm

79:38

the guy.

79:38

>> Some of us have actual meetings to do,

79:40

bro. I mean,

79:40

>> it's all good. It's all good. I got a

79:41

lot going on. I got a lot to announce

79:42

and happen in the coming weeks. But Sax,

79:44

do do take us behind the scene here.

79:46

>> And I think it was hilarious. So, I

79:48

don't mind getting trolled by the

79:49

president. It was great. But how did you

79:51

how did that go about behind the scenes

79:53

that he nailed that joke?

79:54

>> Don't tell him. Leave it. Leave it.

79:56

>> What did you do? I mean, cuz that looked

79:57

like it was workshopped or is he just

79:59

naturally I mean, he's obviously

80:00

naturally comedic, but did you put that

80:02

in with him? Did you have to clear that

80:03

with him? Hey, Dunan, Jal, whatever.

80:06

>> Well, they asked me for the names of,

80:09

you know, my co-hosts and so they could

80:11

do shout outs. So, I gave him the list.

80:14

>> Oh, no. And I just I said and I put even

80:16

Jake House,

80:20

>> but I mean he went for it.

80:23

>> No, he he he got we went through it. So

80:25

>> he got the joke. Okay.

80:26

>> He got the joke. We went through it.

80:27

>> He got the laugh. He got it. He heard

80:28

the laugh and he heard the laugh and he

80:30

doubled down.

80:31

>> I thought it'd be funny. But no, we went

80:33

through everyone's names beforehand

80:35

>> and uh

80:36

>> I mean, talk about EQ. The guy's EQ is

80:39

off the charts, man. He just he's timing

80:42

is great. I suggested I suggested the

80:44

name JCAL and he's like, "No, no, give

80:45

me his full name." He thought it was

80:47

more courteous.

80:48

>> He's actually a very

80:50

>> courteous man. He wanted to use your

80:54

full name, not just your nickname.

80:55

>> I think what he probably realized was

80:56

for my parents who were

80:58

>> just over the moon. So, thank you for

81:00

that. It meant a lot to my dad who's

81:02

>> That's lovely.

81:02

>> Yeah. He's been struggling a bit and it

81:04

it really Let me get a little choked up

81:06

here, but my dad's been struggling a

81:07

bit. And um I got to see him in Brooklyn

81:09

after that and we were on a tech stream

81:11

and and it meant a lot you know cuz for

81:14

a kid from Brooklyn to get a shout out

81:15

from the president of the United States

81:16

is

81:16

>> you made it.

81:17

>> I mean it's just

81:18

>> your father your father should be really

81:20

proud of you.

81:21

>> Thanks man. Appreciate it. I appreciate

81:22

it boys. All right listen for your

81:24

Sultan of science the amazing Dave

81:26

Freeberger who put that event together

81:27

in 10 days and then jumped right in.

81:29

He's got to run a hollow at the same

81:30

time. So I just want to give our MVP of

81:33

the week.

81:33

>> We should give a shout out to the Hill

81:34

and Valley guys for partnering with us.

81:36

Jacob, Jacob Hellberg did a great job.

81:38

>> Love Jacob. I love Jacob.

81:39

>> And Delian and Chris.

81:42

>> Thank you guys. They were our partners

81:43

on the event.

81:44

>> Hillen Valley did a great job. Yeah, I

81:45

love those guys.

81:47

>> But yeah, just I'm giving the MVP of the

81:49

week for of the besties to you, David

81:51

Freeberg. You put a lot of work into

81:52

this. So, and we appreciate it. You're

81:54

running a hollow and then you went right

81:55

into working on the all-in summit which

81:57

we'll be at in a couple weeks. Chimat,

81:58

thank you for buttoning up. We're

82:00

getting a little complaints from the HR

82:01

department about the uh the buttons and

82:03

so we we've now renegotiated that. I'm

82:05

going to I'm going to unbutton three

82:07

buttons now and walk around Forte.

82:09

>> Perfect. And Saxs, I will see you at the

82:11

White House. JD and I will be in the

82:12

commissary. So, we'll invite you to

82:14

lunch with us in JD.

82:16

>> It's called the Navy actually

82:18

>> in the mess. Yeah. And and you know

82:19

what? Lutnick's joining us as well. And

82:21

um who's our energy guy? Chris. Chris

82:23

said he wanted to jump in on that. So

82:25

maybe you can join us. I'll invite you

82:27

now that I am deep into the

82:28

administration.

82:30

>> Thank you for tuning in everybody.

82:31

Allin.com events. The scholarship

82:33

tickets are up. So, if you want to try

82:35

to get one of the very few scholarship

82:36

tickets, we always like our upandcomers.

82:38

Please, if you're if you're of means,

82:40

don't apply for the scholarship. You

82:41

won't get it in. But if you're up and

82:42

coming and you're part of the audience

82:43

and you want to get one of those

82:44

discounted tickets, we have a limited

82:45

number of those available.

82:46

All.com/events.

82:48

Love you besties. Byebye. Love you.

82:50

Byebye.

82:52

>> Let your winners ride.

82:55

>> Rainman David.

82:59

>> And it said we open sourced it to the

83:01

fans and they've just gone crazy with

83:03

it. Love you.

83:04

>> Queen of kinshine.

83:07

[Music]

83:12

>> Besties are gone.

83:15

>> That is my dog taking a notice in your

83:17

driveway.

83:20

>> Oh man, my dasher will meet me up. We

83:23

should all just get a room and just have

83:24

one big huge orgy cuz they're all just

83:26

useless. It's like this like sexual

83:28

tension that we just need to release

83:29

somehow.

83:31

Wet your feet. Wet your feet.

83:34

>> Your feet.

83:36

>> We need to get mer.

83:41

[Music]

83:45

I'm going all in.

Interactive Summary

The video opens with casual banter about nicotine pouches and a side hustle. The main discussion centers around the All-In podcast's AI summit in Washington D.C., including interactions with the president and the administration's AI policy. Key policy pillars for AI, such as innovation, infrastructure, and an executive order against "woke AI" in federal procurement, are discussed. The hosts share their individual highlights from the summit, covering market commentary, energy policy (especially nuclear), and a heated debate on immigration and H-1B visas with JD Vance. A significant portion of the conversation focuses on the future of intellectual property and copyright in the age of AI, exploring concepts of fair use versus compensation for content creators, and the long-term viability of patents. The panel then delves into AI user privacy, prompted by Sam Altman's comments on ChatGPT's lack of legal confidentiality, leading to proposals for certifying AIs as professionals or implementing default encryption for AI conversations. The episode concludes with an economic update, analyzing GDP growth, the Federal Reserve's interest rate decisions, and the substantial impact of new trade deals with the EU and Japan, including discussions on tariffs and their potential inflationary effects.

Suggested questions

7 ready-made prompts