0:00

Hey everybody, it's This Week in

0:02

Google. Leo's out and I'm filling

0:04

in. We've got Jeff Jarvis as

0:06

usual and a special guest at

0:08

Zitron. This week on the show, we talk

0:11

about AI and the new

0:13

groundbreaking legislation the EU passed.

0:15

We also talk about TikTok.

0:18

Will it end up being forced to shut

0:21

down or be sold? We

0:23

also talk a lot about Kara

0:25

Swisher and this great New York

0:27

Times story about how automakers are

0:29

sharing consumers' driving behavior with insurance

0:31

companies. All that and more coming

0:33

up on the show. Stay tuned.

0:38

Podcasts you love. From

0:40

people you trust. This

0:43

is Twig. This is Twig. This

0:45

Week in Google, Episode 759, recorded Wednesday, March 13th, 2024.

0:47

The posters will rise.

1:01

This episode of This Week in Google

1:04

brought to you by Melissa, the

1:06

data quality experts. Melissa has helped

1:08

over 10,000 businesses worldwide

1:10

harness accurate data with their

1:12

industry leading solutions. Processing,

1:15

get this, over 1 trillion

1:17

address, email, name and

1:19

phone records. 1 trillion. G2

1:21

and its 2024 grid report is once

1:24

again recognized Melissa as leaders

1:26

in data quality and address verification.

1:29

Melissa knows money is tight.

1:32

They offer free trials, sample

1:34

codes, flexible pricing with an

1:36

ROI guarantee and unlimited technical

1:39

support. All customers everywhere in

1:41

the world. When

1:43

it comes to data, Melissa treats

1:46

your data like the solid goal

1:48

they protected. In fact, they are

1:50

FedRAMP authorized as an address data

1:52

resource. That means you're getting

1:54

the highest level of security for your

1:56

customer data. It also means governmental agencies

1:59

can use. Melissa with confidence. It's

2:01

also GDPR and CCPA compliant.

2:03

It meets SOC 2 and

2:05

HIPAA HiTrust standards for information

2:08

security management. They got it

2:10

all covered. Improve your e-commerce

2:12

operations with Melissa. Now as

2:14

an SRE partner, Melissa's cloud-based

2:16

tools standardize, validate, verify, and

2:18

enhance data on which top

2:20

businesses rely, ensuring address data

2:22

is seamlessly location optimized. They're

2:25

constantly getting better. Oh, and

2:27

don't forget to try the

2:29

free Melissa Lookups apps. They're on

2:31

Google Play or the iOS app

2:33

store. No signup is required. Get

2:36

started today at Melissa. 1000

2:38

records cleaned for free at

2:40

melissa.com slash twit.

2:43

That's M-E-L-I-S-S-A,

2:45

melissa.com/twit. It's

2:48

time for twig this week in Google. As

2:50

you may have guessed, I am not Leo

2:52

Laporte. In fact, I'm Para Smartenome. Leo has

2:54

abandoned us this week, fleeing the country to

2:57

the shores of Cabo. In his

2:59

stead, I have seized the means of

3:01

podcast production and God only knows what

3:04

will follow. Laughing over there is,

3:06

you know, who is not in Cabo this week?

3:08

Jeff Darvish. I'm never anyplace funny. And

3:10

I, I for one, welcome my new

3:12

master. Thank you. He's

3:14

the director of the Tau Knight

3:17

Center for Entrepreneurial Journalism at the

3:24

School of Journalism at the City University

3:26

of New York. It feels so powerful

3:29

to be able to cue that up.

3:31

Ed is right away. He just said

3:33

Christ. I'm going nowhere. Yeah. I was

3:35

about to say speaking of people

3:37

who are just right immediately on the pulse of

3:39

the show, we've got Ed Zittron

3:42

here, CEO of EZPR and

3:44

host of the Better Offline

3:46

podcast. Ed, welcome. Thank

3:49

you for having me. I will be Leo Laporte

3:51

this episode. That's

3:53

really important. We all at some point are

3:55

going to need to take turns being Leo

3:58

and you know, you'll never. You're never

4:00

going to know which one of us is

4:02

going to suddenly become an AI accelerationist and

4:04

that's really important Yeah,

4:07

that's definitely what I'll be doing Was

4:14

like I we need a we need

4:16

a contrarian Perspective to balance out

4:18

the amount of shows we've had that

4:20

are like AI should have access to

4:22

everything Yeah,

4:25

no Also, even

4:27

if it did what's it gonna do? It doesn't do anything

4:30

Great great point Well,

4:34

the information just had a story about

4:36

that About trying

4:39

to tamp down the enthusiasm because it

4:41

isn't doing enough Yeah

4:43

earlier this week my colleagues at

4:45

the information reported that Amazon

4:48

and Google are quietly trying to

4:50

tamp down expectations around generous AI

4:54

and they are Basically saying

4:56

that the hype about the technology has gotten

4:59

ahead of what it can actually do for

5:01

customers at a reasonable price That's

5:05

the sound of a balloon being blown up ready

5:07

to pop I Know

5:11

somewhere out there on the beaches of Cabo

5:13

Leo's shaking his hand going but my

5:16

you know Cool chat

5:18

GPT thing for Emacs. Well

5:21

Leo they're not here this Paris I'd love

5:23

to know the background of the story Does

5:26

because they're trying to let air out of the balloon Right

5:30

before it gets poppy Was

5:33

that the journalists saying? How

5:38

do I ask this question are

5:40

the companies Putting

5:44

out an active strategy of Lowering

5:47

expectations or did the

5:49

journalists kind of just hear this among

5:51

investors and among others that they cool

5:53

a little bit Oh, I

5:56

actually just I just read this so I'm

5:58

very excited to go found this So

6:00

a lot of it appeared

6:02

to come from direct sales calls and conversations

6:04

with internal people from what I can understand

6:06

saying that they're having to do this and

6:09

they're having to make people step back and The

6:13

weirdest one was this KPMG bought access

6:15

to Microsoft co-pilot for like 40 something

6:17

thousand people But when asked why they

6:20

were like, ah don't really

6:22

know why we're using it. But you know If

6:25

our customers ask any questions, we should

6:28

really know which is just complete

6:31

Nonsense like this isn't

6:34

the balloon it that I don't even think the

6:36

balloon is being Deflated

6:38

I think these companies are trying

6:40

to quietly deflate it while keep

6:42

it publicly interesting. That's what I'm

6:44

saying Yeah, I think there's three

6:46

ways right that they're they're trying to

6:48

deflate it one is Well,

6:51

it can't do everything we said it could do Two

6:53

is well The impact won't be as profound as

6:55

we said it would be and three is the

6:57

revenue won't be as earth-shattering as we hinted It

6:59

would be Right. Yeah,

7:01

and I think the thing that people don't realize

7:04

is Media is kind of

7:06

not help with this exactly either is even

7:08

if it gets The biggest

7:10

problem isn't just that it's not doing enough.

7:12

It's that even when it gets profitable when

7:14

it helps people get profitable I should say

7:18

What if it's only doing like two to four percent more

7:21

profit? Well, how about that

7:23

because generative AI right now is

7:25

exceedingly unprofitable very very unprofitable

7:28

Lastly and all of it

7:31

is flowing back to Google and Microsoft Open

7:34

AI and an anthropic are

7:36

basically in the pocket. They both agreed to

7:38

their own versions of exclusive Contracts

7:40

with them. So by the way, this is all

7:42

the system where just Just Microsoft

7:45

invested what ten billion dollars to basically buy

7:47

ten billion dollars of revenue same deal with

7:49

Google and an anthropic for about three billion

7:52

But The other thing in there was they

7:54

mentioned this Klarna story that Klarna saw

7:56

forty million dollars of saving in there

7:58

That's not what caught. Corona said clone

8:00

A said the was there was a specific

8:03

would have nothing to bring up. Felix This

8:05

is really important and very strange Stamps not

8:07

seen this one years it was. Say.

8:10

Profits. Profits. Improvements:

8:13

An estimated drive as forty million

8:15

dollars us, the in profits, improvement

8:17

to clone us, and Twenty Twenty

8:19

Fourth. So by the way, that

8:22

does not say. Profits.

8:24

And it does not say savings. it

8:26

says estimated in Twenty Twenty Four, which

8:28

is the year where it currently in.

8:32

Some that does that France at improvement

8:34

mean I've never heard of that. It

8:36

means that it could seriously save money

8:38

somewhere that could contribute to the the

8:40

top line profit at some point in

8:42

some form to a company called plan

8:45

and is already not petite to least

8:47

profitable because their entire business bubbles is

8:49

based on people get in free money

8:51

and they're making a billion dollars. Soon

8:54

as every all six seven hundred people. and

8:57

has lame that blame that on a i

8:59

rather than medicine a. And

9:01

blamed it on a I management or

9:03

perhaps not giving every one zero percent

9:05

in press. Learns. All.

9:08

The time red red bull know puts

9:10

with no real credit checks is the

9:12

same pragma from his face since it's

9:14

just. It. All feels like

9:17

we're being Cons: On

9:19

some was interesting because we look back at the

9:21

two thousand and press. Loves it

9:23

was about using Vc money to buy

9:25

audience that wasn't really legitimately their. Yes,

9:29

So now they're using bc money to

9:31

buy what. Well.

9:34

That's the interesting saying. they're using Vc Bunny

9:36

to find a new way to make money

9:38

from their customers. It's the same

9:40

deal it was actually Texas that

9:42

back through his day in this

9:44

question because if it's Microsoft and.

9:47

google. what they're doing is investing money

9:49

in a i companies that them become dependent

9:51

on them for their mobiles to run so

9:54

all mugs of and google doing is buying

9:56

revenue they just that most of the ten

9:58

billion that was invest OpenAI was

10:01

cloud credits. Anthropic agreed

10:03

just before they took $3 billion from Google

10:05

to be exclusively using Google's

10:07

cloud and AI services. So in their

10:09

case, it's just they've created a new

10:11

money stream. It's kind of akin to

10:13

like the cloud boom in the 2010s.

10:16

So it's an old model, equity for revenue. Yes.

10:19

But in the startups case, I don't

10:24

know. I don't know what

10:26

these people, there were some really, like there was some very

10:28

niche like AI things are kind of cool,

10:30

but you see it in Snap. You

10:32

see it in these big type of Facebooks talking about

10:34

generative AI. What's it going to do for Facebook? My

10:38

biggest theory is that Google wants to replace Google

10:40

search with generative AI anyway, but that's a whole

10:42

other point. Yeah, no, no,

10:44

no, I write a whole thing about it, Jeff. You got to read

10:46

it. I

10:49

thought it was interesting in the article my

10:51

colleagues wrote, they mentioned that at

10:53

AWS, they recently gave their salespeople

10:56

kind of a reality check on

10:58

the tech. They had an analyst come

11:00

in during annual kickoff event a couple

11:02

of weeks ago and say that the industry

11:04

is at the peak of the hype cycle

11:08

and that he anticipates the hype

11:10

could veer into quote, a trough

11:12

of disillusionment in the coming years

11:14

as customers realize generative AI's limitations.

11:17

We have

11:20

a downer tone for a company presentation

11:22

about selling AI. I

11:24

have this vision that at the headquarters of

11:26

the information, there's a big dial and the

11:30

BSometer. No,

11:33

but that's a really good idea for

11:35

our interior design. It is. And

11:39

so how far on the BSometer now,

11:41

I mean, does, does, because

11:44

I'd imagine, especially the information you two had, but

11:46

the larger organization of all the wealth and

11:48

resources you have, at some point you just

11:50

all have conversations saying, is this, is this

11:52

real? Is this going anywhere? I

11:55

don't know if I want to comment on that, on

11:57

that one. I got something. So it's about some Of the. Major

12:00

decisions, but. I'd

12:02

be very curious what the information has

12:04

to save face but Busan skinner of

12:06

a I have to look for to

12:08

just lessons coverage or am I mean

12:11

I can't speak to. Her any other

12:13

individual employees lot on it, but for

12:15

me it's. I think it's

12:17

I think has been quite clear on

12:19

the podcast. I think it's quite a

12:21

lot. Of hype. It reminds me of

12:23

said the in a whole medafor. Sprays

12:25

or the crypto N N F T crazes

12:28

need is just the. Same. Cycle

12:30

of the Moments and I haven't

12:32

yet really see any evidence that

12:35

this is a revolutionary technology. It's

12:37

in the way it has been

12:39

described by assuming. This is as

12:41

have the first. Moment.

12:44

We're We're starting to see the eventual fall

12:46

of the same Leo's running down the beach

12:49

right now to give back off on the

12:51

showed scream as it's a. Leo is

12:53

pulling up a big phone and trying

12:55

to dial in to tell us in

12:57

Leo. Leo ad for Contexts has been

12:59

radicalised ever since he went on a

13:01

walk with an unnamed A. I acceleration

13:03

us to convince him that and five

13:05

to ten years A I. Will be

13:08

every thanks. Daddy.

13:10

Daycare the same with that person does come.

13:12

I was good knob joke as I mean.

13:16

I could. The.

13:18

Big difference I think between this and like

13:20

most of us and Crypto is the is

13:22

a product here for the problem is that.

13:25

As one's kind of in on the idea of

13:27

a I being big in the same way the

13:30

autumn of remember parasite only twenty tons of the

13:32

big suffers a service posts and have one Love

13:34

Daves new way that we could make money off

13:36

of software that people's you could sell Nc and

13:38

Sports. The was actually the thing that. Take.

13:41

Never really had. Consistently with the it's

13:43

price was a structured way to sell

13:45

a contract base. Things that had an

13:47

obvious bill of so soon as I

13:50

lay on it was great. I it

13:52

feels more like the cloud boom which

13:54

was good until you realize. The

13:58

same level of hype from the

14:00

Metaverse did Sting Corner which is

14:02

a I is gonna also my

14:04

everything Okay how. Well.

14:06

It will. Okay, but how will it

14:08

do that? Integrations,

14:10

And a p I awesome great but what does

14:12

that mean? I do not know. Are.

14:15

You paid thirty five million dollars to me, I work

14:17

for Mckinsey. I was just here to tell you the

14:19

eyes good but the thing is with the valley right

14:21

now. Is. A

14:24

right now I've one likes this idea because it's

14:26

a new thing is using as one can do

14:28

is using the At one can integrate and see

14:30

rate you can sink through. Tell A I apply

14:32

to a business I worked as. here's the thing

14:34

I don't like. I dunno might that. That.

14:37

Makes sense, that's it's a logical

14:39

thing to sing top. However, When.

14:42

You're really since Think about the actual technical

14:44

side. Damn. How

14:47

much of the single trying to automate. Can

14:50

be messed up because that's a big butts

14:52

messing up things is a big part Jenner

14:55

of Ice And then even if you work

14:57

out a way, this can directly help you

14:59

and integrate their. It is

15:01

expensive, It is unprofitable for the service

15:04

provider rely the way it's unreliable. But

15:06

even it's is find the boundaries and

15:08

there's a flourish Wall Street Journal articles

15:11

few weeks ago. I were

15:13

they were saying gas, amazon and Google

15:15

that old open a i'm Google are trying

15:17

to sell their generate by I solutions

15:19

and they keep running into brom of hallucinations

15:21

and then they said what could we do

15:23

to fix this scenario was like yeah

15:25

we could just tell the i not want

15:28

Serve isn't conference which then led to another

15:30

call from someone saying yeah but at some

15:32

point the I would just go other

15:34

know what.com. Con: Healthy

15:36

Man: I'm sorry, don't trust myself.

15:38

It's just so strange because. Everyone

15:42

there was this insane amount of money to

15:44

know Mark Andreessen should be lost that publicly

15:46

said the money the when it's like chart

15:48

for ai. But. Not as hard as the money

15:51

that went into flow with as Noom and I said that. For.

15:54

So much money's gone into the systems, but. Once.

15:57

the actual thing was that was the thing

16:00

Where's the thingy that I use now? Where is the

16:02

essential thingy now? I don't see you don't think

16:04

a chat GPT is the

16:07

creme de la creme I Don't

16:09

know what it's for. I would love someone to tell

16:11

me how this Richard. I am lazy I don't want

16:13

to do stuff if the computer can do the thing

16:15

for me. I'd be so happy But

16:17

I never use chat GPT for anything you

16:20

said right I've

16:22

tried I've tried to use it for exactly one

16:24

thing I was I forget what I was trying

16:26

to look up something maybe related to niche Healthcare

16:29

industry terms for a story I was working

16:31

on and I was having a really hard time summarizing

16:33

it like a human person Right eventually asked chat GPT

16:36

a number of different questions and it kind of helped

16:38

me I used to the only thing for the

16:40

second time today My best use and I

16:43

said this mission before is to summarize

16:45

strategy I

16:51

Just put it all in there and it gives you two nice paragraphs You

16:54

know and I think I think your point is something here

16:56

that Part of the presumption

16:58

and presumption that and that their vision story

17:01

is that so many people see AI as

17:03

a method to efficiency That's the mistake Because

17:06

that presumes that this general machine can do

17:08

things we do right Somewhere

17:10

in the rundown I put up Zach Seward

17:12

who's the head of AI in the newsroom at the

17:15

Times now He did a really good presentation at Southwest

17:17

Southwest, which I haven't been to for years

17:19

and had no FOMO for all these years But

17:21

I would like to have heard that but as a put it up and

17:24

it's all about Finding anomalies

17:26

in data. That's a journalistic way to look

17:28

at it But you

17:30

know analysis of data and

17:33

finding anomalies and patterns Now

17:35

it seems to work pretty well because

17:37

you're giving it to yourself for judgment,

17:39

but replacing jobs No

17:43

efficiency So you

17:45

can write more power points. That's not efficient. It's

17:48

spamming the world. I Don't know what

17:50

it comes from. I right agree and I

17:53

would love it to reflect my my date my

17:56

PR firm job. There's a lot of spreadsheets a

17:58

lot of boring documents Love anodyne copy You'd

18:00

think this would be what these things were for, but every

18:02

time I ask it to do stuff, it gives me possibly

18:05

the most half-assed work I've ever seen.

18:09

Just like you have all of, like

18:11

you're throwing several zoos, a hundred

18:14

cats and a few trees in there

18:16

just for this query, and you can't

18:18

even give me like a filled-in spreadsheet?

18:20

I can't upload a spreadsheet and have

18:22

you tell me what's in it? I

18:24

thought this was the very basic thing

18:27

this could do. It's not quite as

18:29

infuriating as crypto where there was nothing.

18:31

Yes, yes, yes. Absolutely nothing, same with

18:33

the metaverse. And I grew up writing

18:35

about MMORPGs. NFTs as

18:37

well. This

18:40

thing does something, but at the same time,

18:42

despite everyone losing their

18:45

nut over Sam Altman

18:47

conning the nonprofit

18:49

board into reinstating him with

18:51

several other bebop and rock-steady-style

18:54

cronies, everyone lost their proverbial

18:56

over that. But this company has yet

18:59

to prove itself. I

19:01

saw a journalist say, it's so beautiful, by the

19:03

way, on the night that Sam Altman was brought

19:05

back, felt sick to my son. Oh my God!

19:08

It's disgusting. It's disgraceful to be

19:10

like that, especially for like a

19:12

goblin like Sam Altman. A

19:15

nasty little man. I'm sorry,

19:17

Sam. Ugh. Nasty. Redditors,

19:20

they made nothing off the IPO, nothing off the

19:22

Reddit IPO. Sam Altman could have made a couple

19:24

hundred million dollars. It's nice to see the good

19:26

guys win, huh? Of course.

19:29

I think part of the issue is that when

19:31

people are talking about AI, they describe it, people

19:34

much like our gone

19:37

host, Leo, describe it as

19:39

a thinking, almost feeling

19:41

machine. They humanize it in a

19:44

way and project intelligence onto it

19:46

in a way that is just

19:48

completely untemped. Untethered from

19:50

reality. Paris, I use

19:53

this phrase today. I've written a

19:55

new premise

19:58

or beginning from my book, The Grouper. parenthesis

20:00

on sale now in paperback as

20:02

well. So I had to write about that. And

20:05

I called it the literate, more

20:07

in our reaction to it than the reality, but

20:09

we see it as the literate machine. Oh,

20:12

it can, it can, we can speak to it and it can speak to us. And

20:15

that's what's so shocking to humans

20:17

and media is just that it

20:21

uses our language, not always well, not or

20:23

Lord knows not accurately, but, but

20:26

it doesn't know anything. It's not, I agree.

20:29

It doesn't know what it's saying. Yeah, I'm agree. Just

20:32

to be clear, I'm agreeing with you. Like it's, it

20:35

doesn't know anything. It's why everyone's like, Oh,

20:37

well, the future versions of Sora video generator

20:39

will look so much better. Now they weren't,

20:42

they weren't, they're not going to look bad. Why?

20:44

Because of the hallucinations, which are a feature,

20:47

not a product. These things don't know if

20:49

you can't tell it that a monkey has

20:51

four legs. Sorry, two arms, two

20:53

legs. I'm just regular stupid. I'm meant to

20:55

be able to learn things. It's not making

20:57

any claims

21:02

about being intelligent here. I think no one

21:04

has that mistake, but it's going to keep

21:06

making these errors because even if it makes

21:08

it, and this is the thing that's helpful

21:11

with the AI generated video, especially it's, we

21:13

don't realize how perfect the world around

21:15

us is and that the world around

21:17

us is we interpret it through basic

21:19

semiotics, science of science. And

21:22

we know what a thing like, we know what a

21:24

monkey looks like. If a monkey had three horns, we'd

21:26

be like, what the hell? That's a strange looking monkey.

21:28

But we'd know it was a monkey based on some

21:30

of the features. But when

21:32

you get down to basic things like

21:34

how a human being walks, how something

21:36

we know moves, and you try and

21:38

do that and you make even little

21:40

mistakes, people know that people aren't stupid.

21:42

And I actually think that that's a

21:44

big thing that generative AI people aren't

21:46

realizing people are not done. They're going

21:49

to see, they see it's not do

21:51

anything. And this is also one of

21:53

the problems I have with the claim

21:55

that Sora is going to replace

21:57

filmmaking and generative AI is going

21:59

to. completely wipe out

22:02

the filmmaking industry as we know it. Filmmaking

22:05

is a... To be a director, there's a very

22:07

high bar. Directors of

22:10

major movies are particular when it

22:12

comes to the shots that they

22:14

have. You're not going to

22:16

be able to describe

22:18

a very specific shot

22:21

as well as get the sort of

22:24

performance from your AI actors with the

22:26

same ease you would just

22:28

standing in the room with two humans that

22:30

you have hired for a specific role. It's

22:34

going to be more complicated. No, but I think TikTok,

22:36

imagine TikTok with Sora, right? There's

22:38

stuff that in

22:40

lay hands can

22:42

do more than it could. I think that

22:44

all of this, it's not a business, but I think

22:46

it had all been put out there as a creativity machine.

22:50

And that's all it does is make stuff up

22:53

and you can make stuff up with it and it's going to be good

22:55

or bad and you can have fun with it. That's

22:57

cool. No problem with that. I think there's

23:00

a lot there, but it's this efficiency

23:02

machine that's going to replace search

23:05

and replace people with jobs and

23:07

be smarter than us is such... Yes, I

23:09

got to be on the Dan Libotard podcast

23:12

this morning. French Ambassador. Does

23:14

that mean you've done now three podcasts today?

23:16

In one day. In one day,

23:18

yes. I am. I did four last

23:20

Monday. I've been doing the promo tour for this

23:22

damn show. Anybody can answer.

23:25

So I use language there that I'm not allowed to use here,

23:28

but they started asking, well, what do you

23:30

know? I had all the big

23:32

swinging Richard of the

23:35

AI boys. I quite like that. Yeah. Yeah.

23:40

And I think the reason that

23:42

I think that Google will do such with

23:45

this is not because it's a good

23:47

idea, but because I think that

23:49

Google wants to... I

23:52

think Google has made a big mess. They made a big

23:54

stinky. They've turned the web into

23:56

a big hole with their allowance of SEO.

23:59

And the thing is... Starting aside, AI spam, by the

24:01

way, Google is at fault because they are the ones

24:03

that have catered to the freaks, the SEO industry, and

24:05

the people in charge of the media are the ones

24:08

that have been trying to dance the Google song. I

24:12

think Google will do a generative bot that

24:14

turns them into a form of ISP. I

24:16

think they want Google search to be much

24:18

more controlled. I think they want it to

24:20

chew up and regurgitate the internet. To be

24:22

clear, I think this is a horrible product,

24:24

but I think that Google and Sandal Pichai

24:26

in particular have just become

24:28

so scummy and so

24:30

actively abusive of the people on

24:32

the internet that just I think that

24:35

this is their eventual endgame. I don't think it's good.

24:37

I don't think it's efficient. I don't

24:39

think it's even right most of the time. But

24:41

guess what? Google need more money.

24:44

Google must show growth. The raw economy

24:46

must survive. We must always have more

24:48

growth in tech, even if it's unsustainable,

24:50

even if it's horrifying for human beings

24:52

and indeed the human capital working at

24:54

these companies. Well, that's

24:56

depressing. Yeah. I

24:59

mean, the world is a vampire. I

25:01

mean, it's part of what you get when you

25:03

have to have companies that are always showing quarterly

25:05

growth. That's right. Speaking

25:09

of AI,

25:12

today European lawmakers approved what's

25:14

being called the world's most

25:16

comprehensive legislation yet on AI, setting

25:18

out sweeping rules for developers of

25:20

AI systems and new restrictions on

25:22

how the tech can be used. It's

25:25

called the AI Act, and the rules

25:27

are set to take effect gradually over

25:29

the next several years and apply to

25:31

all AI products sold in the EU market,

25:33

regardless of where they were developed. Some

25:36

of the notable parts of this include

25:39

their prohibitions in the legislation, including

25:42

bans on the use of emotional recognition,

25:44

AI in schools in the workplace,

25:47

and on untargeted scraping of

25:49

images for facial recognition databases.

25:53

The new rules will eventually require providers

25:55

of general purpose AI models like Chat2PT

25:57

to have up-to-date

26:00

technical documentation and publish a review

26:03

of the new model and the new model.

26:07

And makers of the most powerful AI models,

26:09

which is what the EU is deemed to

26:11

have kind of systemic risk, will be required

26:13

to put those models through state-of-the-art

26:17

safety evaluations and notify regulators

26:19

of any serious incidents

26:21

that occur with the models and

26:23

implement, I guess, mitigations for

26:26

potential risks or cybersecurity

26:28

production. What

26:30

do you guys think about this? I think the EU

26:32

is the only force that could stop the tech industry.

26:35

I don't know, I haven't read the fundamental

26:37

of the bill, but I do think regulation

26:39

is necessary here. And I think in

26:41

particular they need to start saying how these

26:43

models are trained and what they're trained on,

26:45

because it's pretty hard to make

26:48

these things forget. You can't really do it, you just have

26:51

to revert the training data. And I think

26:54

that, I'm surprised Sam Altman hasn't

26:56

created the anti-Butlerian

26:58

Jihad Europe for

27:00

this, because this is anything that

27:02

makes OpenAI responsible for their

27:04

training data is fatal for that company.

27:07

There's a reason that Mark Andreessen was freaking

27:10

out about the idea of AI companies having

27:12

to abide by copyright. Same

27:14

reason that OpenAI is so freaked out by them, why

27:16

they're paying people, because if there's anything

27:18

that sets precedent here, they're screwed,

27:22

you can't unring the

27:24

bell of training, they will have to probably

27:26

start again to make it they can't really

27:28

prove that it's not there anymore.

27:30

And if you look at Sora, a great deal

27:32

of it looks very very similar to a lot

27:34

of Shutterstock stuff too. So, I

27:38

think that, I don't know if they have a

27:41

relationship with them, but I'm pretty sure that all

27:43

of these models are based on some form of

27:45

plagiarism. The E-Rocks, I love that they're doing

27:48

stuff like this, it's better than the

27:50

nothing we're getting over here. So

27:53

I'll disagree. I'm

27:56

finally that, I've finally summoned

27:58

someone on this. podcast who agrees

28:01

with me. I think

28:03

you want to go first? You go first. No,

28:06

you go. I'll hear your rebuttal. All right. All

28:08

right. I think the legislation is better than it

28:10

could have been. They were talking about outlawing

28:13

open source, which would be disastrous. It

28:15

would be regulatory capture. They were talking

28:17

about, I think

28:20

that the reason OpenAI is not screaming is

28:22

because it is regulatory capture. The big companies

28:24

with the big money will be able to

28:26

deal with this legislation and regulation and newcomers

28:31

won't. And so that's why you

28:33

always hear Microsoft just

28:35

saying, regulate us. Yes, please. And that's

28:39

what Zuckerberg says too. And that's what

28:42

OpenAI says as well. So I

28:44

think there's issues there. I also think

28:46

that Europe constantly says, we're ahead on

28:48

regulation, but nothing else. And then they

28:50

come back around and they whine, why

28:52

don't we have our innovative companies? Because

28:55

they don't invest, they don't do it. And

28:57

they regulate reflexively. The problem, I think, in

28:59

the end, I'm fine. We have some regulations.

29:01

The regulation is not bad. But

29:03

I also put it on the run down. He says begrudgingly.

29:06

Yes, it's a bit lower. That's my view. But

29:08

I don't think it's going to do much. And

29:11

I think in its later stages, it's a

29:14

very flimsy negligee

29:17

on a, I'll stop that metaphor.

29:19

Jeff immediately retreats. It's his third

29:22

podcast of the day and it's

29:24

gotten too much for

29:32

him. So I put up

29:34

this thing from AI Snake Oil, the Arvind

29:38

Narayanan and Sayash Kapoor, which

29:40

makes the argument that I've

29:42

been on the show before,

29:44

that guardrails are impossible at

29:46

the model level. They say

29:48

that AI safety is not a

29:51

characteristic of models. It's a general

29:53

interest machine. It's like saying Gutenberg,

29:55

you're responsible for everything coming off

29:57

the press. It's like expecting Microsoft.

29:59

to tell us as we're using

30:01

word, no, you can't type that.

30:03

I'm programmed to not let you type that. You can't do that.

30:06

Obviously absurd. And so

30:08

putting the responsibility level in the model

30:12

isn't going to be effective. And so

30:15

it's a veil that makes people

30:17

think we're doing something and we're

30:19

gonna protect ourselves from this dangerous

30:21

technology. The problem I have with

30:23

it is that. No, we should be honest

30:25

about it that you cannot protect

30:27

yourself from this technology and the bad things

30:29

that people will try to do with it.

30:32

Now deal with that. The thing

30:34

I think that is kind of interesting about

30:36

this legislation is

30:38

it's kind of tiered. Originally,

30:41

I believe it was set

30:44

to exclusively focus on developers

30:46

of what they call high-risk

30:49

AI systems, which is

30:51

kind of a catch-all term that I

30:54

guess, I'm gonna break this

30:56

down from a really interesting website.

30:59

I think the EU put together

31:01

called artificialintelligenceact.eu that has kind of

31:03

what looks like, it's almost an AI-generated

31:05

summary of the act itself, but I'm

31:07

guessing it probably isn't. But in

31:09

their prohibited AI systems list, you've

31:11

got a couple of tiers. One is prohibited AI systems,

31:14

where it's like AI can't be used for these sort

31:16

of things, which is compiling

31:18

facial recognition databases, inferring

31:20

emotions, social scoring,

31:23

stuff like that. But

31:25

then most of

31:27

what this act is

31:29

targeting is developers of

31:31

so-called high-risk AI systems,

31:33

which is if

31:36

they profile individuals or include

31:38

automated processing of personal data

31:41

to assess various aspects of

31:43

a person's life, like work performance,

31:45

economic situation, or health, or

31:47

if it is somehow, I guess,

31:49

related to EU law, if

31:52

it is like the government is using something related to

31:54

the system, which I think is kind of an important

31:56

distinction. The

32:00

vast majority of these regulations

32:02

are focused on these type

32:05

of AI, like developers

32:07

and companies, which is a very specific

32:09

subset. For like the general AI systems,

32:12

like a chat GPT or something, that

32:15

is the regulations being

32:18

imposed here are of a much lower

32:20

standard. It's like they've got

32:22

to provide technical documentation and

32:25

whatnot. And for free. Which

32:27

is why. Because I know I think Leo

32:29

also often talks about how this

32:31

is going to squash free, like

32:34

open AI initiatives for like

32:36

free and open license like

32:38

AI models. They only need

32:40

to comply with copyright restrictions

32:42

and publish the training data summary, which

32:45

I think is fair. Also

32:47

I think at this point you should realize,

32:50

Leo, you're not here so of course my

32:52

perfect favorite argument where the person cannot respond.

32:55

But very basic thing here is if you're

32:57

worried about little AI models, you should already

33:00

have realized they've already failed

33:02

their screwed. These massive deals between

33:04

open AI and and

33:06

Tropic and the major cloud providers and

33:09

the fact that they can afford to buy

33:11

these massive data sets, $60 million of Reddit,

33:13

for example, with Google. The

33:15

walls already been lost. The little models can never

33:17

train at the scale. They

33:19

never will be able to. But they do

33:22

need to. Well kind

33:24

of. The question has to be addressed. The

33:26

argument is that that was all big swinging.

33:29

Richard, you didn't need to have the biggest

33:31

models. That's Sam Altman who's saying

33:33

that. Sam Altman is the one pushing back

33:35

now saying, oh yeah, you don't need big

33:37

models anymore. You need a small model. Yann

33:39

LeCun says the same thing. He

33:43

is just one is

33:45

just an extremely annoying character and

33:47

I disagree. Nevertheless, putting that aside,

33:50

otherwise I'll just spend the whole time getting mad at them. It's

33:53

just. Wait a second. Wait, wait, wait.

33:55

Who's the most irritating person in AI? I want

33:57

to hear each of you answer that. Oh,

34:00

what's his name? The

34:05

one that said there would be a Bitcoin virus. Could

34:09

be any number of people. I guess my

34:12

answer is Elon Musk, even though I'd argue

34:14

that he's not in AI. That's a cheat.

34:17

It's an easy answer. It's

34:20

the easiest answer and just because Grok, sometimes

34:24

I'll just be alone at

34:26

home thinking, should be thinking about normal things

34:28

related to my life. And I'll just remember

34:31

the existence of Grok AI. Does

34:33

it exist? Here's a question. Does Grok really exist? Yes. Yeah?

34:38

Grok exists on X slash rate my new

34:40

stop is, the Elon Musk website. And it's

34:42

great because every time you see a screenshot

34:44

of it burning someone, like trying to roast

34:46

them, it does the same thing. It says,

34:48

oh boy, where do I start? It's so

34:50

bad. That's how you

34:52

know it's a sick burn. That's how

34:55

you know it's written by Elon Musk. Yeah.

34:58

It's just like, this is extremely,

35:00

this is a terrible insult. I'm

35:02

going to say, oh boy, where

35:04

do I start here? Yeah.

35:07

So I'm trying to find this

35:09

guy because he was a less wrong guy and he

35:12

pops up. He's like an AI. God damn

35:14

it. I'm going to be thinking about this for a while.

35:16

I think Yarn's a pretty good choice, but I think Sam

35:18

Altman. I truly think

35:20

Sam Altman is the most annoying because he gets

35:23

away with it. He also does not sound eloquent.

35:25

I don't know if you heard Sam Altman speak,

35:27

but for the smartest guy in the bloody room,

35:29

he sounds pretty dumb. I

35:32

don't even mean like Dollyn

35:34

A just not a good public speaker thing.

35:37

He's just doesn't, oh, he

35:39

kind of seems like Chad GPT. Just kind of

35:42

mediocre. He's like, yeah, we'll get big in

35:44

like four years. And I

35:46

want everyone in cnbc.com immediately just

35:48

copy paste the transcript, publish.

35:51

And it's just, it

35:53

frustrates me. Sam Altman frustrates me a great

35:56

deal. He's fallen. That man has

35:58

failed upwards like seven times. I

36:01

mean it is still mind-boggling

36:03

to me that he was ousted

36:06

from the board, ends up getting

36:08

reinstated and accumulates even more power.

36:11

We still don't know why. We still do

36:13

not know why. No, we don't. Nope. We

36:16

haven't. It could be to do with his sister. We'll

36:18

never know. Could be to do with any number of

36:21

things. Well hey, maybe it was the fact that he

36:23

wanted to turn the non-profit wing of OpenAI into a

36:25

profitable thing. And you put Larry

36:27

Summers on the board, so no worries there.

36:32

Yeah, this week Bloomberg reported that...

36:36

Oh sorry, Eliza Yudowski. Eliza

36:39

Yudowski. Elizaziz. He's

36:41

very annoying. Because

36:44

not only that, he is the worst, worst,

36:46

worst of the doomers. In

36:49

the Journal of Moral Panic, otherwise known

36:51

as Time Magazine, that's

36:53

where he writes his screen. Hey, Moral Panic, drink.

36:56

Drink. I

36:59

think Gutenberg deserves a beer and Moral

37:01

Panic line, is my view. Or

37:03

vodka. Go to Diet Coke. So

37:07

Yudowski is the one who's out there screaming

37:09

that paperclips will kill us all. Yeah. And

37:12

that he's the worst of the doomers. He is

37:14

the worst of the doomers. He's

37:17

in the New Yorker story with 99.9%

37:20

that it's going to destroy us and has to stop

37:23

us. And he's just... He would love that. He's

37:25

a big fan of

37:27

Rocco's Basilisk, also known

37:30

as the Weenies version of Pascal's

37:32

Wager. What

37:34

is Rocco's Basilisk? Okay, this is

37:37

beyond our podcast score here. So

37:39

Rocco's Basilisk is this thing that

37:41

we need to start building or

37:44

working on or pleasing a theoretical

37:46

machine god before it comes into

37:48

existence. Otherwise it

37:50

will punish us if we do not

37:52

do so. It is such a dumbass story.

37:54

It's exactly the kind of dumb

37:57

guy's intellectual exercise. Like, what

37:59

if... computer was scary.

38:01

What if the computer was mad at me? Oh

38:04

no, I'm so smart for thinking of this.

38:07

It's a recursive intellectual exercise. I do

38:09

think it's quite bold to reinvent

38:12

religion in the year

38:15

of our glory. It's

38:18

the same thing. H.M. and I brought

38:21

Cosbascala's controversial for a few reasons.

38:25

It relies on a lot of speculation about the

38:27

nature of future AI and its motivations as if

38:29

it has any motivations. The idea

38:31

of AI punishing people in the past

38:33

for not helping create it is seen

38:35

by many as illogical. Yeah,

38:38

yeah, yeah. But it's

38:40

exactly what I'm traveling to. I

38:42

didn't know that. Wait, I'm traveling

38:44

now? I'm into it. It sounds like travel. That's

38:46

fun. What a big test for

38:49

you. Yeah, the time travel into the tea.

38:51

Oh God. Oh, that's really smart. I

38:55

think we need to add more letters

38:57

into test grill. I think we should

38:59

make it more complicated, ideally. Yeah, I

39:01

challenge while Leo's gone is to get test grill

39:03

of the title of the show today. That'll really

39:05

Oh, he would fly back from Cabo to stop

39:07

it in the test grills. Yeah,

39:10

that. So wait, wait,

39:12

wait, wait. So, so Rick, you're Kowski. I didn't

39:14

know this. He's also the co-founder of less wrong.

39:17

Yes. What is that? Right place. If you

39:19

want to find a bunch of libertarian guys barely

39:22

covering up their racism and

39:25

talking about subjects with less articulation

39:28

and a Redditor or a stack

39:30

overflow member and

39:33

a little less anger than

39:35

the average hack and use user. It's a

39:37

very useless place to go. If you

39:39

want to meet a bunch of people that you'll never want

39:41

to meet in real life. Wow.

39:44

Yeah. On the Wikipedia

39:46

page for less wrong.

39:49

The first subhead under

39:51

history is Rocco's Basilisk.

39:53

The biggest thing they've

39:55

created is a

39:57

version of Pascal's way. What's

40:02

great is you go on there now,

40:04

let's read some of these titles. Meta-honesty,

40:06

Fermigoth, Honesty's Edge Cases, oh my god

40:08

go outside! These

40:11

people have never touched grass in their lives.

40:15

There's another one that says no one in my

40:17

org puts money into their pension. Notes

40:19

from the prompt factory, there is

40:22

way too much serendipity. Okay.

40:26

Jesus. Here's the problem.

40:30

These people are getting huge amounts of money. They're

40:34

using money on the major

40:36

college campuses to do clubs and fellowships

40:38

and classes around this crap. They're

40:42

getting the year of legislators

40:47

and they have the year of media. And they're

40:49

idiots! Honestly

40:53

it's like the J. Rosen scam, but for

40:56

AI it's kind of cool. Just

40:58

like you do. Yeah,

41:00

I was going to say, you're talking to

41:02

a... You've got a J. Rosen adjacent

41:04

here. And

41:08

why you... J.

41:10

Rosen is at least an academic. At

41:13

least he has academic cred. At

41:16

least he has spoken to reporters and knows

41:18

them. These people are just like, yeah, thought

41:21

about it really hard, wouldn't it be scary

41:23

if the computer did this? And everyone goes,

41:25

oh, god damn, holy crap. What

41:28

if the computer was smart? But if you look

41:30

at the wider media, they're buying a not much

41:33

more sophisticated story from Sam Altman. Sam

41:35

Altman's like, yes, can automate everything. It

41:37

could be super smart and automate all

41:39

the stuff. He's about that vague. I

41:43

love these... They're

41:45

just wonderful. I want t-shirts for every

41:47

one of them. How

41:49

could I have thought that faster? The

41:52

parable of the fallen pendulum. What

41:55

could a policy banning AGI look like?

41:57

There isn't any AGI, it's not going

41:59

to happen. give it up. Notes

42:01

from a prompt factory. Oh my

42:04

lord, the serendipity one is a beautiful

42:06

one. Yeah, there's one about the COVID-19

42:08

pandemic of course because these guys these

42:11

guys love to think about stuff and

42:13

get really close to saying something racist.

42:16

Highlights from Lex Friedman's interview

42:18

of Yellow Lacoon. Wow.

42:24

What a... How to have

42:26

Holly genetically screened children. Lex

42:28

Friedman is so funny though.

42:31

Lex Friedman is

42:33

awful. Because he buys all the

42:35

BS. Remind me who Lex Friedman

42:37

is. So imagine if

42:39

you will, a very pallid boring

42:42

man that nevertheless has become one of

42:44

the most successful tech podcasters. I will

42:46

depose him. He is

42:48

an anime. Oh that's true, I have seen you

42:50

tweeting about better offline rising in the ranks. You've

42:53

got to take over Lex's spot. I will take

42:55

him on by doing a good job unlike him.

42:57

He does these like two three hour long interviews

42:59

with people. He gets got Elon Musk, he got

43:01

Jeff Bezos and he gets like Pajale of course.

43:04

I was gonna say how many posts of the All In podcast

43:06

do you think you could take in a fight at? How

43:10

many members or posts? How

43:13

many hosts or I guess... How many hosts? Couldn't

43:15

take a shot. He's a ...

43:23

Oh absolutely. Those guys you know those guys make

43:25

some like depressing looking food. But Lex Friedman what

43:27

he does is he does these three hour long

43:29

interviews that are the most... He

43:31

talks like this the entire time.

43:34

He is also just one of the least articulate

43:36

men to ever walk the South. I want to

43:38

read... Why do you like him?

43:41

Like I don't know why people like him. I

43:43

don't know why people like him at all but

43:45

let me read you this question. The transcript of

43:47

the question he asked Jeff Bezos and

43:50

this is the exact way he said it. You went

43:52

to Princeton with aspirations to be

43:54

a theoretical physicist. What attracted you to

43:56

physics and why did you change your

43:58

mind and not become... Why why

44:00

you're not Jeff Bezos the famous

44:02

theoretical physicist. Oh

44:05

no brother man

44:07

brother man and to be clear

44:09

he's reading this question. Oh, yeah,

44:11

yeah, slowly slowly slow. Is

44:14

he's a member of the slow

44:16

talkers of

44:21

America. Maybe he's popular. Maybe

44:25

comes from the culture of people listening to

44:27

podcast and like 1.5 speed. I

44:31

think it's that and yeah, you could probably listen to

44:33

this like 8 speed and get these beans flow. So

44:35

that makes you sound smarter. I

44:38

think he's speaking slowly because that's how fast the

44:40

information comes out. Let

44:42

me read you the description of his

44:45

YouTube channel. Lex Friedman

44:47

conversations about science technology

44:49

history philosophy and the nature

44:52

of intelligence consciousness. Love and

44:54

power. All right,

44:56

did he

44:59

have a young man. We're gonna

45:01

have to go to an ad break. Let's say

45:03

your last Lex Friedman

45:05

thoughts in any any final thoughts

45:07

guys. Michael play 30

45:09

seconds of him interviewing Tucker Carlson.

45:11

No, I don't think we can

45:13

do that. Actually, just just to

45:15

tie this up. He does these

45:18

very long-winded things where he gives

45:20

platforms to incredulous freaks like the

45:22

jolly like talker. He

45:24

lets Elon Musk or yes, they're well.

45:26

There's very very big problems with the

45:29

immigration and the works and

45:32

he's so dull, but he's also a

45:35

he's basically a right-wing guy. He's

45:37

basically just another right-wing tech guy and

45:39

that is a problem. The two of

45:41

the biggest podcasts are center right right

45:43

wing. That sucks. That's bad. Anyway enjoy

45:45

the outbreak. Enjoy the ad

45:48

this episode of this week in

45:50

Google brought to you by Rocket money.

45:53

By the way, thank you very much Paris Martin. Oh

45:55

for filling in for me while I'm on vacation. I did

45:58

want to come back and tell you about Rockin' money

46:00

because they have saved me enough to pay

46:02

for this vacation. Come to think of it,

46:05

did you know nearly 75% of people

46:07

have subscriptions? They have forgotten

46:09

about subscriptions. Things

46:12

like fitness apps and streaming services and

46:14

delivery services. When I started using

46:16

Rocket Money, I couldn't believe how

46:18

much I was paying for that I'd completely forgotten

46:20

about. I found campaign

46:22

contributions from 2022 that I was

46:25

still paying monthly and I cancelled

46:27

it with Rocket Money. I

46:30

found just the other day a

46:32

WordPress subscription I'd forgotten about,

46:34

got $300 back thanks to Rocket Money.

46:37

Rocket Money saves me money on

46:40

all those subscriptions I've completely forgotten

46:42

about. It's a personal finance app

46:44

that finds and cancels your unwanted

46:46

subscriptions but it does a whole

46:48

lot more. Rocket Money also monitors your spending, helps

46:50

you lower your bills, you can grow savings. It's

46:53

really an easy to use app. I can see

46:55

all my subscriptions in one place and if I

46:57

see something I don't want, Rocket

46:59

Money can help me cancel it with

47:01

a few taps. They'll deal with

47:04

customer service for you. Awesome!

47:07

Rocket Money, more than 5 million

47:09

users, it saved a total of

47:11

half a billion dollars in cancelled

47:13

subscriptions. On average it

47:15

saves members $740 a year when you use

47:17

all the app's features. It

47:19

saved me more than that. I love Rocket Money. Stop

47:22

wasting money on things you don't use. Build

47:24

your unwanted subscriptions by

47:27

going to

47:29

rocketmoney.com/twig. That's

47:31

rocketmoney.com/ twig. You got

47:33

to try this thing. It's awesome. rocketmoney.com

47:38

slash twig. Now back

47:40

to this week at Google Paris. You're doing a great

47:42

job. Thanks for filling in for me. Thanks

47:46

Leo. This

47:49

morning the House approved a bill

47:51

to force TikTok to either

47:54

cut ties with ByteDance or be banned in

47:56

the US. The bill is

47:58

now moving to the Senate where it's... The prospects

48:00

are less clear. Its

48:03

passage to the House was crazy fast.

48:05

It was only introduced last week and

48:08

passed by a House panel on a

48:10

unanimous vote. TikTok

48:12

has tried to lobby against the bill, asking

48:14

users whenever you use the app, you've probably

48:16

seen the pop-up come up, it's

48:19

asking users to call their representative to

48:21

protest. Apparently that's already

48:23

alienated some Congress people, who

48:25

would have guessed. Biden has said that

48:28

he'll sign it if it reaches his desk, it's

48:30

unclear if that will happen. If

48:33

it gets there, implementation will likely be delayed

48:35

because TikTok's expected a file suit to block

48:38

it. Meanwhile, China's also expected

48:40

to block ByteDance's attempts to

48:42

sell TikTok if it gets to this. Jeff,

48:46

I think before the show, you were saying you're

48:48

gonna get mad about this, so get mad, yeah.

48:52

I actually heard someone on the message we

48:54

see this morning say,

48:57

well, yeah, it was

48:59

Mikey Sherrill from New Jersey, who

49:02

said, TikTok is

49:04

against free speech, so what are we gonna do? We're gonna

49:06

take away the free speech. The

49:09

House logic here is just awful. It's

49:11

moral panic, drink everybody, meeting

49:14

political nihilism. It

49:18

is ridiculous, as Morning Joe goes on,

49:20

about all the Chinese propaganda, I don't

49:22

see any Chinese propaganda there, and give

49:24

some credit to citizens that in a

49:26

democracy, we can hear it and laugh

49:29

at it and get rid of it.

49:31

We're not sheeple. And

49:33

a lot of people and a lot of

49:35

voices who otherwise are not heard in big

49:37

old white mass media now have their stage

49:39

and they're gonna take it away, and so

49:41

it pisses me off majorly.

49:44

Could we get the moral panic clip to

49:46

play, folks? Nice.

49:50

Oh, no. I got a bad feeling about this. Where's the audio?

49:54

There we go. Thank you. It's a moral panic,

49:56

everybody. Cue the moral panic. One

50:00

of the most interesting things about this is, I

50:02

mean, this isn't the first time TikTok has,

50:06

the US has tried to ban TikTok,

50:08

but apparently TikTok was blindsided

50:11

by the bill. The

50:13

Wall Street Journal reported that just two

50:15

weeks ago executives from TikTok's US operations

50:18

flew to their company's headquarters

50:20

in Singapore to meet with

50:22

their bosses and tell them everything was A-OK

50:24

with the app and that they totally weren't

50:27

in danger of being banned. Obviously

50:30

a big miscalculation. Inside

50:33

TikTok, some leaders were apparently aware that

50:35

lawmakers were working on legislation, but they

50:38

didn't expect it would win so much

50:40

support so quickly. Some of the people

50:42

familiar with the matter told the Wall

50:44

Street Journal. What

50:46

do you think young voters are going to say about this? They're

50:49

not going to understand the nuances of it and they're going

50:51

to get mad they can't watch TikTok. I

50:54

mean, yeah, they're probably going to blame the

50:56

Biden administration and be quite pissed. I

51:00

thought it was interesting. My colleague, Kaya

51:02

Uryev, who writes our

51:05

Creator Economy newsletter, was at South by

51:07

Southwest this week and kind

51:09

of did a little scene report from,

51:12

you know, the various panels

51:14

with creators. Apparently, creators are totally

51:17

unfazed with this. A

51:20

TikToker named Remy had

51:23

said at a panel, a lot of creators

51:25

are probably ignoring this right now because we've

51:27

heard it so many times and it's possible

51:30

that it could happen. Which I mean,

51:32

I think that that's probably the approach that most

51:34

users in the app are taking. Like this is

51:36

not the first time. It's happened many times over

51:38

the last four years that TikTok, the, oh, TikTok

51:41

is going to be banned, panicked. I

51:43

don't know. Do you think that they actually have

51:46

the political clout to do this this time? It

51:49

seems unlikely that the Senate would approve this, but

51:51

I also wouldn't have expected the House to unanimously push

51:54

this through. I think there's momentum behind

51:56

it. I'm not a policy guy, but

51:58

clear the House surprisingly. surprisingly fast, it popped

52:01

up surprisingly fast, I

52:03

just don't know if any of the old people

52:05

in the Senate really care. I think they'll be

52:07

like, I don't like China. So damn,

52:12

sure. Who gives a damn? I'd

52:14

be against child porn, be against China

52:16

on that list. I mean, being

52:19

against child porn is the unanimously agreed

52:21

upon thing I would hope. But the

52:24

China one is kind of

52:26

like, that one is beginning to get, I

52:28

mean, the Tom Cotton part of those hearings, where he

52:30

was just repeatedly racist.

52:33

It's that one guy who was very much

52:35

not, wasn't he like, I'm from Singapore. What

52:38

are you talking about? Yeah, the fight dance

52:40

executive who he was trying to accuse of

52:43

being a Chinese guy. What

52:46

are you talking about? Where are you from?

52:48

Where are you really from? Where are you

52:50

really from? I think the

52:52

journalists should have gone all race purity

52:55

on Tom Cotton, they

53:00

should have like traced his background back to another

53:03

country that wasn't America just to mess with

53:05

him, because that is the same thing. It's

53:07

disgraceful that that happened. And this bill was

53:09

kind of disgraceful, too. It doesn't seem to

53:11

actually be fixing any problems. It seems to

53:13

be setting up standards to

53:15

punish other apps. While also,

53:18

yeah, if the CCP actually has access

53:20

to TikTok data, that's incredibly dangerous. Like I

53:22

think the people need to realize

53:25

that. But for

53:27

the average citizen, how dangerous is it? I

53:31

don't think the harm, where's the danger? I

53:33

don't think it's the danger now. It's the danger that could

53:35

be there. But at the same time,

53:37

when they come in, they'd been really, but

53:39

that's kind of like, it is a good

53:41

point. Like, what can they do with that

53:43

data? What does that data say? Now, the

53:45

example given by one of the witnesses in

53:48

the Biden's hearings was that the CCP used

53:50

it to chase down dissidents in Hong Kong.

53:53

And that is disgraceful, disgusting, and

53:56

could be bad. Like, I mean, doesn't

53:58

China have... Every Chinese student

54:00

I've ever had is extremely savvy

54:03

about this and very careful. It

54:05

doesn't matter what the platform is. It doesn't matter anything else.

54:07

And they're not going to come on and say, I'm going

54:09

to criticize the regime on TikTok

54:11

and think I'm safe. They know better. Does

54:13

that mean that we shouldn't protect them? Yeah.

54:16

But I'm saying that there is presumption

54:18

that that's how the Chinese government is

54:20

going to use it is naive. But

54:23

it's how they've been potentially using it

54:25

already. Well, we know we've had back

54:27

in the day, one of the first horrible things that

54:29

happened is Yahoo died in 2018. A

54:31

dissident. And he went to jail for

54:33

many, many years. I

54:36

just think that this is not the

54:38

way to do it, but something has to be done. But

54:40

I am not aware of what that

54:42

would be. And the speed at which this has

54:44

been pushed through and the way it's being pushed

54:47

through is harmful for America more than it is

54:49

harmful for China or TikTok. And

54:51

also Coldblur over at 404 Media

54:53

had an interesting take on it today. I

54:56

mean, yeah, the goat. That

54:58

the US, the headline of the article was the

55:00

US wants to ban TikTok for the sins of

55:02

every social media company. And I

55:04

think that that is a really relevant point here

55:07

is that we wouldn't be in this place where

55:09

we are today with this

55:12

really coalesced animus towards TikTok if

55:14

it wasn't incredibly politically expedient for

55:16

politicians on both sides of the

55:18

aisle here in the US to

55:21

be rallying their forces against big

55:23

tech generally. I think we see

55:25

this with every of these nonsense

55:27

social media hearings, one of which

55:30

our very own Jeff Jarvis participated

55:32

in it. I

55:35

don't really know how we go back from this.

55:38

I also think that perhaps I don't know. Here's

55:40

an idea. So there was not there's an executive

55:42

order that was going out that was going to

55:44

restrict the flow of sensitive data through intermediaries like

55:46

data brokers to foreign countries

55:48

like China. America's fine,

55:50

though. No need to

55:53

protect Americans data from American companies.

55:55

That's what's really bothering me about

55:58

this. There are things. The

56:00

Tick Tock has done or might

56:02

do or theoretically could do. The.

56:05

American companies do today. Yeah

56:07

I know my satisfying as. Canada,

56:10

Canada wants to stop the data from coming

56:12

here. The you want to stop so different

56:14

from here for exactly what you're say. That

56:16

is because America's best Us. And

56:19

is just like how about matter.

56:21

Why? Did we make Mark Zuckerberg

56:23

divest from Mehta? Before. His

56:25

comments or up his place when Cambridge

56:28

Analytics santa. Why was they know that

56:30

allow? Don't get me that occasionally with

56:32

a bunch of Bs regards. I mean

56:35

yes. Space protests. Facebook has. Even when

56:37

you put their site, Facebook has stopped

56:39

providing a service. Facebook has stopped. Facebook.

56:42

Has taken advantage of the people using the

56:44

platform. Skyn. A lot

56:46

of we. The end result Mehta. Have to be

56:49

salivating at the thought of tic toc been banned.

56:51

Someone's. Is not months ahead or for.

56:53

Medicine. Oh no, it's a precedent that there's

56:55

Well then I can be very happy about Isis How

56:57

know and they'll let me. I did that. Events has

57:00

has had like a county. Who is? Their

57:02

current focus is trying to get

57:04

users to. Spend a

57:06

lot more time on Instagram, really feel free to

57:08

don't really a tick tock phone and or wanted

57:10

avoid Dell. That's true. It's

57:13

just it's frustrating because this kind

57:15

of shows that this not real

57:17

concerned about the like like Tyson

57:19

Caboose thanks. V. The things that.

57:21

Most. Social Networks there. And.

57:26

It. Sucks that the actual problem isn't

57:28

that the problem is China. Just

57:30

of kind of vague xenophobia. Personal are

57:32

not cent of army of them Ship

57:34

was what took office Racism. Xenophobia. Just

57:36

call it what it is. There.

57:38

Is a degree here just to

57:40

steer of shine off ruled as

57:43

the Scary China it's. It's.

57:45

Disgraceful. And also it

57:47

isn't protecting any one. No one is safer

57:49

because of this. All. It's

57:51

going to do is allow Opie goddamn

57:54

contact. Who. oversaw the destruction

57:56

of activision blizzard who oversaw

57:59

multiple allegations of sexual harassment, a

58:01

horrifying culture and activation. Now he's saying,

58:03

oh, I might buy TikTok. That's

58:05

a better situation. You want to trust that greasy

58:08

pervert? These

58:11

are our saviors. It just frustrates, all of this

58:13

frustrates me because it isn't being done

58:15

for anyone other than a vague

58:18

vibe of political goodness or

58:20

social, not even social, just

58:22

economic goodness. Yeah.

58:24

The Wall Street Journal reported earlier

58:27

this week that Kotick, the former

58:29

CEO of Activision, is

58:32

looking for partners at a dinner at an

58:34

Allen & Co. conference earlier this week. So,

58:36

it's for Journal rates. We're all looking for things

58:38

happening. Kotick floated the idea of partnering to

58:40

buy TikTok to a table of people that

58:43

included OpenAI CEO Sam

58:45

Altman. Right. OpenAI could

58:47

use TikTok to help train its AI

58:49

models if a partner such as Kotick

58:51

could raise the capital for such an

58:54

acquisition. It

58:56

all seems to converge. Right. They're

58:58

not going to sell. Of

59:01

course. They're not going to sell.

59:03

And I mean, if this someone's passing the

59:05

Senate and ends up on Biden's desk where

59:07

he signs that we're going to have a

59:09

protracted legal battle before

59:12

it ever gets to a place where someone could buy

59:14

it. Yeah, about the first effing amendment. Yeah.

59:17

This is a matter of freedom of expression

59:19

in this country. I

59:22

guess so. What's the matter in that? Is

59:24

that, is it a matter of freedom of expression in

59:27

the company? Oh, I absolutely think so. Not

59:29

a bite dance, but of all the people,

59:31

all the citizens on TikTok

59:34

whose speech is cut off. There

59:37

is a class action, I think, there. It

59:39

said that you're taking away our platform. Our

59:42

means to choose to a private enterprise.

59:46

Well, no different.

59:49

I mean, I don't believe the

59:51

Internet is a medium, but now

59:53

putting that aside, it'd be no

59:55

different from saying that Simon &

59:58

Schuster is a... Outlawed

1:00:00

To the authors that point by the

1:00:02

government. The authors that point of action.

1:00:05

Their. Platform for speaking is gone.

1:00:07

Take. Away Wordpress. Take.

1:00:10

Away of. Ah,

1:00:15

And open source platform. Is

1:00:18

A for it's a matter of expression and

1:00:20

where we'd that we don't read A We

1:00:22

respect books, We don't respect Tv. we're We're

1:00:24

so excited rates play. We're not directly Titus

1:00:26

I, I I am not a lawyer. On.

1:00:29

Show any of us are know. They.

1:00:31

Have to coal mine, regulate. Vote for things I

1:00:34

said. Screw up this. Isn't

1:00:36

your thought x She's very sorry

1:00:39

he didn't receive further is protected

1:00:41

under free speech by the way.

1:00:43

Margin: Call you back from his

1:00:45

opinion See: Email me Bobby, We'll

1:00:47

get we'll get you on the

1:00:49

positive nasty freight anyway and. I.

1:00:53

Saw a whole thing with a postman

1:00:55

those that was restricting individual speeds by

1:00:57

banning a platform with that counts. I'm

1:00:59

honestly would be fascinated either Was a

1:01:02

class action. Iran's a home. This.

1:01:05

Event or Concept. Is

1:01:07

also interesting at my colleagues. At

1:01:10

the information reported yesterday that

1:01:12

ah, unlike. All. The

1:01:14

previous times that Tix Heart has

1:01:16

been threatened with her advance. The

1:01:18

investors are the doesn't come with

1:01:21

according to the rescue, the science

1:01:23

and select. Few of the by

1:01:25

dance investors. That a previously lobbied for

1:01:27

tic toc are doing it now and probably deadly

1:01:29

said that they don't see any reason to

1:01:31

speak out in such a polarizing issue. Or

1:01:34

he. Did. They could be foot perceived as

1:01:36

it's disloyal to the U S. if they.

1:01:39

Come out in support of textile orders

1:01:41

a sixth of senate as.to dissolve Performative.

1:01:43

The house wanted to get there ya

1:01:45

ya as the senate's not gonna pass

1:01:47

it and so we don't worry about.

1:01:50

the it seems like they it's

1:01:52

a couple there's a couple different

1:01:54

parts are one is they're worried

1:01:56

about wasting like venture capitalists seem

1:01:58

to worried about waste their political

1:02:00

clout and want to pick their

1:02:02

battles because they say that they

1:02:05

have upcoming battles

1:02:08

around AI and defense tech that they

1:02:10

think are really important and want to

1:02:13

save their political capital to engage

1:02:15

with. But they've also said,

1:02:17

and this is quoting from the

1:02:19

article, that after

1:02:22

four years of an on and off discussion about

1:02:24

the ban, there's a growing chance that tech talk will

1:02:26

be forced to withdraw from the US. They noted

1:02:28

that a US ban wouldn't end TikTok. The

1:02:31

app could still operate outside the country,

1:02:33

although its ability to make money would

1:02:35

be sorely diminished given the US is

1:02:37

the world's biggest ad market. And some

1:02:39

VCs even play down the financial impact

1:02:41

that TikTok has on ByteDance,

1:02:43

arguing that the Chinese tech company

1:02:46

might in some ways be healthier without it. I

1:02:50

think it's very interesting that this seems

1:02:52

to be a moment where some of

1:02:55

the company's biggest supporters are also bowing

1:02:57

to pressure. I

1:02:59

was a bit surprised by this sudden pivot too.

1:03:03

The internet is very unpopular right now. I

1:03:06

mean it is, but also I think

1:03:08

that the venture capital firms might be

1:03:10

a little bit scared of the government

1:03:14

because if perhaps they went and looked

1:03:16

at how much money they put into

1:03:18

things like crypto or metaverse

1:03:20

or how much money they put into

1:03:22

just completely non-existent, I think like

1:03:24

a weight loss startup got $450 million, $500 million in 2021. There's

1:03:29

a lot of things that VCs did that

1:03:31

definitely flaunted very basic due

1:03:33

diligence. Look at FTX for example. And

1:03:37

I think that they're scared of the exposure. I don't blame

1:03:39

them. I wouldn't want it. And

1:03:42

also... Yeah, especially after the huge boom time

1:03:44

we're just coming off where anything kind of

1:03:46

goes, it is now crunch time to

1:03:48

say the least. That's not a sympathetic

1:03:51

character. It's a multi-billion

1:03:53

dollar monstrosity

1:03:56

algorithm machine. And Mark Zuckerberg hates

1:03:58

it because he was never... able

1:04:00

to create an algorithm this good. But

1:04:02

everybody here likes it. Of

1:04:05

course, I'm not even, I'm not saying it's bad,

1:04:07

I'm just saying the perspective that people are

1:04:09

coming from. I find it upsetting.

1:04:11

It makes me feel 100 years old. I load

1:04:13

it up. I was just saying, do you use

1:04:15

TikTok? I assume you don't make

1:04:17

TikToks, but do you watch them? I

1:04:20

can't understand it. I

1:04:22

use it and I watch. No,

1:04:24

it really is. I'm getting old. I've never been

1:04:26

good at video either, but also

1:04:28

there's something that gives me anxiety that the

1:04:31

feed never ends, that there's always more. I

1:04:33

mean, that's every website now though. No, but

1:04:35

it's worth saying, isn't it? But

1:04:38

it's also, I'm being fed stuff that I didn't

1:04:40

ask for, which also really upsets me. I

1:04:43

don't like the algorithm interfering as much as

1:04:45

it does, or at least I like it

1:04:47

to be a little less overt with it.

1:04:49

Instagram is bordering on unusable for the

1:04:51

same reason. TikTok is just hyper aggressive with

1:04:53

it. I kind of respect the fact that they're so honest with

1:04:56

it. It's just like, yeah, we control this. You're going to see

1:04:58

whatever we want. We want you, little pig. Jess,

1:05:00

you use, you watch TikTok through

1:05:02

the app, right? Yeah. I'm

1:05:05

curious if you've had the same experience. I

1:05:07

feel like the main point of TikTok is

1:05:09

they hyper curate your feed to be everything

1:05:11

you'd want and more. But

1:05:13

I've noticed the last maybe dozen times

1:05:16

I've opened TikTok for the past couple

1:05:18

of weeks, my feed is bananas. It

1:05:20

is completely random. It is like all

1:05:22

preferences they've previously had for me have gone

1:05:25

out the window. I'm not seeing any more

1:05:27

contact about the guy with the eel pit

1:05:29

or the woman who is legally digging

1:05:31

tunnels, which are my two favorite TikTok

1:05:33

subjects. Suddenly, I'm just

1:05:35

seeing random stuff out there. There are still

1:05:37

cat videos usually, which is what's keeping me

1:05:40

going back, or strange musical content. Have you

1:05:42

noticed any change to your home? My

1:05:45

theory, yes, I have, but it's

1:05:47

sporadic. I find that like every

1:05:49

third day, boy, it's feeling

1:05:52

weird today. And the next day it's stuff I

1:05:54

like. I think they try stuff

1:05:56

out on you to see what

1:05:58

happens. they what I've

1:06:01

never managed to get it dialed in it's

1:06:04

never engaged me I've never been like I must

1:06:06

watch this but also I just don't consume that

1:06:08

much video like I'm not even saying it's a bad

1:06:10

app I don't like video that much I don't

1:06:13

I am similar I don't watch YouTube

1:06:15

videos ever is my famous

1:06:17

statement in the pie I watch no YouTube

1:06:20

videos or like 12-minute long ones hmm

1:06:24

and but it's just this whole bill I think we

1:06:26

can wrap it up by saying who's

1:06:28

it help who's this actually protecting who

1:06:31

does this benefit bill yeah yeah

1:06:34

I mean I think that it held the centralization

1:06:36

of power in the hands of few guys white

1:06:38

guys within the tech industry yeah

1:06:41

white homogeneity pushed

1:06:43

at scale it's going to make rich

1:06:45

men even richer it's going to centralize

1:06:47

the same tools by the way the

1:06:49

same things are basically done

1:06:52

using American social networks look at what Jason

1:06:54

Kobler is tweeting right now and

1:06:56

it sucks it sucks because if there

1:06:59

was a real problem I'd love to see it and

1:07:03

we should be doing a supporting open source

1:07:05

we should be supporting math done and

1:07:07

and pushing blue sky and even threads to

1:07:09

open up blue sky just opened

1:07:12

up its its moderation structure good name

1:07:14

ozone I like that I love blue

1:07:16

sky brilliant I love you too I

1:07:18

like it a lot I haven't been

1:07:20

posting there as much I've been tweeting

1:07:23

a lot lately getting

1:07:25

back into it but I've got to start blue

1:07:28

sky blue sky threads the

1:07:30

problem is people I like around threads but

1:07:33

I don't like friends as much it is

1:07:35

a putrid place it is like it's

1:07:37

like someone created linked in off

1:07:39

the dark which is not that

1:07:41

fun that sounds like someplace Jason

1:07:43

Kelz telekines would be no he

1:07:45

goes on Twitter and

1:07:48

he goes you know what racism is bad

1:07:50

but if we thought about why and it

1:07:52

would just be like a horrible thread where

1:07:54

he's giving freedom someone very racist he did

1:07:56

a whole thing about Alex Jones where he

1:07:58

was like why don't a good But

1:08:00

why don't I agree? And

1:08:02

he says this, I'm just asking the

1:08:04

questions thing, this vile right-wing thing, like

1:08:06

a classic right-wing trope. And

1:08:09

he's like, I wasn't saying I even

1:08:11

want Alex Jones back, but now he's

1:08:13

coming back. Isn't that

1:08:15

interesting? It's just, it's

1:08:18

very frustrating, all of that stuff. Completely different

1:08:20

subject, I realize. Switching gears

1:08:22

a little bit, there was a really interesting story from

1:08:24

Cashmere Hill up in New York Times this week. Yes,

1:08:26

I'm glad you're doing this. Sort of his thought on

1:08:29

how the auto industry is

1:08:31

collecting reams of data about consumers'

1:08:34

driving habits and sharing it with insurance

1:08:36

companies. The opening

1:08:38

anecdote in the story follows a guy

1:08:41

named Ken Dahl, who drives a leased

1:08:43

Chevrolet Bolt and says, he's a careful

1:08:45

driver. He's never been

1:08:47

responsible for an accident. And Dahl

1:08:50

was surprised in 2022 when

1:08:52

the cost of his car insurance skyrocketed by

1:08:54

21%. Quotes

1:08:56

from other insurance companies were also high,

1:08:58

and one insurance agent told him

1:09:01

that his LexisNexis report was a

1:09:03

factor. LexisNexis is this

1:09:05

New York based kind of global data broker.

1:09:08

They catered to the auto insurance industry

1:09:10

with this one kind of division they

1:09:12

have, and has traditionally kept tabs on

1:09:15

car accidents and tickets and like helps

1:09:17

determine potential insurance rates from that. Dahl

1:09:20

wanted to know what was causing his

1:09:22

insurance increase, so he requested info from

1:09:24

LexisNexis under the Fair Credit Reporting Act. LexisNexis

1:09:27

sent him a 28 page, no 258

1:09:29

page, sorry, 258 page consumer disclosure report, and

1:09:36

what it contained stunned him from

1:09:38

the Times. More than 130

1:09:40

pages detailing each time he or his

1:09:42

wife had driven their Bolt over the

1:09:44

previous six months. It included the dates

1:09:47

of 640 trips,

1:09:49

their start and end times, the

1:09:51

distance driven, and an accounting

1:09:54

of any speeding, hard braking, or sharp accelerations.

1:09:56

The only thing it didn't have was where

1:09:58

they had driven their car. And

1:10:01

so according to this report, the trip details

1:10:03

had been provided to LexisNexis by

1:10:05

GM, the manufacturer of the

1:10:07

Chevy Bolt. And a

1:10:09

bunch of insurance companies had requested

1:10:11

information about Dahl from LexisNexis over

1:10:13

the previous month. A quote

1:10:16

from Dahl said, it felt like a

1:10:18

betrayal. They're taking information that I didn't realize

1:10:21

was going to be shared and screwing with

1:10:23

our insurance. And

1:10:25

I think kind of the money quote in this story is, automakers

1:10:30

and data brokers that have partnered to

1:10:32

collect detailed driving data from millions of

1:10:35

Americans say that they have driver's permission

1:10:37

to do so. But

1:10:39

the existence of these partnerships is nearly

1:10:41

invisible to drivers. This consent is obtained

1:10:43

in fine print and murky privacy policies

1:10:46

that few read. As

1:10:48

someone who doesn't drive, I feel it's inappropriate

1:10:50

for me to react to this story. So

1:10:52

I'm turning it over to the two, I

1:10:54

assume, drivers on this podcast

1:10:56

today. Ed? I

1:10:59

mean, welcome to the machine. I mean, this

1:11:02

has been going on since Metro Mile came

1:11:04

in since they, progressive over

1:11:06

a decade ago, started doing an OMD thing.

1:11:09

You click into your car that would lower

1:11:11

your insurance. They've been collecting this data forever.

1:11:13

How do you think the insurance rates have

1:11:17

changed over the years? They've changed because they've been collecting

1:11:19

this data. It's I

1:11:22

assume that Tesla with their insurance does

1:11:24

the same thing. They're fairly

1:11:26

sure they do something like this. And

1:11:29

guess what? This is the

1:11:32

future. This is why having a car

1:11:34

full of stuff, full of tech, full

1:11:36

of apps is actually a really bad

1:11:38

thing. I don't like that much information

1:11:41

out there. Yep, there it is. I

1:11:44

hope. See this, so this, by the way,

1:11:46

this feels like something that the government and

1:11:48

the house and Congress need to get their

1:11:50

up in arms about. Basic

1:11:52

transparency in data collection and its

1:11:55

use. That is a this

1:11:57

story should be bigger than TikTok because

1:11:59

this is. affecting more Americans right now

1:12:01

today. The cost of car insurance is

1:12:03

an incredible burden on the average man. I

1:12:06

have a beautiful Volvo XC4 electric car.

1:12:08

It's got mechanical doors, technical

1:12:11

doors, they open the closed, it goes forwards, backwards, side

1:12:13

to side, it's an amazing vehicle. Does it

1:12:15

open the screens that does all you can like

1:12:17

have an app? It also has a beautiful physical

1:12:21

knob. Fun little side, I met

1:12:23

Steve Wozniak the other day and he was talking

1:12:25

to me about how electric car sucks. They all

1:12:27

have screens, right? And now we hate inclusive motors

1:12:29

because they've got big screens and everything's got a

1:12:31

screen. Mercedes got a screen. And

1:12:34

I think that that's part of the problem. But the

1:12:36

other problem is how much of that data, because a

1:12:38

lot of them have like sensors about driver awareness, for

1:12:40

example. Driver how engaged you are with

1:12:42

the road. These are the real, these are

1:12:45

the things that are going to really screw people. Don't

1:12:47

be scared of China coming into the TikTok app. Be

1:12:49

scared of the American

1:12:51

manufacturers that are currently tracking

1:12:53

how much you are seeing

1:12:55

the road. How engaged

1:12:58

you feel. These are the scary

1:13:00

things. Especially because these tools aren't

1:13:02

always accurate. I know with the

1:13:04

driver awareness tool, when I

1:13:06

reported on Amazon, they had kind of installed

1:13:08

these in the cabs of some of their

1:13:10

trucks and, you know,

1:13:13

driver vans. And there was kind

1:13:15

of massive wave of complaints from

1:13:17

drivers, not because they were like,

1:13:19

oh, we want to not pay

1:13:21

attention to the road, because they're like,

1:13:23

I'll be sitting there driving, paying attention

1:13:25

to the road. And it will ding

1:13:27

me five times for inattention, perhaps because,

1:13:30

you know, the person driving is

1:13:33

Asian or just not white.

1:13:35

And the camera can't properly

1:13:37

attune itself to their eyes.

1:13:39

This has been a slippery

1:13:41

slope. Well, this is the smart

1:13:43

car. It's

1:13:46

probably a middle school car. It's what,

1:13:48

10 years old. So probably

1:13:50

doesn't have all of this. No, I

1:13:53

had to spend six hundred dollars to get Android auto. What

1:13:57

is that? Android auto, same as your Apple.

1:14:00

home. I can look at the map on

1:14:02

the cars screen. It has a screen. First,

1:14:07

let's underscore, as we like to say in

1:14:09

journalism, great reporting

1:14:11

by Cashman Hill. This is

1:14:13

the kind of privacy story. Instead of doing the moral

1:14:15

panic about the internet,

1:14:19

it's taking all of our data and I don't

1:14:22

know my data and it's non-specific enough to the

1:14:24

point. This is specific. It

1:14:26

is a violation of your

1:14:28

own space and what

1:14:31

you do. Point one. Point

1:14:33

two, none of this required the internet. This

1:14:36

is about evil data

1:14:38

brokers and evil insurance companies.

1:14:41

And the data is going to be out

1:14:43

there. What's needed for legislators, as

1:14:45

you say Ed, what's absolutely needed now is

1:14:47

for them to come up with and what they can do

1:14:49

is they can forbid the use of this data. You

1:14:52

can't, it's going to be collected for

1:14:54

various reasons, for various things. Okay, whatever.

1:14:56

You can forbid it being collected. You

1:14:59

can forbid it being used for insurance.

1:15:01

That's what legislators are for. Or

1:15:03

even at the bare minimum mandate

1:15:05

that they disclose, it's being collected and

1:15:07

give you the opportunity to opt out. But

1:15:10

also to your point earlier,

1:15:13

Harris, like remember

1:15:15

Connect, Microsoft's camera, like

1:15:18

the thing for the Xbox, it couldn't see black

1:15:20

people. The

1:15:22

panopticon systems that many contractors

1:15:26

use, the ones that they use to

1:15:28

tell if you're actually watching the screen,

1:15:30

they continually see things like dreadlocks

1:15:33

or even just the hair of African-American

1:15:35

people as things that suggest they're not

1:15:37

there. These systems

1:15:40

are biased. They are racially biased. And

1:15:42

that's before you get to the obvious,

1:15:44

like this is just invasive. These systems

1:15:47

are biased against people like many algorithmic

1:15:49

systems, like the mortgage system that was

1:15:52

something like 80%. There

1:15:54

was an insane amount of bias in the mortgage system. Using

1:15:58

datasets to make calls like this is...

1:16:00

insane and by the way I'm correct

1:16:02

Tesla does have a safe driving system

1:16:04

for their insurance they give you a

1:16:06

score kind of similar to Metro mile

1:16:08

these companies want to do this and they're

1:16:10

doing it because guess what they can just

1:16:12

they can now hyper justify how

1:16:15

they judge you but guess what if these

1:16:17

are the systems then to be clear there

1:16:19

is a separate thing here there is the

1:16:21

I'm tracking where you're going and what you're

1:16:23

doing that is separate to

1:16:25

racial discrimination and the engagement

1:16:28

systems those I don't know if

1:16:30

they're using that data but just

1:16:32

a very basic driving here that

1:16:35

is going to have a class issue as well

1:16:37

you're driving through rougher neighborhoods which

1:16:39

as judged by an insurance company

1:16:41

if you're driving in such a way

1:16:44

that suggests your profile is now

1:16:46

risky which is racially biased these

1:16:48

systems are going to hurt so many people

1:16:50

and they already are and

1:16:53

yet here we are chasing tick I

1:16:55

just right right right right it's just

1:16:57

like this is a thing hurting people

1:16:59

today cashmere does a great job at

1:17:01

this cashmere's done great work

1:17:03

on data brokers companies like spoke yo and

1:17:05

tell us these companies

1:17:09

are actually actually start axios axiom but kill

1:17:11

me the next business one all they have

1:17:13

who knows

1:17:17

but we got these companies these companies

1:17:19

are actual scum they're actually dangerous to

1:17:22

Americans they are brokers you can pay

1:17:24

someone 25 bucks a month to

1:17:26

look up everything they can find out your

1:17:28

phone number your address in many cases sounds

1:17:30

like a moral panic it actually really isn't

1:17:32

companies like a bean who made good money

1:17:35

deleting that information from the internet but

1:17:37

there's only so much you can do these

1:17:39

are things to target somebody shouldn't

1:17:41

they are brokers should be illegal but

1:17:44

here's the other issue well

1:17:47

two things one day workers go way back when

1:17:50

I worked at time we used to workers like

1:17:52

crazy and that's how we did subscription models and

1:17:54

all this kind of stuff I used to scare

1:17:56

students and say let me go

1:17:58

to axiom not axios And I'd

1:18:01

say, let me find out the names and

1:18:03

addresses of, just to be really creepy, women

1:18:06

between the ages of 21 and

1:18:08

35 who live within n

1:18:10

miles of this place who have a

1:18:12

college degree and a car. And

1:18:15

it'll give it to me. Far more than the

1:18:17

internet ever does. Far more. But

1:18:20

here's the other issue. I just said we need laws, we need

1:18:22

legislators. But of course, government is the worst

1:18:24

enemy of privacy. Because what's going to happen is,

1:18:26

fine, this data doesn't tell the insurance company where

1:18:29

you went. But it's discoverable

1:18:32

and certain states are going to come in and find

1:18:34

out whether or not you went to an abortion clinic.

1:18:37

And the data's there. And

1:18:39

that's the real issue. Exactly.

1:18:42

And so, there

1:18:44

needs to be transparency. There needs to be the

1:18:46

right to erase it. Again, it's not the internet.

1:18:48

It's your car and the insurance company and your

1:18:51

state's government that are the real problems here. Yeah,

1:18:53

the internet can do stuff too. But it's more

1:18:55

anonymized there. It's not in your personal space. This

1:18:59

is about,

1:19:02

I wrote a book about privacy called Public Parts. And

1:19:05

so, I learned about all this

1:19:07

stuff. Wow, a rare drink opportunity.

1:19:09

We never get private parts. Yeah,

1:19:11

public parts. Public parts is Howard

1:19:13

Stern. I

1:19:18

dedicated it to Howard Stern as a result because

1:19:20

he was part of the title. There

1:19:25

is no right to privacy in the Constitution.

1:19:28

It was invented by case law

1:19:30

over time. It's something that,

1:19:33

and I believe in the value of publicness and being

1:19:35

out there and talking, I believe in the value of

1:19:37

privacy. We've got to protect it. But

1:19:39

it's a... And

1:19:43

the problem I have is that when

1:19:45

you have crap like Shoshana Zuboff and

1:19:49

surveillance capitalism, it

1:19:51

ruins it because it just goes overboard with, I'm

1:19:54

going to say it again, moral panic in a

1:19:56

way that's not really about harm. It's not really

1:19:58

about the real problems. Attraction,

1:20:01

but it's this kind of stuff. Don't pay attention to

1:20:03

zoom off pay attention to cashmere Hill Yeah,

1:20:07

Carl bowed Wrote

1:20:09

a really good breakdown of cashmere's

1:20:12

article in tech dirt Which

1:20:14

all I'll quote from a little bit here Again

1:20:18

like countless past scandals This is the

1:20:20

direct byproduct of a country that has

1:20:22

proven to corrupt to pass even a

1:20:24

baseline privacy law for the internet era

1:20:26

To corrupt to regulate data brokers and

1:20:29

obsessed with steadily defanging the funding understaffing

1:20:32

and curtailing the authority of

1:20:34

regulators tasked with overseeing corporations

1:20:36

the broad and reckless disdain for US

1:20:38

consumer privacy and safety Yeah,

1:20:42

I do think he has a point whole

1:20:44

rocks top college

1:20:47

Call it much nuts that fella.

1:20:49

He's absolutely ruthless. He's one of the

1:20:51

few people was on Elon Musk early

1:20:54

He really he was pushing very hard Of

1:20:58

course not just want to be clear

1:21:00

that Laura Kolodny very early on the

1:21:02

Elon Bay. Absolutely ruthless big up Laura

1:21:04

yep, but he's done an excellent job

1:21:06

of Really

1:21:08

getting to the heart of how angry you should be I

1:21:12

used to and I feel like this

1:21:14

thing is a lot of What

1:21:16

the tech industry does that leads to things like

1:21:18

this is they turn everything into a very complex

1:21:20

shell They make it sound like magic goes back

1:21:22

to the AI think goes about to tick tock

1:21:24

all this Not many sensors could really

1:21:27

explain to you why tick tock's back. They're just

1:21:29

like ah China Right, right.

1:21:31

I know China's scary. They're in my phone This

1:21:34

thing with the cars sounds

1:21:36

complex was signed in the complex way, but

1:21:39

is fairly simple They are invading your privacy

1:21:41

and I feel like there Should

1:21:44

be someone in Congress House there

1:21:46

I say even the president who

1:21:48

could just say these things like

1:21:50

Carlos Cole call is actually very

1:21:53

cleanly spoken and furious about the stuff as he

1:21:55

should be And I feel like more

1:21:57

people in tech should be I'm

1:22:00

angry every time I read about the stuff I get

1:22:02

angry because you see the world burning around you and

1:22:05

you see The fire

1:22:07

engines all rolling up to a house that isn't

1:22:09

on fire And

1:22:11

it's just yeah, but who wants to go after

1:22:13

you're going after in this case the

1:22:17

auto industry the insurance industry and

1:22:20

the data broker industry which all politicians

1:22:22

use for all of their direct mail

1:22:24

and course advertising

1:22:29

and They say well, they're Alex. They can. Yes.

1:22:31

Well, they're gonna be Well,

1:22:33

that sucks on a on another positive

1:22:36

note. I think that we should use this opportunity

1:22:38

to go to an ad break Jeff

1:22:41

Jarvis at Zitron everybody and

1:22:44

now you're gonna go to an ad

1:22:46

break You

1:22:51

need parts O'Reilly auto parts has

1:22:53

parts need them fast We've

1:22:56

got fast no matter what you

1:22:58

need We have thousands of professional

1:23:00

parts people doing their part to

1:23:02

make sure you have it product

1:23:04

availability Just one part that makes

1:23:07

O'Reilly stand apart the professional parts

1:23:09

people Um,

1:23:18

I want to talk next about a A Great

1:23:21

article that the New York Times

1:23:23

did on taking a deep look

1:23:25

into Musk's Charitable

1:23:29

foundation and I want to also

1:23:31

throw it to Ed for kind of an overview

1:23:33

of this a little bit because I know you

1:23:35

Just did a podcast episode about this where you

1:23:37

interviewed One of

1:23:40

the authors of it David Serenthold

1:23:42

I guess can you tell me a little bit about the article and

1:23:44

like what I got was like, yeah

1:23:46

absolute absolute legend David he jumped on

1:23:48

the phone the day after the article

1:23:51

went out What a legend so

1:23:53

long story short Elon Musk part about

1:23:55

five six billion dollars of Tesla shares

1:23:57

to be clear They didn't buy them

1:24:00

just chairs he had in the Musk

1:24:02

Foundation, a non-profit. Now part of the rules

1:24:04

of doing this is you have to spend

1:24:06

5% of it every

1:24:08

year to qualify for the $2

1:24:10

billion tax break

1:24:13

he got. He regularly fails to do

1:24:15

so. He regularly fails to invest enough

1:24:17

money to do so. But he

1:24:21

also, unlike, say, Larry Page's foundation, he

1:24:23

doesn't dump it in a donor advised

1:24:25

fund, which is kind of a black

1:24:27

box for giving grants to people. No,

1:24:29

he has no staff, he

1:24:32

has two unpaid volunteers, and he

1:24:34

is the other person on the board, and

1:24:36

it isn't really obvious where

1:24:39

the money goes. When it does

1:24:41

go somewhere, it goes in either very

1:24:43

small amounts and does very little, or

1:24:45

very large amounts and also does very

1:24:47

little. There is a school called Ad

1:24:49

Astra, which is literally inside a SpaceX

1:24:51

compound where Musk's own children go. It

1:24:54

used to be outside, now it's inside

1:24:56

a SpaceX compound. They

1:24:58

got several million dollars. He gave $100

1:25:00

million from the Musk

1:25:03

Foundation to another non-profit just

1:25:05

called The Foundation. It's really

1:25:07

firing on all cylinders there.

1:25:09

The Foundation.

1:25:12

It really is good. And red. It

1:25:15

bought a bunch of land out

1:25:17

very close to a boring company

1:25:19

site in Texas. Very

1:25:21

cool, very good stuff there. He

1:25:23

also donated $55 million to a

1:25:25

cause of a guy who auctioned

1:25:28

off seats on a SpaceX flight,

1:25:30

who then immediately bought three more

1:25:32

seats on another SpaceX flight. And

1:25:35

I think one of the interesting things that David,

1:25:37

the reporter, brought up in your

1:25:39

interview is part of what

1:25:42

is supposed to be happening here is if

1:25:44

you are getting this tax

1:25:47

break from being a charitable organization, you have

1:25:49

to do charitable things. Meaning you have to

1:25:51

do things that benefit the public good. And

1:25:53

this article, the thing that it kind of

1:25:55

hits again and again, is that it's unclear

1:25:57

if any of you have any questions about

1:26:00

of kind of the Musk

1:26:02

Foundation's investments

1:26:04

or philanthropic endeavors,

1:26:08

it's unclear who they benefit outside of

1:26:10

Elon Musk or Elon Musk's

1:26:12

employees and customers, which is

1:26:14

incredibly unusual for a charitable

1:26:16

foundation. Well, a good one he did

1:26:19

as well as he said he was going to fix the water contamination

1:26:21

problems that are playing Flint, Michigan. Promised

1:26:24

he'd do so. He actually tweeted at one point that he'd

1:26:26

already done so and then deleted the tweet in 2018. He

1:26:30

gave him about one point two million dollars, which is

1:26:33

it's good he did something. They bought

1:26:35

water filters for the school. They bought laptops for

1:26:37

the school. Great idea. They then responded to him

1:26:39

with a four page plan basically saying, here's how

1:26:41

you do. Here's how

1:26:43

we will do the thing you promised to help us with.

1:26:46

And then you can fix the

1:26:48

water as you promised. He sent

1:26:51

a Tesla development executive to

1:26:53

Flint, Michigan, who gave people

1:26:55

rides around the around

1:26:58

the parking lot. What

1:27:01

of the city hall? He

1:27:03

arrived in the Tesla. That's incredibly

1:27:05

charitable. And then he said, hey, well,

1:27:09

we might build an office out here. You'll never

1:27:11

guess what happened. He didn't build anything. He

1:27:14

didn't fix. Flint, Michigan.

1:27:18

At all. Some

1:27:21

of the details in the story are just wild.

1:27:23

I mean, as you'd expect from an Elon Musk

1:27:25

story, some of them are kind of all

1:27:28

quote from here. Among the donations

1:27:30

the Musk Foundation has made, there was a

1:27:32

fifty five. There was fifty five million to

1:27:34

help a major SpaceX customer meet a charitable

1:27:36

pledge. There were millions that went to Cameron

1:27:38

County, Texas after a rocket blew up and

1:27:41

there were donations to two schools closely tied

1:27:43

to his business. One that

1:27:45

was literally physically walled off inside a

1:27:47

SpaceX compound. And

1:27:50

the other, like you mentioned, is located next

1:27:52

to a new subdivision for Musk employees. I

1:27:55

just want to go into a little bit more about the one you

1:27:57

mentioned, Ad Astra, which is Latin for to

1:27:59

the. Ours, ostensibly was

1:28:01

founded by Musk as a

1:28:03

nonprofit school to explore new ways to

1:28:06

teach math and science. And this is

1:28:08

from the Times again, but that school

1:28:10

too would serve a personal purpose for

1:28:12

Mr. Musk. In its first year

1:28:14

of operation out of his home in the

1:28:16

Bel Air neighborhood of LA, five of

1:28:19

Ed Astra's 14 students were

1:28:21

his own children. The

1:28:24

headmaster said the only criteria

1:28:26

for admission were quote, kindness and

1:28:29

eagerness to learn and parents that

1:28:31

worked at SpaceX. Company

1:28:35

store. I just

1:28:37

want to again underscore, I made fun

1:28:39

of the New York Times using the verb underscore

1:28:41

the other day for what reporters really want to

1:28:43

say to themselves. So anyway, I'll

1:28:45

underscore Cashmere Hill and David

1:28:47

Ferrenholt. The Times

1:28:49

is driving me completely batty lately with

1:28:52

his credulous coverage of fascism and

1:28:54

its Biden, his birthday polls

1:28:56

and all this stuff. But

1:28:59

the Times is also the repository for people like

1:29:01

Ferrenholt and Hill who do this kind of

1:29:03

great work. And we just more of this, please, less of

1:29:05

the other crap. Yeah, it's

1:29:07

how I feel. And David was great

1:29:09

as well. He was a real, he was a

1:29:11

president. But he was also very,

1:29:13

he was very, he was

1:29:15

very entertaining. Like he was very into the story

1:29:17

and you could tell it frustrated him. And

1:29:20

you could tell he'd like physically gone out to

1:29:22

where Ad Astra was meant to be. But

1:29:25

he wasn't able to get any of that was a, he

1:29:28

said this in the podcast, but there was just a sign

1:29:30

basically saying, do

1:29:32

not pass. I forget what the nonsense

1:29:35

they say, but it's like, no, no, no

1:29:37

unauthorized entry, blah, blah, blah, blah, blah, for

1:29:39

a school, just to be clear. And

1:29:42

it's just one of the

1:29:44

more depressing things he said, though, is the people

1:29:46

that would regulate this are like the IRS and

1:29:50

other parts of the government that are

1:29:52

underfunded. And it sucks. It sucks

1:29:54

because Musk will get away with this. Musk

1:29:57

will get away with this. And Ryan Mack.

1:30:00

who was the co-reporter on the piece as

1:30:02

well. He did a good amount of haven-founding

1:30:04

and he basically found that no one talked

1:30:06

to Musk about the foundation. Sorry,

1:30:09

the Musk Foundation I should say. No

1:30:11

one, he didn't bring it up. It was not

1:30:13

something that he talked about. It's not something he

1:30:15

thought about. It's just a big tax dodge. And

1:30:18

something else that Farinhold added at the end of

1:30:21

the podcast was, he was saying, and I mentioned

1:30:23

it earlier, Larry Page is the same deal. He

1:30:26

just dumps it into a donor-advised fund but he doesn't

1:30:29

invest as much as he needs to. And it's just

1:30:31

a way for them to get away with it. And

1:30:33

I think that Musk probably doesn't want

1:30:35

to spend that money because he doesn't

1:30:38

want to liquidate billions of dollars of

1:30:40

Tesla stock. What's also disgusting is everyone

1:30:42

who covered every single time he claimed

1:30:44

he would do anything with the foundation

1:30:46

while nothing happened. Every major

1:30:49

publication, Wall Street Journal headline was, Musk

1:30:51

gives 11.6 million Tesla shares

1:30:54

to charity, failing to

1:30:56

put in that headline the

1:30:59

words, his own. Failing to

1:31:01

follow up. This stuff has been out

1:31:03

there a while. A

1:31:05

lot of this is publicly available stuff.

1:31:08

We all cut lines. Not my job. I'm running a PR firm.

1:31:10

But, Jim, this should

1:31:12

have been on this already. But I think there is, if

1:31:15

not a fear of Elon Musk, there

1:31:18

is still this want to believe that

1:31:20

he means well. And Kara Switzer has

1:31:22

been particularly scummy on this one. I think it's

1:31:24

the same sort of dynamic that we saw with

1:31:26

Trump, where it's like the

1:31:28

traditional media apparatus is

1:31:31

not equipped to

1:31:33

handle someone

1:31:35

who manipulates the truth at

1:31:38

such a, you know, regularity.

1:31:41

They are equipped. All you need to do

1:31:43

is refer to him as a con artist.

1:31:45

But I realize all you need to do

1:31:47

is put his statements in context with reality.

1:31:50

But I remember seeing this right

1:31:52

after I think, you know, when his star

1:31:54

was rising with the acquisition of Twitter, there

1:31:57

was some completely after.

1:32:00

and I'm tweet or something that he'd put out that

1:32:03

Bloomberg or someone wrote a whole story about it

1:32:05

as if it was true and of course it

1:32:07

just ended up being something he tweeted and

1:32:09

never ended up revisiting

1:32:12

every must story every must story in

1:32:14

every must he was gonna six billion

1:32:16

dollars world hunger we neural

1:32:18

link customer we put it in the first

1:32:20

thing tweet I'm gonna

1:32:22

buy Wikipedia tweet every otherance it becomes a

1:32:24

story sure except

1:32:30

it's dumber with Elon Musk he

1:32:32

is a like musk musk and Trump

1:32:34

are both proven lies they're both dangerous

1:32:37

dangerous to society in both different ways

1:32:40

with musk tech

1:32:42

industry knows better but I think the tech industry

1:32:44

look if you want to just be an enthusiast

1:32:46

industry that's fine be happy with everyone

1:32:49

no critiques or anyone you can't critique some but

1:32:51

not all musk

1:32:53

has posted multiple great

1:32:55

replacement theories things he's

1:32:58

continually doing this right-wing

1:33:00

firebrand nonsense this anti-immigrant

1:33:02

stuff he is out

1:33:04

and out racist he posted something very

1:33:06

racist against african-americans a couple months ago to

1:33:09

do with sex which I'm just not gonna go

1:33:11

any further on and

1:33:13

nobody did anything the great

1:33:16

replacement thing was a few weeks ago nobody

1:33:18

covered nobody but they will

1:33:21

cover when he fart since this I'm

1:33:23

going to I'm gonna remove likes I've

1:33:25

been a very much yes

1:33:28

and we're going to make the tweets

1:33:30

go sideways now and it's and

1:33:33

the thing is you know who actually has done a really good

1:33:36

job on this I'm gonna give a lot of credit to TechCrunch

1:33:38

TechCrunch who really did start off as kind of

1:33:40

very rah rah for tech has been damning on

1:33:42

musk those people like Amanda Silverling

1:33:45

does an excellent job sadly I've

1:33:47

Darin Essington's head moved on from there now

1:33:49

but they're doing a

1:33:52

great job calling out but CNBC isn't

1:33:54

CNBC other than Laura Kolodny who is absolutely

1:33:56

amazing and does a hell of a

1:33:58

job there you can't also You

1:34:00

can't, it undermines great reporting

1:34:02

like that when you sideline it

1:34:04

with just Musk's vague promises and

1:34:06

outright lies and then you don't

1:34:09

revisit them. And Paris, I

1:34:11

get your point. The comparison with Trump is

1:34:13

true. Except Trump, as a former president and

1:34:16

presidential candidate, there is a difference

1:34:18

there. Now, he's still

1:34:20

covered terribly, but there is a difference. Elon

1:34:22

Musk is a private citizen worth billions of

1:34:25

dollars. Rip into shreds! Take

1:34:27

his arse down! He

1:34:29

only went after media matters because he knew they

1:34:31

didn't have much money. He won't go after actual

1:34:33

journalists. Other than, here's the thing,

1:34:35

the real abdication of authority with Elon Musk is

1:34:38

in the hands of Kara Swisher. Kara Swisher should

1:34:40

be right across to you. She's doing a media

1:34:42

tour for Burn Book right now. Which

1:34:44

is an astonishing book. I mean, just

1:34:46

the fact that she's being interviewed by

1:34:48

Sam Altman and Reid Hoffman

1:34:51

about her book. It's an insult to

1:34:53

journalism. So you read it so we

1:34:55

don't have to. Say more. Yes. You

1:34:57

should read Paris Marx's review of it.

1:34:59

I was just Googling that right now.

1:35:01

Paris Marx, another great guy. But

1:35:04

with those, he makes the point that she kind

1:35:06

of acts like she criticizes people

1:35:08

but never really got down to it. She'd have a

1:35:10

moment to talk about his review. It's excellent. But

1:35:12

the long and short of Kara's book is, yeah,

1:35:15

all these people suck, but

1:35:17

I was at their parties. Right. With Elon Musk, she

1:35:19

only turned on him because he called her an asshole.

1:35:22

And to be clear, the thing he called her an

1:35:24

asshole for was because he misread a tweet where she

1:35:26

was supporting the US government paying him for Stalin in

1:35:28

Ukraine. Something he promised to do for free. People

1:35:32

like Kara have the ability to stop

1:35:34

men like this. They have

1:35:36

the ability to say, you are lying. They

1:35:38

have a huge... Stop. Same

1:35:41

deal. Oh, thank you.

1:35:43

Oh, my bad. Noire.

1:35:46

Why are you getting such fucking attention? I

1:35:49

want... I'll be honest, like... They

1:35:52

shouldn't get that. Better off when it's better

1:35:54

than they let it go out. Well, I mean, they get

1:35:56

attention because they are hyperbolic. Because they... No,

1:35:59

because they appeal to... very center liberal audience

1:36:01

of people who don't really have

1:36:03

morals but they do like posting

1:36:05

Instagram things. They appeal

1:36:07

to people who don't really want to believe in anything

1:36:09

other than that which feels convenient and

1:36:12

doesn't make them think too much. Kara is

1:36:14

the arbiter of that kind of information and

1:36:16

on top of that with Musk, she

1:36:19

was still defending his ass well into

1:36:21

2022. She was the one

1:36:23

saying he's hard to unpack. Oh he's

1:36:25

hard to unpack. He's so smart and we

1:36:27

know he's a flipping dullard. He's

1:36:30

a big wobbling candle of boringness

1:36:33

and he's not invented any of

1:36:35

the things he claims he invented. His

1:36:37

company's boring company, constantly goes to cities,

1:36:39

doesn't do anything, burns

1:36:41

people in Las Vegas. He's hurt

1:36:43

so many laborers in a very blue

1:36:45

collar city. Elon Musk

1:36:47

is a scumbag and calling him a scumbag and you

1:36:50

don't even have to do it in the newspaper. You

1:36:52

just say it in a talk. Maybe

1:36:54

you don't give him a talk. Aaron Ross-Sawkin shouldn't be

1:36:56

sitting there for the times and going, oh Elon, tell

1:36:58

me about your things. Tell me about your travails. No,

1:37:01

sit there and roast him and if he won't do

1:37:03

the interview, screw him. Say

1:37:05

he won't do the interview. It's time to

1:37:07

ask these people accountable. Anyway, sorry, stop. No,

1:37:09

this is fantastic. I could let you go

1:37:11

on. I was like, we got it. We

1:37:13

got to give a moment

1:37:15

to breathe. The great long,

1:37:18

I guess, review of

1:37:20

Kara's book, Burn Book by

1:37:22

Paris Marx is called Kara Swisher's

1:37:24

Reality Distortion Field. I posted a

1:37:27

link to it in the discord

1:37:30

chat. Here's a quote from

1:37:32

it. In the

1:37:34

end, Burn Book is the story of

1:37:36

one of Silicon Valley's most prominent access

1:37:38

journalists who took a page from the

1:37:40

billionaire she covered and created her own

1:37:42

narrative to see and present herself as

1:37:44

something else entirely. The story Swisher is

1:37:46

trying to tell helps to distance herself

1:37:48

from the decades she spent boosting companies

1:37:51

that have now been quote, disastrous, as

1:37:53

she herself admits. But it also works

1:37:55

for the industry. Letting Swisher present

1:37:57

herself as a tough reporter allows Silicon Valley to

1:37:59

be a Valley to pretend it was

1:38:01

being held to account this whole time

1:38:03

when really Swisher was along for the

1:38:06

ride and bought into the tech determinist

1:38:08

worldview guiding the industry. What

1:38:11

do you think of her, Jeff? Oh, I

1:38:14

say probably she blocked me long ago. And

1:38:16

I wasn't even saying anything that was that

1:38:18

awful. She's very, she acts all tough and

1:38:20

she's very oversensitive to criticism

1:38:23

herself. I

1:38:28

think she could have been a good analyst 20 years

1:38:30

ago. I think she could have figured things

1:38:32

out. But

1:38:36

I'm not sure where it went south. Was it putting people

1:38:38

on the red chair and making a lot of money from

1:38:40

that? I think

1:38:42

that there is this like version,

1:38:45

I mean, it's something that many people

1:38:47

are, can

1:38:50

happen to many people. Is when you move

1:38:52

from being a down

1:38:54

in the trenches reporter to

1:38:57

being a quote unquote commentator with

1:38:59

like a capital C and your

1:39:01

job is to have tech luminaries

1:39:04

or luminaries in the industry sit

1:39:06

across from you at a big

1:39:08

conference and ask them questions, part

1:39:10

of what is going on there

1:39:12

is you're existing in an economy

1:39:14

where your continued success relies on

1:39:17

the fact that you've got to

1:39:19

get those people to pick up your

1:39:21

calls and come back for next year's conference.

1:39:24

And I think that that causes a little

1:39:26

bit of brain rot in the sense of it's hard to be

1:39:30

truly pushing the envelope in

1:39:33

those industries. I think

1:39:35

in comparison to I've watched Galway, it was

1:39:37

a different case. I

1:39:39

used to go to the DLD conference in Munich and

1:39:41

he always did this spiel, you know, 100 slides

1:39:45

and 20 minutes and I'm going to be really funny the

1:39:47

whole time. And then

1:39:49

I was on MSNBC once with him

1:39:51

on his first appearance on TV and

1:39:53

I saw the drug shot right into

1:39:55

his vein. He

1:39:57

was a god. at

1:40:00

being on TV. It's no big deal. I was like,

1:40:02

you know, on daytime show, nobody sees it. My mother

1:40:04

didn't notice, you know, who cares? But he was, this

1:40:07

was it. This was the drug for him. He wanted

1:40:09

fame. And so it wasn't, in his

1:40:11

case, it's a little bit different. He gets invited to

1:40:13

give hugely expensive talks at places,

1:40:16

but he's not an, he's not an

1:40:18

access journalist so much as he,

1:40:20

as he's a faux-rye voice. And

1:40:23

so the two of them together share that

1:40:25

BS of acting like they're tough critics

1:40:27

of the world. He's a terrible predictor.

1:40:29

He's wrong about crap. She's not a

1:40:32

tough critic at all. It's

1:40:34

an act and people buy the act.

1:40:37

And that's the thing. I mean, I think part of it, obviously I

1:40:39

think they have, are responsible for their

1:40:41

own actions, bad takes and

1:40:44

bad predictions and bad interviewing. But I

1:40:46

also think that part of it is

1:40:49

what happens when you

1:40:51

are, have chosen to exist

1:40:54

within an industry that demands your

1:40:56

constant opinions and hot takes on

1:40:58

everything on a daily and weekly

1:41:01

and monthly basis. Like, I mean,

1:41:03

we've got this podcast, it's long,

1:41:05

but it's once a week that

1:41:07

I am expected to give commentary

1:41:10

and opinion on a very

1:41:12

specific niche amount of events that are changing.

1:41:15

She and other commentators that like

1:41:17

Echelon are doing this every single

1:41:19

day. That is how they're ready

1:41:21

to get to the point. They have to get on TV to do it. And you

1:41:23

have to get on TV and you've got to get clicks and

1:41:26

it is, it rots the brain.

1:41:28

It's the attention economy.

1:41:33

I think you're right, by the way. There's one other

1:41:35