0:00

This podcast is supported by Raisin. Raisin is

0:02

here to help you earn more on your

0:04

savings. With a single no fee account, you

0:06

can secure some of the best rates in

0:08

the nation on CDs and high yield savings

0:10

accounts from an exclusive network of federally insured

0:12

banks and credit unions. Your

0:14

streamlined dashboard allows you to fund and

0:16

manage products from multiple institutions, maximizing your

0:19

deposit insurance coverage. Best of all, the

0:21

minimum deposit is just $1, and

0:24

there are no fees. Visit raisin.com and

0:26

get started in minutes. You

0:28

see WWE's coming to Netflix. I

0:31

did. Are you excited? I

0:34

think it could be a good thing. I think it could

0:36

make wrestling into a thing that people didn't have to apologize

0:38

for watching. Now, this is a thing, because I used to

0:40

like wrestling back when I was a teen. And

0:43

then I would talk

0:45

about it with people and they'd be like, you watch

0:47

wrestling, and they would kind of roll their eyes or

0:49

their eyebrow would raise a little bit. They shamed you.

0:51

They shamed me. And that's

0:53

not the reason I stopped watching, but that

0:55

was one of the reasons. But do you

0:57

feel like a social stigma as a result of

0:59

your wrestling fandom? Yes, absolutely.

1:01

But the interesting thing about Netflix is

1:04

like Netflix has taken so many things

1:06

that seemed obscure or weird and just

1:08

made them popular. So whether it's like

1:11

Formula One racing with Drive to Survive

1:13

or golf with Full Swing or Suits

1:15

with Meghan Markle, it's like the

1:18

Netflix algorithm just takes stuff that nobody

1:20

used to watch and then makes it the side of the world. So

1:23

wrestling could have a huge 2025. You heard it

1:25

here first. When are

1:27

we going to do a tag team wrestling

1:29

act? I mean, as soon as the All In podcast

1:32

is ready to rumble. That's

1:34

two on four. You think we could take them? Absolutely,

1:37

yes. Well, you're very tall, so we've got a height

1:39

advantage. And you've got a lot of heart. I'm

1:46

Kevin Roos, a tech columnist for the New York Times.

1:48

I'm Casey Noonan from Platformer. And this

1:50

is Hard Fork. This week,

1:52

how AI is fueling a

1:54

newspocalypse. Then, Andresin Horowitz, investor

1:56

Chris Dixon makes the case that crypto still

1:58

has a bright future. Do we have

2:00

questions? And finally, it's time for

2:02

the year's first installment of Hatch RPG. All

2:23

right, Casey, we have to start this

2:26

week by talking about what is happening

2:28

in the media industry, because it has

2:30

been a very bad last couple of

2:32

weeks for a lot of these sort

2:34

of iconic publications. Just

2:36

in the last week or so, the LA

2:38

Times has conducted a big round of

2:40

layoffs. More than 20% of

2:43

the newsroom in total has been laid

2:45

off by that newspaper. Pitchfork,

2:47

the iconic music publication, is

2:49

being wrapped into GQ with

2:52

its top editor departing Sports

2:54

Illustrated, laid off most of its

2:57

unionized staff after some complicated maneuvering

2:59

by the corporate owners. Time

3:02

Magazine also conducted layoffs. And

3:05

Artifact, the sort of news aggregator app

3:07

that we talked about last year on

3:09

the show, started by the co-founders of

3:11

Instagram, it announced that it was

3:13

also shutting down, basically saying, we don't see

3:16

a future where there's a sustainable model for

3:18

a product like this in the news industry.

3:20

Yeah, I think it is really important.

3:23

The past couple of weeks seemed like a

3:25

time when more and more people are going

3:27

to just visit websites they used to visit

3:30

and realizing that they're not there or they've

3:32

lost a huge chunk of their staff. And

3:34

I think the web and, frankly, our culture

3:36

just feel less vibrant right now than they

3:38

did even just a year ago. Yeah,

3:41

and I think we should just start by

3:43

saying the problems that the news business has

3:45

today that we're seeing play out in all of

3:47

these layoffs and all this consolidation, there's

3:49

no sort of one single reason for

3:51

them. This is a long story that sort

3:54

of goes back 20 or 30 years some

3:58

parts of it. There's also a piece of it that

4:00

has frankly not that much to do

4:02

with technology. We've seen a lot of

4:04

kind of, you could call it financialization

4:06

of the media business. A lot of

4:09

struggling, especially local and regional newspapers

4:11

and TV stations have been bought

4:13

up and consolidated by hedge funds

4:16

and private equity firms that

4:18

have sort of strip mined them for

4:20

cash and left these sort of desiccated

4:23

husks out there in the world.

4:25

So this is not purely a story about

4:28

sort of Silicon Valley and tech power kind

4:30

of destroying what was left of

4:32

the sort of legacy media business. But

4:35

it does have a lot of crossover with

4:37

what's happening in tech. And so I thought

4:39

today we should just talk about the kind

4:41

of troubles that we're seeing in the news

4:43

industry, what's behind them and kind of where

4:46

we think things should go from here. So Casey,

4:49

what's your sort of capsule

4:51

theory about why the media is

4:53

struggling so badly right now? I

4:56

mean, the internet has just

4:58

profoundly changed both how people find their

5:00

news and how that news is monetized.

5:02

That's the story in a nutshell. We

5:05

can dive into those things, but at

5:07

the end of the day, we're just

5:09

living through a historic

5:11

shift in the way that people get

5:13

their information. And I do think unfortunately

5:15

that a lot of folks in the

5:17

media industry have been too slow to

5:19

wake up to those changes and get out in

5:21

front of them. Yeah, I would agree with

5:24

that. I think there have been a lot of media companies

5:26

that have sort of clung to business

5:28

models long after it was clear that

5:30

those business models were endangered. I don't

5:32

think we've seen as much like experimentation

5:35

from some of the media companies as I

5:37

might've liked. But I

5:40

don't wanna put this all at the feet of

5:42

the journalists and the media companies who are trying

5:44

to navigate this because there are some larger shifts

5:46

going on in the world that I think relate

5:48

directly to the struggles that the media business is

5:50

having. Yeah, and I mean, the journalists are doing

5:52

great. One of the things that makes these stories

5:54

so upsetting is that none of these journalists are

5:56

losing their jobs because they were bad at being journalists. They're all great

5:58

at being journalists. It was just their boss. woke up one day and said,

6:00

I don't want to pay for you anymore. So anyways, well,

6:02

look, there's a bunch of things we can

6:05

talk about, but maybe let's just start with

6:07

ads, Kevin. Would that be a fair place

6:09

to start? Yeah, I think this is the

6:11

sort of the sort of classic first explanation,

6:13

the sort of prime mover in all of

6:15

this is that the media industry has basically

6:17

been reliant on ads and, you know, local

6:19

businesses in their city advertising,

6:21

you know, use cars or apartments

6:24

for rent, classifieds were a big

6:26

source of revenue. And we

6:28

know what happened with all of that. It all moved

6:30

online to platforms like Craigslist and

6:33

Facebook. And all of

6:35

a sudden, you just didn't have that sort of plank

6:37

in your business model if you're a media

6:39

company. Yeah, well, also, the internet just enabled

6:41

much more effective forms of advertising, right? It

6:43

used to be that, well, if I wanted

6:46

to sell a used car in Phoenix, Arizona,

6:48

I would buy an ad in the Arizona

6:50

Republic, the local newspaper there, and I would

6:52

just trust that whoever was looking for a

6:54

used car there would find it. Well, now

6:56

I can just go on to one of

6:58

these big tech platforms and say, show me

7:00

people who you think are actively in the

7:02

market for used cars who make this much

7:04

money, who live in this zip code, and

7:06

you can reach them directly. And

7:09

it is Google and Facebook in particular that

7:11

figured out that particular mousetrap. And so over

7:13

the past 20 years, we've just seen the

7:15

two of them build these incredible ad businesses.

7:17

And it has come at the expense of

7:19

those media companies that used to make money

7:21

off of used cars and everything else you

7:23

just described. Right. But that's like a fairly

7:26

old story that's been going on for 20

7:28

years at this point. I think

7:30

the question I have is why things seem to

7:32

be falling off a cliff right now. And I

7:34

think, you know, looking into this a little bit

7:36

talking to some folks at different publications

7:38

that have been affected by the

7:40

ad market, it really seems

7:43

like there's been a pullback in the

7:45

broader ad market by advertisers. They are

7:47

not spending as much on any ads,

7:49

whether it's Google, or Facebook, or, you

7:52

know, newspapers and magazines or TV stations. But

7:55

I think the second big sort of

7:57

macro trend that can help us understand

7:59

what's going on in the media business right now,

8:02

is the changing nature of the

8:04

relationship between publications, people who put

8:06

things on the internet, and the

8:09

platforms that distribute that information. Well,

8:11

interesting you say that. I

8:13

thought you were gonna say the changing relationship

8:15

between publications and

8:17

their audiences. Well, say more about that.

8:19

So I think in the early days,

8:22

the internet seemed really exciting, because all

8:24

of a sudden, if you were a

8:26

local publication, your total addressable market went

8:29

from Phoenix, Arizona to the entire world,

8:31

right? And so stories that once might have gotten

8:33

a few hundred or thousand views, all of a

8:35

sudden could get millions. That was really exciting. But

8:38

when social networks came along, and

8:40

as Google became more predominant, people

8:42

stopped visiting the home pages of

8:45

those publications. And so they could

8:47

no longer rely on that direct

8:49

relationship they had with their audiences.

8:52

And as people became more used to

8:54

using Facebook and using Google, that is

8:56

where all of the traffic started to

8:58

come from, which was really exciting, because

9:00

you could get those millions of views, but

9:03

at the same time, it was destroying the

9:05

relationship between the publisher and the audience. So

9:07

publishers no longer really had audiences. What they

9:09

had instead was traffic. They had a steady

9:11

stream of drive-by users who were coming in.

9:13

Often, I believe, people didn't even know what

9:15

publications they were reading, because it was just

9:17

the first thing that popped up on Google,

9:19

or it was just a little link they

9:21

happened to see on Facebook. And

9:23

for a while, again, publishers were OK with

9:25

this, because they were seeing more views for

9:27

stories than they'd ever seen in their entire

9:29

lives. And if you've been in

9:31

the advertising business, really goal number one is

9:33

just to aggregate as many eyeballs as you

9:35

could. And so in the 2010s, you see

9:38

the rise of these publishers like BuzzFeed, which

9:40

were getting hundreds of millions of page views

9:42

every single month, and which looked like they

9:44

were going to be unbeatable, sort of iconic

9:46

media companies for decades to come. But

9:48

then the spigot got turned off, and that is

9:51

what started a lot of the problems we're seeing

9:53

today. Yeah, and I think for people who are

9:55

not in the news business, it can be kind

9:57

of hard to sort of imagine. how

10:01

influential these tech platforms are and

10:04

how one tweak made by

10:06

a team of engineers at

10:08

Facebook or Google can really

10:10

cascade throughout the entire news

10:12

ecosystem. So, for example,

10:15

Facebook in recent years has made a

10:17

conscious and deliberate decision to sort of

10:19

pivot away from news. They

10:21

were sort of mad at the media

10:23

industry for being so mean to them in coverage.

10:26

This is real. Well,

10:28

I mean, they also just got like

10:31

unending blowback from everyone who didn't like

10:33

everything about news on Facebook. Right. Right.

10:36

And I think we should also say that the

10:38

people who run these some of these tech platforms

10:40

have made a conscious decision to deprioritize news for

10:43

their users. So Facebook famously has

10:46

pivoted away from news. They were

10:48

getting all kinds of criticism for,

10:50

you know, promoting and boosting kind

10:52

of polarizing political content around the

10:54

2016 and 2020 elections. They

10:58

were sort of responding to users

11:00

who were complaining about all this, like

11:02

polarizing news showing up in their feeds.

11:05

They were also mad at the news industry for

11:07

being mean to them in coverage over the years.

11:09

And so Facebook decided to dial

11:11

down news so that people who were going

11:13

onto Facebook would see less of it. And

11:16

that has had a tremendous impact

11:18

on digital publishers. There

11:20

have been some recent reports that

11:22

traffic to news publishers from social

11:25

media apps like Facebook and

11:27

Instagram has plummeted. CNBC recently

11:29

reported using data from Chartbeat that

11:32

meta owned apps were accounting for

11:34

just 33 percent of

11:36

overall social media traffic to news publishers.

11:38

That was down from 50 percent

11:41

just a year ago. So

11:43

that's sort of the industry wide figure. But

11:45

at certain publishers, especially publishers

11:47

that focus on political news or that

11:49

have a more sort of partisan point

11:52

of view, the declines you're

11:54

hearing about or even steeper, Mother Jones,

11:56

these sort of progressive political outlet based here

11:58

in the Bay Area. said that

12:01

it had experienced a 99% plunge

12:03

in traffic from Facebook compared to

12:05

several years ago. You

12:07

know, I'm no digital traffic expert, but a 99% plunge

12:09

seems bad. Yeah.

12:12

And, you know, there are a couple things

12:15

I want to say about what you just

12:17

said. One is that, in addition to your

12:19

recap, which I agree with everything, some countries

12:22

have also started to pass laws saying that

12:24

if companies like Facebook and Google want to

12:26

include links to publications on their platforms, they're

12:28

going to need to strike deals with

12:30

publishers essentially to just pay them off and

12:32

make them go away. And I

12:35

think that this has actually been another reason

12:37

why Facebook in particular has said the heck

12:39

with you guys, because we've already seen we

12:41

can take news off our platform and our users don't care.

12:44

So that's something. The other thing that I

12:46

would say is, while I think it is

12:48

great that there was this period where particularly

12:50

nonprofit outlets like Mother Jones could

12:52

do a bunch of fundraising by

12:54

using this cheap or free Facebook

12:56

traffic that they were getting, at

12:59

the same time, it got too

13:01

many publishers addicted to this firehose.

13:04

And as they were getting addicted to it, they

13:06

were losing those direct connections with their audience. What

13:09

do I mean by direct connection? I mean something

13:11

that is not being mediated by some sort

13:13

of ranking algorithm, right? So I did

13:15

start a newsletter because I did see

13:17

the Facebook thing in particular and it

13:19

scared me. If digital

13:21

media is going to have a future, it

13:24

is going to be that readers know what

13:26

publication they are reading when they are reading

13:28

it. They feel positively about it and they

13:30

want to support it in some way. So

13:33

I do think a lot of what we've

13:35

seen in the past couple of weeks is

13:37

that particular chicken coming home to roost is

13:39

that there are now publications that got addicted

13:42

and got huge on the backs of this

13:44

traffic that was being delivered to them by algorithms.

13:46

And when the algorithm changed, the rug got pulled

13:48

out from underneath them. Right. So

13:50

the companies right now are saying like

13:52

we don't want to be dependent on

13:54

Facebook and Google for our continued existence.

13:57

They're looking to pivot to different kinds

13:59

of distribution strategies. that aren't as

14:01

reliant on the staying in the good graces

14:03

of social platforms, but it's been a really

14:05

hard transition for a lot of Publications and

14:07

they're not going fast enough and and we

14:10

should talk about what what is happening with

14:12

Google as well Google is the other big

14:14

fire hose of traffic and even as Facebook

14:16

has declined as a traffic source Google is

14:19

still right up there and it whatever publications

14:21

you are reading listeners I guarantee you that

14:23

Google is providing a huge amount of traffic

14:25

to those websites and the traffic is usually

14:28

not somebody Googling the name of the publication

14:30

it is them googling a subject or how

14:33

to do something or looking for the best

14:35

laptop of 2024

14:37

and so much of the digital media

14:39

infrastructure that we have is supported by

14:42

Google sending that traffic But for reasons

14:44

that we should get into right now

14:46

that the nature of that traffic is

14:48

quite Procarious and I think every publisher

14:51

should assume that very soon. It is

14:53

going to start declining quite rapidly Yeah,

14:56

and let's talk about that Google factor

14:59

But first I just want to acknowledge this sort of

15:01

elephant in the room, which is that these problems are

15:03

not affecting all Publishers in

15:05

the entire media ecosystem equally right

15:07

Ezra Klein I call it the

15:09

times had a great column recently

15:12

about the sort of news Apocalypse

15:14

and he pointed out that these

15:16

changes that you're describing to the

15:19

news ecosystem They have hit the sort of

15:21

middle of the news industry the hardest You

15:23

know, the New York Times is doing quite

15:25

well by the standards of the news industry

15:29

Also, there's a real bright spot in your

15:31

part of the news industry Which are these

15:33

sort of like newsletters that go direct to

15:35

consumer that you pay for as a subscriber

15:38

That business is still quite good So

15:40

we are in some ways sitting at

15:42

the ends of this kind of barbell

15:44

shaped news ecosystem right now We

15:47

are the lucky ones. We say we feel

15:50

that we feel grateful for that totally But

15:52

there's this entire world out there of kind

15:54

of midsized publications that is really really struggling

15:57

right now So let's

15:59

talk about the AI piece of this because I think this is a

16:01

part of the conversation that is starting to be discussed

16:04

is the effect that AI and

16:06

generative AI are having on the

16:08

business models of what remains of

16:10

the media industry. So you recently

16:12

wrote about Pitchfork, the music publication

16:14

that was folded into GQ, and

16:17

you sort of made a case that this is in

16:19

some ways related to the rise of AI. Walk

16:22

me through your argument there. Yes, I

16:24

should say I'm a huge fan of

16:26

Pitchfork and while it remains to be

16:28

seen what happens to it, it does still exist in some

16:30

form and I'm grateful at least for that. Pitchfork has

16:33

turned me on to so much amazing music over the

16:35

years. The

16:37

thing that younger folks may not recall is that

16:39

new music used to cost $18. You used to

16:41

have to go to a store and buy it.

16:44

And so that created essentially a job for

16:46

Pitchfork as a publication to go out there,

16:48

scour the landscape, listen to everything, and then

16:51

come back and say, hey, here is what

16:53

is actually worth your time and money. And

16:56

then in 2011, Spotify launches in

16:58

the United States and all of a sudden,

17:00

me, a person who used to say

17:02

like, oh, yeah, like, yeah, okay, that's getting a good review.

17:04

Like is this worth going out and spending money on? Okay,

17:06

I'll spend some amount of money on this. All

17:09

of a sudden it was like money was

17:11

out of the equation. I had my Spotify

17:13

for 10 bucks a month. I was listening

17:15

to anything I wanted for pennies a day.

17:17

And all of a sudden, I didn't feel

17:19

that same need to seek out a Pitchfork

17:21

review because as good as the writing is

17:23

and remains, it wasn't serving that same function

17:25

that it used to for me. I used

17:27

to look to pitchfork to sort of educate

17:29

me musically, tell me about artists from

17:31

the 60s, 70s, 80s, before I was paying

17:33

attention to pop music and did an amazing

17:36

job at that, right? Spotify,

17:38

for the most part, would show me new stuff. Well,

17:40

Spotify, as it has improved its own AI

17:42

systems, is now really good at saying like,

17:45

oh, hey, Casey, like you like the Beatles,

17:47

like here are some other acts from the

17:49

60s that were sort of like

17:51

contemporaries of theirs that made some great music

17:53

as well. And so now Spotify is playing

17:56

the role that a music publication that once had

17:58

a team of 20, 30, 50. Who knows

18:00

how many people who were all trying to

18:02

educate me about music that is now just

18:04

all being done by math. And I don't

18:07

want to say that it is perfect or

18:09

even better than the human beings. But in

18:11

practice, because Spotify is the point where I'm

18:13

consuming all music already, it is just effortless

18:16

for me to let it do that job. And

18:18

I think it creates a kind of existential crisis if

18:20

you're a music critic, it is

18:22

asking you to effectively change your job. That's an

18:25

editorial problem more than it is a business problem.

18:27

But I do think we should point it out

18:29

because the story of tech and

18:31

media isn't just the money dried out.

18:33

It is that as we go through

18:35

these technological shifts, they call for new

18:38

kinds of publications to answer that moment.

18:40

Yeah, so I think that's sort of

18:42

the first shoe to drop with AI

18:44

and the news business is kind of

18:46

the fact that people are now turning

18:48

to AI to kind of curate their

18:50

information, but also their taste. The sort

18:52

of new taste makers are not critics

18:55

and publications. They are algorithms and AI

18:57

DJs. And I think that you can

18:59

extrapolate that across the industry. But

19:03

there's this other newer factor that I

19:05

think is starting to become a real

19:07

issue, which is that generative AI is

19:09

rising up through search. So

19:12

if you do search on Google

19:14

for something like Best Laptop 2024, some of the

19:16

results that

19:19

you might get are in fact

19:21

AI generated sort of pages from

19:23

these kind of AI content farms.

19:26

And I think we should explain like why

19:28

that's a potentially very bad thing for

19:30

traditional publishers. Yeah. So

19:33

like Best Laptop 2024 is one of the

19:35

most valuable searches since there probably is on

19:37

Google. People pay a lot of money

19:39

to be the first results or the first ad

19:41

shown to people who search for that term. That's

19:43

one reason. The second reason is if you are

19:45

a publication and you're even vaguely associated with tech,

19:48

you have a huge incentive to write an article

19:50

called Best Laptop of 2024. And

19:53

whatever laptop you decide, you're going to put an

19:55

affiliate link on that. So if I'm in the

19:57

market for a new laptop, I find it on

19:59

some website. I click the link, I

20:01

buy it off of Amazon or whatever, Amazon

20:03

is going to share some of that money

20:05

with the publisher. And this has become a

20:07

huge source of revenue for publishers over the

20:09

past decade or so. As the advertising market

20:11

has declined for them, affiliate link revenue has

20:13

been a bright spot. It's helped

20:16

to prop up a lot of these publications, helped to preserve a

20:18

lot of jobs and journalism. Well now

20:20

you introduce AI into the equation. And Google

20:22

has this product, SGE, search generative experience. And

20:24

it gets that best laptop 2024. It

20:27

could just show you 10 websites that wrote a

20:30

best laptop 2024 article. Or

20:32

it could summarize them, pick out its own best

20:34

laptop 2024, show its own affiliate link and keep

20:36

all of the revenue for itself. And the question

20:39

that you should be asking if you're a publisher

20:41

is why would Google continue to send traffic to

20:43

you in the future when it can just keep

20:45

the money for itself? Right.

20:47

So I think a lot of publishers are

20:50

very nervous about these kind of generative AI

20:52

powered search engines. The New York Times actually

20:54

cited an example very much like the one

20:56

you just provided about the best laptop of

20:58

2024 in its lawsuit against open

21:00

AI and Microsoft that was filed late

21:03

last year. I'm basically saying these AI

21:05

chatbots that are hooked up to the

21:07

internet and can go retrieve sort of

21:09

real time data. They are

21:11

essentially a substitute for journalism because

21:14

if you can say, you know,

21:16

what is the wire cutters best

21:18

recommendation for a, you know, a

21:20

ceiling fan? Is that

21:22

a... Yeah. I

21:25

will not buy a ceiling fan without checking with a

21:27

wire cutter. Right. So if you go to

21:29

the wire cutter... That article was decapitated by a bad one. So

21:32

if you go to the wire cutter looking

21:34

for their best ceiling fan recommendation and you

21:36

end up buying a ceiling fan through that

21:38

affiliate link, the New York Times, which owns

21:40

wire cutter may get a cut of that.

21:42

But if you can just go to chat

21:44

GPT with browsing enabled or one of these

21:46

other AI search products, you might

21:48

just get that kind of pulled out from

21:51

the wire cutter article with no affiliate link

21:53

and no revenue shared back with the

21:55

wire cutter. Yeah, that's right. So, okay,

21:57

to basically sum up the sort

21:59

of diagnosis... of the problem here. We have

22:01

these kind of challenges with

22:04

media that date back at least a

22:06

decade and probably two of sort of

22:08

the decline of their advertising businesses, a

22:11

shift away from having direct relationships

22:13

with readers to now being sort

22:15

of intermediated by these tech platforms.

22:17

And we now have these AI

22:20

tools that are threatening various facets

22:22

of the industry from AI generated

22:25

spam showing up in Google results to chat

22:28

bots that can just kind of summarize information

22:30

rather than actually directing people to the websites

22:32

of publishers. And we didn't even talk about

22:34

the fact that like AI spam is

22:36

now outperforming like so much of them in new

22:39

stories. Yeah, say more about

22:41

that. Well, so this month there

22:43

was this group of researchers based in Germany

22:45

who released a year long study. It was

22:47

called, is Google getting worse? A longitudinal investigation

22:49

of SEO spam in search engines. And

22:52

they were trying to figure out if they

22:54

could measure empirically whether Google search results were

22:56

getting worse by turning up

22:58

more low quality content. So they researched a

23:01

bunch of product searches to sort of like

23:03

the best laptop 2024 thing. And

23:06

they found that yes, Google results for

23:08

product terms did seem to pull up

23:10

a lot of spam and low quality

23:12

content. My strong suspicion is that a

23:14

lot of this was AI generated and

23:17

if it wasn't already, certainly it will

23:19

be in the future. And so now

23:21

if you are the publication that's still

23:23

writing best laptop 2024 articles, you're

23:26

not just competing against the many other newsrooms on

23:28

the internet that are creating that sort of thing.

23:30

You're now also going to be competing against AIs

23:32

and it could be a lot of AIs. It

23:35

could be infinite AIs, right? And that is a

23:37

challenge that I think Google has truly not even

23:39

begun to answer in part because they've said, we're

23:41

basically okay with AI created webpage. We don't have

23:43

a rule against AI created webpages. I'm not even

23:45

sure that they should have a rule against AI

23:47

created webpages, but I do know that they have

23:50

a strong incentive to have good search results, at

23:52

least until they replace all their search results with

23:54

AI. This

23:56

seems like a short term problem. It will eventually be

23:58

replaced by a much bigger, longer. problem.

24:01

So if that's the diagnosis of what

24:03

is happening in the media industry today,

24:06

what is the prescription? Like where

24:08

can news publishers look if they

24:11

want to build sustainable businesses in

24:13

the year of our Lord 2024? I think it all has

24:17

to start with reimagining the cost

24:19

structure of a business. So we

24:21

used to think about media businesses

24:23

in terms of we will hire

24:26

a newsroom of this many reporters.

24:28

We will have a back office

24:30

function with human resources and book

24:32

keepers and accountants and an in-house

24:35

lawyer and everything. And if

24:37

you are a medium-sized publication or you want

24:39

to start a medium-sized publication, that I think

24:41

is the thing that is just the hardest

24:44

to do right now. We do

24:46

not really have any visibility into

24:48

how you do that sustainably, right?

24:50

So where I would like to

24:52

see more attention going is toward

24:54

these very small newsrooms and

24:57

these journalist-owned collectives. One of my

24:59

favorite stories in media, one that

25:01

makes me most optimistic, is this

25:03

site called Defector, which was started

25:06

by a bunch of former writers

25:08

at Deadspin, the old sports blog.

25:11

And these folks were so talented and their

25:13

audience loved them so much. They had a very

25:15

authentic connection with our audience who loved the way

25:17

that they talked about sports and politics. And so

25:19

when they started up their own thing and they

25:21

say, hey, come support us, a lot of people

25:23

were willing to come, show up, and pay them.

25:25

And they actually share all their financials every year,

25:27

which I think is amazing. And in

25:30

their most recent report, you can see they're doing

25:32

pretty well. They share most of the revenue they

25:34

make with the people who are writing the website.

25:36

It feels like they have found a path forward.

25:39

It's not maybe as big as

25:41

like Pitchfork was. It's probably not

25:43

as big as Deadspin, but it is

25:45

bigger than Platformer and it is doing

25:48

meaningful work. We're seeing other people adopting

25:50

this model. Some of my favorite reporters

25:52

from Vice News' old Motherboard vertical, which

25:54

was their vertical that covered tech, they

25:56

recently started something called 404 Media.

25:58

It's the same thing. very small nucleus

26:00

of incredibly talented reporters. And they are

26:03

now just asking people to come support

26:05

their work directly. They've already broken so

26:07

much news. So, you know, if you

26:09

are a journalist and you want to

26:11

find a sustainable path in this industry,

26:13

I would be thinking about these collectives,

26:15

you know, I also am a big

26:17

believer in the, the, the, the solo

26:19

creator model or the maybe one, two,

26:21

three person newsroom model, which I'm like

26:23

trying out myself. Now, I talk with

26:25

journalists all the time who are considering

26:27

making that move. I have been trying

26:29

to provide the best assistance I can to some

26:31

folks that are and some of them have actually

26:33

thrived. And, you know, the final thing I would

26:35

say Kevin is I would love to see the

26:37

millionaires and the billionaires take a look at what

26:40

folks like defector and 404 are doing, right? The

26:42

classic move for a billionaire is to be like,

26:44

what is some, uh, prestigious

26:46

old property that's losing a lot of money that

26:48

I could take over and maybe help it lose

26:51

slightly less money and just sort of bask in

26:53

the fact that I own the Washington post or

26:55

something like that. Right. These

26:57

folks could lose so much less

26:59

money and could have such a bigger

27:01

impact on journalism and democracy. If they just went

27:03

to 404, they went to defector, they went to

27:05

one of these other collectives and they said, you

27:08

know what, I'm just going to give you $5

27:10

million a year, right? See

27:12

what you could do with it. Maybe you decide you want to hire 10 more

27:15

people. Maybe you want to invest in a

27:17

really cool website. Maybe you want to watch

27:19

some really expensive investigations, right? But man, if

27:21

you are a rich person and you just

27:23

want to be a patron of some journalism

27:25

organization, I'm telling you it is in these

27:28

small collectives where you are going to get

27:30

so much more bang for your buck. Yeah. I

27:32

think there's something to that. I think

27:34

especially for local and regional news, it

27:36

makes sense to consider alternative

27:39

ways of funding yourself. One of my

27:41

favorite examples of a local news startup

27:43

that I really love, um,

27:46

is, uh, city side, which is out in

27:48

the Bay area where I live. Um, there's

27:50

a, it started with a website called Berkeley

27:52

side, which was sort of a hyper local,

27:54

uh, digital publisher that was sort of just

27:56

a handful of people who were sort of

27:59

covering, you know, news in

28:01

Berkeley, California. There's now a publication

28:03

called Oakland side. So they're expanding

28:05

to other cities in the area.

28:08

And they really have sort of an

28:10

interesting mix of funding, you know, some

28:12

of it comes from grants and foundations.

28:15

Some of it comes from kind of

28:17

wealthy benefactors who decide to support this as

28:19

sort of a quasi philanthropic exercise. And I

28:21

think it's just a really interesting model for

28:23

local news. I don't know if that

28:25

works at the same way on the kind

28:28

of national or global scale. And

28:30

so my recommendation for the news

28:33

industry, when it comes to these

28:35

bigger national and global publications,

28:39

is to sort of get back to basics,

28:41

right? There's this famous saying in

28:43

Silicon Valley, it's sort of the motto of the

28:45

Y Combinator startup accelerator, they

28:47

like to say make something people

28:49

want. And I

28:51

think that has to be at the center

28:54

of what media organizations are trying to do.

28:56

We cannot have a sustainable news industry

28:58

if it is dependent on the largess

29:01

of billionaires, on the availability of grants,

29:03

I've even heard some people talk about,

29:05

you know, maybe the government should fund

29:07

news production, maybe we should have employers

29:10

bundling subscriptions to, you know, offering subscriptions

29:12

to news publications as part of their

29:15

employee benefits packages. And I just really

29:17

think that all of those strategies are

29:19

sort of missing the fundamental point, which

29:21

is that in order to succeed, you

29:23

have to make something that people want

29:26

to pay for people should not have

29:28

to be guilted, or pressured or cajoled

29:30

into subscribing to their favorite favorite publication,

29:32

it should be something that they want

29:34

to do. And I

29:37

think in the last decade or two, since

29:39

I've been in media, journalists

29:41

have really lost a sense of kind

29:43

of entrepreneurship and creativity, it's almost become

29:45

kind of a dirty word in certain

29:48

parts of the media industry to care

29:50

too much or to know too much

29:52

about how your company or your journalism

29:54

organization makes money. We've sort of cleaved

29:56

that off into like a separate part

29:58

of the industry and look,

30:00

there are good reasons for that.

30:02

I don't necessarily want the investigative

30:04

reporters covering Congress

30:06

to be thinking about their web

30:09

traffic or how many subscriptions their

30:11

stories are converting. But

30:13

I do think that we as an industry could

30:15

benefit from a little more entrepreneurial thinking and

30:18

just trying stuff and thinking about not only what

30:20

is the best story to tell here, but what

30:23

is the thing that people are actually willing to

30:25

pay for? Yeah, I wanna say something

30:27

about the make people want point because

30:30

it is a really important one. For a

30:32

long time now, when new publications have been

30:34

started, it is often started around an advertising

30:36

category, right? Like I'm not gonna name the

30:38

publications over the past years that were started

30:41

by companies thinking like we need a tech

30:43

vertical, right? And they would spin something up,

30:45

they would lavishly fund it for a while,

30:47

but then once the ads didn't turn out

30:49

to be there, the plug got pulled on

30:52

it. Well, that was because there's already a

30:54

lot of tech coverage. And if

30:56

you don't show up in the market with some

30:58

sort of unique editorial insight, the market is probably

31:00

gonna reject you unless you just happen to have

31:02

the most talented people in the world working there.

31:04

So, and again, this is like one

31:06

of the reasons why media is such a hard business

31:08

is it is not even

31:11

enough to have a good business model,

31:13

which is already enough of a challenge

31:15

on its own. You have to have

31:17

an existential reason for being and it

31:19

is just in the nature of media

31:21

for the existential reason you find to

31:24

not last that long. You actually have to

31:26

constantly reinvent yourself. When I started writing

31:28

the newsletter, it was just about that day's problem at Facebook

31:30

and there were a lot of them and it

31:32

sustained my newsletter for years. But then one day

31:35

that was an old story and the newsletter had

31:37

to be about something else, right? And publications find

31:39

themselves in that spot all the time and it

31:41

is super hard, but if we want to rebuild

31:43

that middle of media, we need folks that build

31:46

publications that have a dedicated reason for being and

31:48

they need to be able to evolve it over

31:50

time. Yeah. You know, Casey,

31:52

I think a point that often goes like

31:54

under appreciated in these discussions is that news

31:57

as kind of a standalone product has very

31:59

rarely been. profitable on its own

32:01

without some kind of adjacent business kind of

32:03

subsidizing it. So there used to be this

32:05

saying at the New York Times, I don't

32:07

know if it was ever true or not,

32:10

that the crossword puzzle paid for the Baghdad

32:12

Bureau, right, you have these like kind of

32:14

lightweight sections, like games and

32:16

puzzles and comics that sort of get people

32:18

in and then you use the revenue generated

32:20

by those things to pay for the very

32:22

expensive, you know, foreign desk. And

32:24

this is true at a lot of different kinds

32:27

of publications and

32:29

media organizations, you know,

32:31

radio stations hosting music

32:33

festivals. I remember a

32:35

few years ago when Buzzfeed's Tasty section,

32:37

their sort of cooking recipe section

32:40

of their website got very popular, they

32:42

sort of started a line of cookware

32:44

that they would sell in stores. So

32:46

there have always been these kind of

32:48

attempts by media organizations to sort of

32:50

use an adjacent business to subsidize the

32:52

the news gathering one. Yeah, which is

32:55

why we're adding comics to platformers. So

32:57

look forward to Marmaduke starting next week.

33:00

So what is it what is an adjacent product that

33:03

you think could make money for news organizations today? I

33:07

think pro wrestling obviously comes

33:09

to mind. Honestly, it could work like

33:11

you could pay people would pay to

33:13

see, you know, David Brooks and Maureen

33:15

Dowd go at it. I mean, you

33:17

know, so much of digital

33:19

media is just people dunking on each other. Why not

33:22

put it in the squared circle and, you

33:24

know, see everyone at SummerSlam. I love

33:26

this occasion match. That's

33:29

better than my idea, which was an app

33:31

that would allow you to tap the expertise

33:33

of world class writers to write your dating

33:36

app profiles and messages for you.

33:39

Call this Cyrano. Yes, writers

33:41

famously good at dating and interrelationship.

33:44

Those are the people you

33:46

want writing that for you.

33:49

When we come back, crypto investor

33:51

Chris Dixon talks about blockchains, web

33:54

three and where the industry went

33:56

wrong. AI

34:07

may be the most important new

34:09

computer technology ever, but AI

34:11

needs a lot of processing speed,

34:13

and that gets expensive fast. Upgrade

34:16

to the next generation of the cloud. Oracle

34:19

Cloud Infrastructure, or OCI. OCI

34:22

is the single platform for

34:24

your infrastructure, database, application development,

34:26

and AI needs. Learn

34:29

more and spend less like Uber, 8x8,

34:32

and Databricks Mosaic. Take

34:34

a free test drive of

34:36

OCI at oracle.com/NYT. oracle.com/NYT.

34:41

I'm Julian Barnes. I'm an intelligence reporter

34:43

at the New York Times. I try

34:45

to find out what the US government

34:47

is keeping secret. Governments

34:50

keep secrets for all kinds of reasons. They

34:52

might be embarrassed by the information. They

34:55

might think the public can't understand it.

34:57

But we at the New York Times

34:59

think that democracy works best when the

35:01

public is informed. It

35:03

takes a lot of time to find

35:06

people willing to talk about those secrets.

35:08

Many people with information have a certain

35:10

agenda or have a certain angle, and

35:12

that's why it requires talking to a

35:14

lot of people to make sure that

35:16

we're not misled and that we give

35:18

a complete story to our readers. If

35:20

the New York Times was not reporting

35:23

these stories, some of them might never

35:25

come to light. If you

35:27

want to support this kind of work, you can

35:29

do that by subscribing to the New York Times.

35:35

So Casey, we have talked a

35:37

lot on this podcast about what's

35:39

happening in the cryptocurrency world and

35:41

all of the various scandals and

35:43

misdeeds and Ponzi schemes and collapses

35:45

that have befallen big crypto companies

35:47

over the last few years. We

35:49

do. In fact, let's do a live

35:51

update on the price of Dogecoin, Kevin. Right now, as we record

35:53

this, the price of Dogecoin is 7.9 cents. So

35:56

I hope that helps you as you make your investment decisions. We

36:00

spend a lot of time joking and

36:02

being glib about what's going on in

36:04

crypto because there have been just so

36:06

many disasters over the past few years,

36:08

especially. But I think we

36:10

also have to treat this as a serious

36:12

subject because there are a lot of people

36:15

in Silicon Valley still to this day who

36:17

believe that despite all the scandals, despite the

36:19

fact that, you know, FTX and Binance have

36:21

collapsed and Sam Bankman freeds in jail and

36:24

the crypto market, you know, is a fraction

36:26

of its former self. There are

36:28

people who are influential in the tech industry who

36:30

think that all this is due for

36:33

a rebound and that crypto will actually

36:35

be an important part of our technological

36:37

future. And would you say that one

36:39

of the things that those folks have in common is that they

36:42

have a lot of money riding on the outcome? Some

36:44

of them do, yes. And some of

36:47

them just believe in kind of the

36:49

theoretical underpinnings of crypto, the sort of

36:51

idea of decentralization and taking away the

36:53

intermediaries of big tech who have been

36:56

controlling the experiences online of billions of

36:58

people. So this

37:00

week we had an opportunity to talk to

37:02

one of the sort of enduring crypto optimists

37:05

in Silicon Valley. And I think he's actually

37:07

someone who both you and I have respected

37:09

and relied on as a source of wisdom

37:11

and information for a long time. And

37:14

that's Chris Dixon. Chris is

37:16

a partner at Andreessen Horowitz, the

37:18

venture capital firm founded by Mark

37:20

Andreessen and Ben Horowitz, who has

37:23

become one of the biggest investors

37:25

in crypto and Web3. The

37:28

information called him Andreessen Horowitz is

37:30

crypto king. In addition to being

37:32

an investor, Chris is also someone who's been

37:34

blogging and thinking about the future of the

37:36

Internet for a very long time. And

37:39

his firm has led investments in

37:41

a bunch of leading crypto companies,

37:43

including Coinbase, OpenSea and Yuga Labs,

37:45

which is the company that makes

37:47

the board ape yacht club and

37:50

collection. And to

37:52

this day, he remains optimistic that crypto

37:54

is not dead. In fact, blockchain

37:57

networks, which is what he prefers to call it,

37:59

are going to be a big to be an

38:01

important part of the future of the Internet and

38:04

kind of a way for ordinary people to take

38:06

back power from some of these centralized tech platforms.

38:08

Yeah. And I should think,

38:10

Evan, like I actually just don't share that view

38:12

at all. I feel burned by everything that I

38:15

wrote about crypto. And Chris was

38:17

one of the people who made me take crypto

38:19

seriously because I thought, look, if this person is

38:21

serious about it, then I need to assume that

38:23

something positive is going to come out of this.

38:25

And I have not had a chance to talk

38:27

to him since all of this stuff started to

38:29

shake out in the market. So when we found

38:31

out that he had a new book coming out, we

38:33

said, well, maybe this is the chance where we can

38:35

ask him the hard questions that we've had. Yeah. So

38:38

Chris just wrote a new book sort of detailing his vision

38:40

for the future of crypto and blockchains. It's called Read, Write,

38:42

Own. And I thought, well, maybe this

38:44

is our chance to sort of ask a

38:46

person who's really in some ways responsible for

38:48

a lot of the big investments that have

38:50

propped up the crypto world. Maybe

38:53

this is a chance to understand why

38:55

he's still optimistic about where crypto is

38:57

headed, get all of the

38:59

evidence that suggests that this industry is in

39:02

real trouble. Chris

39:07

Dixon, welcome to Hard Fork. Thanks for having me.

39:10

So we're going to talk about your new book, Read,

39:12

Write, Own. But first I want

39:14

to recall or recollect a conversation that you and

39:16

I had about 10 years ago.

39:20

I don't know if you remember this, but we were, you know, I'd

39:22

been following. I think we were in SF or something. I remember being

39:24

on a peer or something. Is that right? Is

39:26

this correct or no? Yeah, we were getting

39:28

coffee at the Ferry Building in San Francisco.

39:31

Uh oh. Did I say something that that was

39:34

I made a bad prediction? Well,

39:37

you told me that you'd gotten really

39:39

interested in crypto. And I remember

39:42

this conversation because this was during the first

39:44

sort of big Bitcoin boom. And

39:47

there was a lot of exuberance. People were

39:49

making tons of money. And I

39:51

remember you telling me like you guys

39:53

in the tech media, you need to

39:55

pay attention to crypto because it's a

39:57

really fascinating story and it's a lot.

40:00

bigger than just Bitcoin. Like these blockchain

40:02

things, they're going to be a really

40:04

big deal. A lot of the smartest

40:06

programmers in Silicon Valley are working on

40:08

this stuff. And it's going to turn

40:10

into real valuable products that will sort

40:12

of revolutionize the tech industry. And I

40:14

remember that conversation for me was a

40:16

moment where I started to kind of

40:19

look at crypto more seriously and cover

40:21

it in a more serious way. And

40:25

fast forward a decade, I think you

40:27

were right that crypto has been a

40:29

very interesting story, but I think not

40:32

for the reasons that you and I

40:34

would have predicted a decade ago. You

40:36

know, we've got these massive scams, exchanges

40:39

collapsing, executives going to jail, you know,

40:41

billions of dollars being made and lost

40:43

overnight. And meanwhile,

40:46

as I kind of look around,

40:48

I can't actually see any tangible

40:50

examples of where blockchains have transformed

40:52

much of anything. And a lot

40:54

of the really smart programmers that

40:57

you were talking about have left

40:59

crypto and are pivoting to building

41:01

AI. So my first question

41:03

to you is just what happened? Yeah,

41:06

like I do think it went off. I mean, I'm not disagreeing

41:09

with you. I think it went it did not go

41:11

the way that I wanted it to go. To do

41:13

something you want to get into it, I would describe

41:15

the last kind of 2020 to 2023 as starting

41:18

off very promising for somebody like me who

41:20

was interested in applications and utility. I thought

41:22

there were some really interesting things built. And

41:24

then I think it went off the rails. And honestly,

41:27

like that, like part of writing this book, like I

41:29

started writing this book after that, I was like, reflecting

41:32

on, do I want to keep

41:34

doing this? Like, I don't want to be part of

41:36

the this sort of casino culture, you know, is that

41:38

something fundamental to the technology? Or is it just

41:41

so far the wrong people, the wrong products

41:43

have been built? You know, like, I believe

41:45

it's the second, it's the latter of those.

41:47

And that and like, I'll just also put

41:50

it in the context of like, there

41:52

are many millions of people who are very

41:54

excited about the idea that you can have

41:56

a financial service that's that the

41:58

way there is no company behind it. where it's truly

42:00

owned and operated by a community. Like I think the

42:02

one thing I was right about, Kevin, is that idea

42:05

has resonated deeply. And if there's another

42:07

proposal for how to fight consolidation of the Internet,

42:09

like I'm open to it, you know, but this

42:11

is my proposal, I guess. Yeah. And, you know,

42:13

we'll get into could blockchain still realize the vision

42:16

that you had back then? But, you know, I

42:18

just want to say, Chris, like I've known you

42:20

for a long time. You invested in a lot

42:22

of stuff before crypto that I thought was super

42:24

interesting. And when you got into crypto to me,

42:27

this was one of the big signals for me

42:29

as it was for Kevin, it's like it's time

42:31

to take this seriously. And over

42:34

the next few years, I think what happened is I

42:36

tried to be one of the journalists keeping an open

42:38

mind about stuff and not assuming that everything was a

42:40

fraud or a scam. But when I look back on

42:42

what I wrote in twenty one and twenty two, the

42:45

stuff that I covered was at best irrelevant

42:47

and at worst was stuff that people lost

42:49

a whole lot of money on. So when

42:51

you're coming back around with a book and

42:53

saying, like, let's take the blockchain seriously, I

42:56

just have this lingering hangover of like I

42:58

tried it and I like deeply regret what I

43:00

were about before. Yeah, but Casey, I

43:02

mean, in fairness, like I started an company in 2008.

43:05

It's called Hunch. I sold it to eBay in 2011.

43:07

And frankly, it wasn't the success that

43:09

I hoped it was. And if you'd

43:11

asked me in 2012, I would have given you a list

43:13

of reasons why it didn't work. And in reality, it was

43:15

timing. It was timing. And like the first neural network paper

43:18

was 1943. Right.

43:20

It was McCulloch and Piss. And then there was

43:22

Turing and his paper was in 1950. I

43:24

think he predicted in 20 years we'd have, you

43:26

know, intelligent machines pass the Turing test. There

43:29

was a there was a so-called A.I. bubble in the 80s. And

43:32

then it worked. And why did it work? Because, you know,

43:34

there's a lot of smart people, everything else. But a lot

43:36

of it was because of video games, because video games drove

43:38

GPUs and it got better. I think

43:40

that from the outside, you guys are tech journalists

43:43

and you know this. But from the outside view, sometimes these

43:45

things look like they're sort of immaculate conception. In

43:47

reality, there's these messy background stories.

43:50

So so, like, I agree with you. Again,

43:53

like you're right. It hasn't realized its

43:56

potential. Does that mean it will never realize its

43:58

potential? I don't think that's the case. So,

44:01

Chris, this book Read, Write, Own is sort of

44:03

a collection of your

44:05

thoughts about crypto

44:07

and blockchains and what you think

44:09

the future of decentralization might be.

44:13

What is the main argument in a

44:15

couple of sentences of this book? So

44:18

the main argument is that the original

44:20

promise of the internet was to be

44:22

a democratically controlled system where creators, entrepreneurs,

44:24

companies could build direct relationships with their

44:27

audiences. In the meantime,

44:29

a set of large companies

44:32

stepped in the middle and

44:34

became intermediaries. And that's,

44:36

you know, Facebook, Apple, Amazon, Google,

44:38

etc. And that has

44:41

made been incredibly beneficial

44:43

to those companies and I think has significantly

44:46

hurt the rest of the participants of the

44:48

internet. And so I am very concerned that

44:50

the internet is going to become like old

44:53

broadcast TV where you have sort of ABC,

44:55

CBS and NBC. And

44:57

that's it. And I think

44:59

that is a bad outcome for creators,

45:01

for entrepreneurs, by the way, for venture capitalists.

45:03

We are in the business of funding startups

45:06

and having a dynamic internet. And I think we're headed there.

45:09

And then we have this technology come along, which

45:11

I think with blockchains as the kind of

45:14

core argument of the book is there, you

45:16

know, if web one protocols like email or,

45:18

you know, sort of thesis and

45:20

web two systems like Facebook or antithesis, I

45:23

see blockchains as sort of synthesis. They're kind

45:25

of the best of both worlds. I use

45:27

the phrase, the societal benefits

45:30

of protocol networks of early kind of

45:32

web protocol, web and email protocols, but

45:34

some of the competitive advantages of the

45:36

web two networks. That's my argument. And

45:39

so I think you could build. So for example,

45:41

I think a very critical moment that led to

45:44

the consolidation of the internet was the fall of

45:46

RSS was the fact that so in 2007 or

45:48

eight, for those who don't know RSS is an

45:50

open protocol, kind of like email and the web,

45:52

which allows for social networking features. People

45:55

forget this now, but in like 2008 or so

45:58

RSS was a legitimate competitor to. Twitter

46:00

and Facebook. RSS is

46:02

now using podcasting, it's using kind of niche

46:04

things, but it's clearly not like dominant.

46:08

It has no user base that's

46:10

a fraction of the size of Twitter and Facebook and things like

46:12

this. And the aggregate revenue

46:14

of the top social networks last year was $150

46:16

billion. That money went

46:18

to those companies. Had RSS1, I believe

46:21

the vast majority of that money would have

46:23

gone to the edge of the network. That

46:25

has huge economic... So these topics of like,

46:27

how do you architect these networks has

46:29

profound economic consequences.

46:32

It has profound cultural consequences, political

46:34

consequences. Who gets to decide who

46:37

has a platform? Who gets to decide with how

46:39

the algorithms work? Who gets verified? Like

46:42

should it be a $8 thing? Should

46:44

it be because you know a friend at Twitter?

46:46

I think all of these systems, I don't think

46:48

that because you happen to build a popular social

46:50

network, you should get to decide how global economics,

46:52

culture and politics works, which is the system that

46:54

we have today and we're heading towards. So

46:57

the core argument is there is a

46:59

new technology, which admittedly has a sullied

47:01

name because of the casino and other

47:04

kinds of things, but there's

47:06

another side to the coin. He asked for a

47:08

couple sentences and you really couldn't tell him. Yeah,

47:12

those were sentences in like the William

47:14

Faulkner sense of the word. I

47:20

think maybe something that could help make

47:22

this discussion a little less abstract is

47:25

talking about sort of particular projects or

47:27

particular companies. And I want to ask

47:29

you about a company and a project

47:31

in crypto that I wrote about a

47:34

few years ago called Helium.

47:36

This is also a company that

47:38

your firm invested in. This was

47:40

a company. They were making these

47:42

essentially wireless routers that you could

47:44

put in your house or on

47:46

your roof or at your office.

47:48

And these things would basically provide

47:51

connectivity to devices around you.

47:55

And at the same time, they would

47:57

also mine a cryptocurrency token. idea

48:00

was kind of that you would, instead

48:02

of Comcast or whoever, like spending billions

48:04

of dollars to build out a new

48:07

wireless network, you could essentially incentivize ordinary

48:09

people to do this by giving them

48:11

some crypto in exchange for providing these

48:14

routers to the network. And at the

48:16

time, I thought, like, that sounds like

48:18

kind of a cool idea. It's kind

48:20

of an original way to bootstrap a

48:22

kind of big, expensive infrastructure project. So

48:25

I wrote about it and I was pretty positive

48:27

about it. And then afterwards, like, it became clear

48:29

that this whole thing was essentially

48:31

a house of cards that no one, barely

48:34

anyone was actually using the network. The

48:36

company insiders and executives had like secretly

48:39

given themselves a big percentage of the

48:41

tokens. And I don't agree with

48:43

the criticism. Well, let me just let me just finish the

48:45

question. And you know, essentially,

48:47

this company had made itself look more

48:49

successful than it really was the token

48:52

price crash and the subreddits and places

48:54

where people who were big, you

48:56

know, proponents of helium started to

48:58

fill with just people complaining that they had lost

49:00

all this money or spent all this money on

49:02

routers that weren't giving them back anything in return.

49:05

People write to me all the time, even

49:07

today, who feel burned by helium and are

49:09

mad at me for writing about it. But

49:12

you actually write in your book that

49:14

helium is a good example of cryptos

49:16

usefulness. I'm curious, like, what in your

49:18

mind did helium prove? Yeah.

49:20

And so I think in the

49:22

book, I say helium solved half of the

49:24

problem. And let me explain that. So

49:27

this has been an idea that I've heard about

49:29

for 20 years, which is the idea of wouldn't

49:32

it be an interesting entrepreneurial idea if we could

49:34

create kind of a grassroots horizon? But the challenge

49:36

is the chicken and egg problem, which is the

49:38

challenge with a lot of tech businesses, right? It's

49:40

like, how do you you got to do two

49:42

sides of the market? And so

49:44

the insight that helium had is that

49:46

you can use financial incentives to help

49:48

with one

49:50

side of the market and the supply side of

49:52

the market. They did a good job on the

49:54

supply side, which is farther than any other attempt

49:57

got. They had not built out the

49:59

demand side. building it out now and

50:01

I believe they're making progress. I'm not personally

50:03

deeply involved but I mean I just idea

50:05

that it was gonna happen overnight like it's

50:07

a really big ambitious idea. I

50:10

understand that the supply and the demand sides of

50:12

that market but I think another way to describe

50:15

what happened with helium is that

50:18

they used casino like economics to

50:21

build an infrastructure project that no

50:23

one wanted or that very few

50:25

people actually had real use cases

50:27

for. So I don't... Well

50:31

they started off with this... I forgot the name

50:33

of it. It was Laura something.

50:35

It was some weird... Laura

50:38

Wan of course. Which

50:40

is a mistake and they switched it to 5G.

50:42

So like that was the first thing. And now

50:44

they have... you can go get it. You can

50:46

go... there's a service where you can go

50:49

sign up and I think it backhalls... it

50:51

backfills with T-Mobile or something.

50:53

It uses helium wherever it can and then

50:55

it backfills with... if it doesn't have coverage

50:58

in that area it backfills with T-Mobile and

51:01

it... you know... and they're building

51:03

that out and so and that just simply

51:05

takes time. You know

51:07

I guess I don't understand the criticism

51:09

that... Look I mean one of the

51:11

points I make in the book is

51:13

that subsidization is a very very old

51:16

tactic among internet companies.

51:18

Why do all these companies have to raise many

51:21

many many billions in venture capital because they subbed...

51:23

a lot of their proposal pitches they subsidize. I

51:26

see helium as okay now the

51:28

forces of openness have financial tools.

51:30

But someone did subsidize the building of

51:32

the helium network. It was the people who

51:35

thought that they could get rich on helium

51:37

tokens and bought a bunch of these useless

51:39

routers and like hooked them up to

51:41

the internet. Like it... we

51:44

can move on from helium but I think that

51:46

is a very clear example of how something that

51:48

looks to people in crypto like

51:51

a successful example of bootstrapping some project

51:54

that wouldn't have gotten funding otherwise or

51:56

would need venture capital otherwise. If you

51:59

drill down... down one layer, it actually

52:01

is just spreading the risk and the

52:04

subsidies around to a bunch of normal people

52:06

who just got snookered into participating. Wait a

52:08

second. I think for a couple things I'll

52:10

say. First of all, we have not ever

52:13

been involved with a project that sold tokens

52:15

to the public, just to be clear.

52:18

I don't know what these people are referring to, but

52:20

Helium never, I guarantee you, because no company we're involved

52:23

with, nor would we ever, would we allow them to

52:25

do this, sells tokens to

52:27

the public. No, you would earn

52:29

tokens by plugging in your router. Everyone else

52:31

earns them through airdrops. Okay. So

52:33

that's just to be clear. So like this idea

52:36

that they're like somehow selling these things to retail

52:38

investors, that's simply not true. I mean, I'm on

52:40

these, I can show you the posts on these

52:42

subreddits that you want, but there are people who

52:45

spent thousands of dollars because someone told them that

52:47

you plug in this router and it makes you

52:49

crypto and you get rich

52:51

off this. And

52:54

I mean, I don't think the company

52:56

said that. No, no, no, but

52:59

in an environment where you have a token that's sort of...

53:02

Anyway, we can move on, but I can't

53:04

say I want to let you go. So

53:06

let's like shift away from Helium, but I

53:08

do think that there is an idea here

53:11

worth exploring, which is what happens when you

53:13

financialize a network. As

53:15

I've observed some of these networks rise and

53:17

fall, it seems to me that when you

53:19

at least plant the seed with people that

53:21

they can make good, maybe not get rich,

53:23

but just get paid

53:26

by doing something, you wind up

53:28

attracting a different kind of person than if

53:30

you're just like YouTube and you just say,

53:32

hey, show up here and upload your videos.

53:34

I mean, the thing that I'm thinking of

53:36

is Axie Infinity, which is a company, an

53:38

Andreessen Horowitz company that I wrote about. If

53:40

you haven't heard of it, it was a

53:42

game reminiscent of Pokemon. Let players earn cryptocurrency

53:45

from battling their little monsters, which are called

53:47

Axies. Those Axies are also NFTs.

53:49

You had to buy one to play the game

53:51

or you could rent them from their

53:53

owners. When I wrote about it in 2021, it

53:55

was taking off in the Philippines. People were earning

53:57

more than the minimum wage by playing it. interesting

54:00

and I wrote about it, but the financial

54:02

nature of it essentially required that new

54:04

investors would come in to prop up

54:06

the price just like you would find

54:09

in a traditional Ponzi scheme. And

54:11

then eventually $625 million worth

54:13

of tokens were stolen, the game's whole

54:15

economy collapsed. So like, I'm trying

54:17

to understand, well, you know, obviously

54:19

we would wish that the hack hadn't

54:22

happened and, you know, maybe we learned

54:24

some lessons about the network. But

54:26

like, as far as I can understand it,

54:28

like, Axi is your dream of a

54:30

blockchain-based future where people have skin in the

54:33

game, they have a piece of the action.

54:36

And it just seems like it is

54:39

not doing good things. No,

54:41

so, okay. So like this, the way I view

54:43

it is, and I think the

54:46

Axi team would say this, the focus is way

54:48

too much on the financial aspects. One

54:50

of the lessons from things like Axi is that the

54:52

primary goal needs to be a fun game outside

54:55

of the financial aspects, right? So you need something,

54:57

let's just say like Roblox or Eve Online. And

55:00

I think there needs to be policy approaches.

55:02

By the way, like I say this in

55:04

the book, and we'd say this publicly, like,

55:06

I think, for example, having much, much longer

55:08

lockups on all tokens would be a

55:10

significant improvement because I think you would just remove

55:12

a whole class of kind of mercenary behavior. Like

55:15

I don't think we should be against people having a

55:17

profit motive. I think the real harm

55:20

is in the short-term profit motive. And

55:22

I think some of that probably has to be

55:24

achieved through policy. Some of it has to be

55:26

achieved through better products. But to me, it's throwing

55:29

the baby out with the bathwater to say that

55:31

because some early attempts over financialized things, there should

55:33

never be digital services where you can have peer-to-peer

55:35

economies where the users can earn financial

55:38

rewards. I don't think it means we

55:40

have to kill blockchains. I do think it means we

55:42

have to be thoughtful about how

55:44

we incorporate those legitimate criticisms into

55:47

better designs. Well, I was

55:49

thinking, you know, my view is that

55:51

this stuff should be banned forever. It's

55:53

not even that people should stop noodling

55:55

on this thing. That is

55:57

fine to me. I just feel like until

55:59

some... one makes one cool thing that is

56:02

not just like remittances

56:04

with a blockchain, I'm just kind of inclined to

56:06

ignore it. Because when I look at the amount

56:08

of talent, capital and time that has gone into

56:11

it, there's basically never been a technology in the

56:13

history of Silicon Valley, as far as I can

56:15

tell, where more time, money and talent went into

56:17

it and produced as little as we have so

56:19

far gotten out of the blockchain. Um,

56:23

I don't know. I mean, like, I think, like when I

56:25

started my career in the early 2000s, people

56:27

would have said that about, about, you

56:29

know, telecom equipment. And, you know, I don't know, I've,

56:31

I feel like I've gone through my whole career with

56:33

these ups and downs. And people say we over invested

56:36

in something. And, you know, cryptos, I

56:38

think, I don't know, I don't know the exact percentages,

56:40

but I don't think it's a huge percentage of total

56:42

venture investment these days. Yeah, we

56:45

talked recently on the show about the

56:47

Bitcoin ETF that have just been approved

56:50

for trading, basically letting people buy and

56:53

sell, you know, things that

56:55

attract the price of Bitcoin in their

56:57

retirement funds or whatever. And that's something

56:59

that a lot of people in the

57:01

Bitcoin world have been optimistic about. I

57:04

know you're not a Bitcoin laserized guy,

57:06

but as you look out across the

57:08

crypto landscape today, like, what are the

57:10

projects that make you optimistic? A

57:14

lot of my career has been focused

57:16

on creative people, you know, the most

57:19

disliked area I'm still the most excited

57:21

about, which is NFTs. I think NFTs

57:23

are a really powerful way for creative

57:26

people like to take musicians as an example.

57:28

Like, so, you know, music is an area

57:30

where users love music, users

57:33

love to pay for music. There's lots of

57:35

talented musicians. The problem today is these intermediaries

57:37

and it's labels, it's Spotify, it's a whole

57:39

bunch of different layers of intermediaries. So

57:42

we have like two companies, for example, doing music

57:44

related NFTs. And it's like you get an NFT and

57:46

it's a digital collectible and it's a backstage pass and

57:48

it's proof that you were the first one that liked

57:51

the musician and you can go to a special event.

57:53

And like, there's just a whole bunch of experiments happening.

57:56

But I think that this is a very powerful tool.

58:00

you know, blockchains and NFTs and these things

58:02

for thinking about new business models for creative

58:04

people. And I mean, I

58:06

guess I should say is like, I like I

58:09

should be rooting for that idea. And I guess

58:11

I am like, I mean, if what you just

58:13

described comes to pass, then hopefully that's something I

58:15

would be able to take advantage of, right? Like

58:17

I'm doing a lot of the things that you

58:19

say, I do run my own business, I am

58:21

a solo creator, I have been able to forge

58:23

a direct relationship with an audience using email, like,

58:25

it's awesome. Like everything that you're describing is beyond

58:27

my wildest dreams. I have a

58:30

community on Discord that gives me ideas and pushes back

58:32

on me when I run, like, all that stuff is

58:34

amazing. I just keep coming back to the fact like,

58:36

I do it using today's internet. And

58:38

I've even been able to do it

58:41

with a pretty low take rate. So

58:43

I'll just be curious to see like,

58:45

what winds up being on the blockchain

58:47

that enables, you know, me to accelerate

58:49

that I'm still open to the idea.

58:51

Yeah, I think making

58:53

money would be one and that's NFTs

58:55

and other kinds of new interesting experiments.

58:58

I assume Casey, you're still dependent on

59:01

these big social networks for distribution and things

59:03

too. Like you can't write an email lets

59:05

you deliver the thing, but how do you

59:07

acquire new users, right? And like Discord at

59:09

some point, like maybe they deprioritize links the

59:11

way the Twitter is. And maybe I don't

59:13

know. And I think in that world, do

59:15

you want people entrepreneurs building alternative sets of

59:17

tools that might

59:20

provide an outlet? That's, that's my argument,

59:22

right? Like, yeah, Chris,

59:25

before we let you go, I think

59:27

there will be people who listen to

59:29

you talking about this stuff and think

59:32

that makes a lot of sense to

59:34

me. I think there will be other

59:36

people who say, this is just someone

59:38

who has invested billions of dollars into

59:40

crypto companies, who is sort of talking

59:42

his own book and going on this

59:44

press tour to essentially do marketing for

59:46

the blockchain and web three investments of

59:48

Andreessen Horowitz crypto. I'm

59:50

curious, like if you didn't have money

59:53

on the line, if you didn't have

59:55

skin in the game, both your own

59:57

personal capital and your investors might. Would

1:00:00

you still be this optimistic? I don't

1:00:06

know. I mean, I've spent been

1:00:08

optimistic kind of my whole career. I mean, I

1:00:10

started in the internet downturn, you know, and so

1:00:12

I started my first company 2003-04. And that,

1:00:14

you know, that

1:00:17

was not a popular time to be

1:00:20

doing internet startups. It really wasn't. It

1:00:22

was quite hard to raise money. And

1:00:25

so, look, I've just seen my whole career every time

1:00:27

there's been a bunch of smart people. It doesn't work

1:00:30

for a while. And then eventually it does. And I

1:00:32

have yet to see the really smart people eventually not

1:00:34

get it to work. Okay. So I

1:00:36

don't know, Kevin, like, also, I would say, like, if you

1:00:38

read this book, I think a lot of people are surprised

1:00:40

when they read this book. It's not like seven cool things

1:00:42

written by fund manager kind of book, like I try to

1:00:44

be, you know, give pros and cons. And as

1:00:46

I said, it was sort of a test for

1:00:48

myself, right? And so look, I

1:00:50

mean, look, I obviously I have an interest in this. I

1:00:52

mean, you could go you could argue the causality goes the

1:00:54

other way. Like, I also work in this business because I

1:00:56

am optimistic. I believe in it and the firm is you

1:00:58

know, so you could argue either way. But yes, we have

1:01:01

skin in the game. I also

1:01:03

think we have a bigger skin in the game, which

1:01:05

is I do think that like, if we don't have

1:01:07

blockchains, like, I think it really, really will hurt tech

1:01:09

startup activity. So there's this, like, you know,

1:01:12

the whole venture industry is sort of based

1:01:14

on this cyclical new companies, you know, the

1:01:16

daddy line dies and the baby lines born.

1:01:19

And like, if we just have a bunch of

1:01:21

like daddy lines forever, like that's not good for

1:01:23

venture capital. So yes, I'm self interested. But I

1:01:25

would argue the causality goes the other way. I'm

1:01:27

in the business because I do believe in tech.

1:01:29

I've also like, I grew up on Unix,

1:01:31

I grew up on the web. I am

1:01:33

a like open source maximalist, I would like to

1:01:35

see everything open source and I see blockchains as

1:01:38

an extension of that. But I would hope people

1:01:40

could read the book and they have a critique

1:01:42

critique the book and not sort of the you

1:01:44

know, the modus behind it. Yeah,

1:01:46

yeah. Well, we'll end

1:01:48

there. Kristiksen, thank you for coming

1:01:50

on. The book is Read, Write, Own, Building the

1:01:52

Next Era of the Internet and you've been a

1:01:55

good sport here while we grilled you. So thanks

1:01:57

for coming on. Thanks, Chris. Yeah, there was a

1:01:59

tough for questions but they were good

1:02:01

you guys were fair thank you. Alright

1:02:03

thanks for coming. Casey

1:02:10

it's time to pass

1:02:12

the hat we're playing

1:02:15

a game of hat GPT.

1:02:37

Our segment on the show where we take news

1:02:39

headlines from the week out of a hat and

1:02:42

generate some plausible sounding language about them until one

1:02:44

of us tells the other one to stop generating

1:02:47

and this week we actually have

1:02:49

in our possession a brand new

1:02:51

hat GPT hat. That's right we

1:02:53

had a wonderful listener who went

1:02:55

and sent this to us now

1:02:57

if you are not watching the

1:02:59

YouTube show this is a cloth

1:03:02

bucket hat that I think started

1:03:04

as a plain white bucket hat

1:03:06

but was tie dyed with blue and

1:03:08

yellow and says hat GPT in the

1:03:10

hardfork font is frankly a beautiful hat.

1:03:12

It's a beautiful hat it was sent

1:03:14

to us by listener Guan Yip who

1:03:16

said that he designed it using generative

1:03:19

AI so thank you Guan

1:03:21

and we will use this hat

1:03:23

today and hopefully more hats will keep pouring in and

1:03:25

we'll just develop a little bit of a hat collection.

1:03:27

Now if it was designed using generative AI does that

1:03:29

mean that there's potentially a hidden code in this? I

1:03:33

don't know check the brim. Yeah well

1:03:35

oh actually it does say the future

1:03:38

is already here on the inside of the hat

1:03:40

so there is a bit and if you flip

1:03:42

it around just say stop generating. Wow what a

1:03:45

good hat. Hats off to Guan Yip thank you.

1:03:47

Alright so here are our news stories for the

1:03:49

week put them in the hat. Alright

1:03:51

and you can mix them up and

1:03:55

let me pull out the first one. Yeah you pull the first one. Okay

1:03:58

let's see Google. Google

1:04:00

employees blast profoundly boring management

1:04:02

which lacks visionary leader among

1:04:05

layoffs. This is from

1:04:07

the New York Post and it

1:04:09

is a story about a post

1:04:12

on LinkedIn from a Google employee

1:04:14

named Diane Hirsch Thoreau and

1:04:16

she wrote on her LinkedIn page quote,

1:04:19

from the C-suite to the SVPs to

1:04:21

the VPs, they are all profoundly boring

1:04:23

and glassy eyed. The full post really

1:04:26

is worth reading in its entirety. She

1:04:28

goes off and basically says this company

1:04:30

lacks all vision. So this post is very

1:04:32

long and it basically outlined a lot of

1:04:35

the problems that this Google employee sees with

1:04:37

the kind of sclerotic

1:04:39

bureaucracy at Google and it's quite a

1:04:41

thing to write in LinkedIn about your

1:04:43

bosses and your bosses. Well it's very

1:04:45

good writing too. You know, I mean

1:04:48

this is a post that comes in response to Google's

1:04:51

recent layoffs and I think you

1:04:53

know the point that this employee is making is

1:04:55

it really seems like these layoffs are being done at

1:04:57

random. Like this employee did not feel like there is

1:04:59

a grand strategy that is being executed and so

1:05:01

now this person like so many other people at

1:05:04

Google are just like how secure

1:05:06

is my job here? You know they feel like

1:05:08

they don't know where things are going. So you

1:05:10

know look in an organization of tens of thousands

1:05:12

of employees are always going to have some folks

1:05:14

there that think that the company doesn't have a

1:05:16

vision or whatever. I don't work at Google. I

1:05:18

don't know how true I would find this if

1:05:20

I worked there. But what I will say is

1:05:22

that in the meantime Diane Thoreau has done the

1:05:24

impossible. She has made LinkedIn seem interesting. Thank you

1:05:26

Diane. This was an incredible

1:05:28

post and we would love to see more posts

1:05:30

like this. Yeah more sounding off on

1:05:32

your bosses on LinkedIn please. Go off

1:05:34

Queen. Stop generating. Okay my turn. Alright.

1:05:38

Pass the hat. Here's

1:05:40

the hat. YouTube and

1:05:43

Spotify won't launch Apple Vision Pro

1:05:45

apps joining Netflix. This story is

1:05:47

from Bloomberg. Basically

1:05:49

YouTube, Spotify and a few other

1:05:51

major developers like Netflix have said

1:05:53

that they are not going to

1:05:56

make versions of their apps for

1:05:58

the Apple Vision Pro. which is

1:06:00

of course what they're calling a spatial computing

1:06:02

headset that comes out from Apple in

1:06:05

just a few days. Yeah, that's right.

1:06:07

And this is a surprise because for

1:06:09

the past decade plus when Apple would

1:06:12

come out with a new device, whether

1:06:14

it would be the iPad,

1:06:16

the watch, developers were racing

1:06:18

to be on this thing because they thought

1:06:20

there's gonna be a big new market there

1:06:22

and if we can be the first movers,

1:06:24

we can guarantee that we will reap the

1:06:26

rewards. And now you have two of the

1:06:28

biggest companies that exist, you just sort

1:06:30

of take for as a given that there's gonna be

1:06:33

a Netflix app or a YouTube app and these companies

1:06:35

have said, you know what, we're gonna pass on this

1:06:37

one for right now. Yeah, it's really surprising in part

1:06:39

because part of what the spatial computing headsets

1:06:41

might be really good for is like watching

1:06:43

a movie or watching a video in kind

1:06:45

of like an IMAX like experience with your

1:06:48

headset. So I don't

1:06:50

know, do you think the signals that developers

1:06:52

don't think this product is going to be

1:06:54

a success or there's another theory that was

1:06:56

advanced by Ben Thompson this week that this

1:06:58

is actually kind of developers sort of getting

1:07:00

back at Apple for the kind

1:07:02

of drama around some of their app store

1:07:04

decisions over the past few years and saying,

1:07:07

well, if you guys are gonna insist on

1:07:09

taking, you know, a 30% cut

1:07:11

of our revenue, we're gonna fight back

1:07:14

by refusing to make these apps that

1:07:16

might convince people to buy your $3,500

1:07:18

face computer. Yeah,

1:07:20

look, developers don't build apps out of good

1:07:23

wealth. They build apps to make money and

1:07:25

Apple has instituted a lot of rules that

1:07:27

ensure that some percentage of the revenue generated

1:07:29

goes to Apple and not these developers. And

1:07:31

so if you are a

1:07:33

Google thinking about the YouTube app or your Netflix,

1:07:36

you might just say, you know what? First of

1:07:38

all, Apple's not even gonna sell that many of

1:07:40

these things. You know, it's like the estimates are

1:07:42

maybe it sells a half a million of them

1:07:44

in the first year. How much incremental revenue can

1:07:46

a YouTube or a Netflix generate from there being

1:07:48

a half a million of these new devices? So,

1:07:50

you know, in the end, if there wind up

1:07:53

being tons and tons of these vision pro headsets

1:07:55

everywhere, I'm sure that YouTube and Netflix will come

1:07:57

around. But in the meantime, I do think it's

1:07:59

interesting They are needling Apple a little bit and

1:08:01

saying, hmm, let's wait and see. Yeah. Now,

1:08:04

we can say you will be able to look at video in the

1:08:06

web browser if you buy one of these things. So you would still

1:08:08

access YouTube, you can still access Netflix, but you're not going to be

1:08:10

able to do it with the app store. Yeah. Right

1:08:13

now it's my turn. Yeah, your turn. All

1:08:15

right. New

1:08:19

Hampshire investigating fake Biden robocall meant to

1:08:21

discourage voters ahead of the primary. This

1:08:24

is from the AP. The New Hampshire

1:08:26

Attorney General's office Monday said it was

1:08:28

investigating reports of an apparent robocall that

1:08:30

used artificial intelligence to mimic President Joe

1:08:32

Biden's voice and discourage voters in the

1:08:34

state from coming to the polls during

1:08:36

Tuesday's primary election. One

1:08:38

of the quotes here was apparently fake. I

1:08:41

Biden said, quote, save your vote for the

1:08:43

November election as if we

1:08:45

are only allowed to vote in one of the

1:08:47

two. It's not true. Also, this happened

1:08:49

in New York. So there was a

1:08:51

generated audio of a Manhattan Democratic boss,

1:08:53

Keith Wright. This is according to Politico.

1:08:56

Keith Wright was apparently talking smack about

1:08:58

another assembly member in this audio, but

1:09:01

the audio itself, which was this 10

1:09:03

second clip, was apparently a fake.

1:09:06

And so here, Kevin, we've talked about the

1:09:09

threat of deep fakes to elections. And I

1:09:11

think we've talked a lot about images and

1:09:13

video. But here it is already showing up

1:09:15

in national and local elections and it is

1:09:17

audio. Yeah. Well, what's interesting

1:09:19

to me is like it doesn't sound like this deep

1:09:21

fake that was played for voters in New Hampshire of

1:09:24

Joe Biden telling them not to go to the polls.

1:09:27

Like no one actually was fooled that this

1:09:29

was actually Joe Biden calling that we've just

1:09:32

become so accustomed to like these sort of

1:09:34

pre-taped robocalls. But I confess like

1:09:36

this arrived faster than I thought it would. Yeah,

1:09:38

I would agree with that. What I would say

1:09:40

if you're wondering, you know, how do I protect

1:09:42

myself against robocalls in this election? I would just

1:09:44

suggest not answering your phone until January. Okay. My

1:09:47

question is like who is picking up the phone in the year

1:09:49

2020? If you don't pick up the phone, if you don't recognize

1:09:51

the number, it's nothing good for you. Okay.

1:09:54

Nobody who nobody has a number is not already in your phone has anything good to tell you

1:09:56

on the phone. It's true. Yeah.

1:09:59

Let them maybe send you an email. generating. HP

1:10:06

CEO evokes James Bond style hack

1:10:08

via ink cartridges. This is from

1:10:10

Ars Technica. Last Thursday,

1:10:12

HP CEO Enrique Lores addressed the

1:10:15

company's controversial practice of bricking printers

1:10:17

when users load them with third

1:10:19

party ink. He said,

1:10:22

we have seen that you can embed

1:10:24

viruses in the cartridges. This is apparently

1:10:27

supposed to justify why you have to

1:10:29

buy their ink cartridges instead of ordering

1:10:31

one from like a third party on

1:10:33

Amazon, or right, which would probably be

1:10:35

cheaper and just as good. You know, there

1:10:37

was a time Kevin, when you could put

1:10:39

like a lot of different cartridges into your

1:10:41

printers, but then the printer companies wised up

1:10:43

and they said, well, why don't we lock

1:10:45

them down and create essentially like these, you

1:10:48

know, Keurig machine style things where only Keurig

1:10:50

pods work in your Keurig coffee maker. Well,

1:10:52

same thing for the printer, but instead of

1:10:54

just saying, we're incredibly greedy and we'll stop

1:10:56

at nothing to take all of your money

1:10:58

for printer ink. They said, if you use

1:11:00

another cartridge, people could literally

1:11:02

die. It

1:11:04

could infect your computer with a virus. Now I'm

1:11:07

not a security expert, but

1:11:09

that doesn't seem very plausible to me. Do

1:11:12

you think there's any way that a virus

1:11:14

could actually infect you through your printer cartridge?

1:11:16

I let me just say there's never been

1:11:18

anything I've been less scared of than a

1:11:20

virus in my printer. Like I truly am

1:11:22

not scared. You have a printer? I actually

1:11:24

don't have a printer and I need one.

1:11:27

Wow. Yeah. Do you have a printer? I

1:11:29

have a laser printer, but it doesn't take

1:11:31

ink cartridges. It takes like toner cartridges. And

1:11:33

what's the difference between toner and ink? I

1:11:36

have no idea. Okay, great. I couldn't tell you. Now,

1:11:46

this one is an activity. Oh, boy. Kevin,

1:11:48

your colleague, Stuart Thompson, published a quiz. We

1:11:50

have to guess which faces in this quiz

1:11:52

were made by AI. And we

1:11:55

are invited to try it out and see how

1:11:57

we do. If you'd like to play along at

1:11:59

home, we're going to put a link to it

1:12:01

in our show notes. Okay, so it's a 10

1:12:03

question quiz. For each question, it shows you a

1:12:05

face and asks you if the face was made

1:12:07

by AI. So should we take it together? Let's

1:12:09

take it together. I'm going to share my screen.

1:12:11

Okay, number one, this is a man I'm going

1:12:13

to say fake. That one looks

1:12:15

like AI to me. His eyes are too close together. All right. And

1:12:18

that's a real person. I apologize

1:12:20

to whoever's eye distance. I accidentally

1:12:23

just insulted. Okay,

1:12:25

this one kind of looks like

1:12:27

a high school gym teacher. I'm going to say real. And

1:12:30

we're wrong again. That is an AI image. Okay. We're

1:12:33

a zero for two. Another. So

1:12:35

I did hear once that the way to tell

1:12:38

on an AI face is by looking at the

1:12:40

ear lobes. Because AI is not good at doing

1:12:42

ear lobes that look like real people to your

1:12:44

lobes. So how do these ear lobes look? Well,

1:12:46

let's see. I'm

1:12:48

going to say fake. They look a little too connected. Fake

1:12:52

lobes. Yes, we got it.

1:12:54

Okay, this one. What do you think

1:12:56

about this one? I think this one

1:12:58

is fake because before we started, I

1:13:00

said it was real and it was

1:13:03

fake. So

1:13:05

you cheated. I thought we were going to take the

1:13:07

test separately. Okay. Correct. That

1:13:09

is an AI image. Okay, let's try another one. Okay.

1:13:12

All right. Now we've got a young lady. She is

1:13:14

sort of blonde. Blue

1:13:18

eyes. I would

1:13:21

say... Can't

1:13:23

really see the ear lobes. I'm

1:13:25

going to say real. Real. No,

1:13:29

AI. Oh my God. This is so hard.

1:13:32

There's another woman. This

1:13:34

has a lot of bokeh in the background. What's

1:13:37

bokeh? That's sort of blurring in the background and

1:13:39

like... Look at you Ansel Adams. I

1:13:42

think this is real. There's

1:13:45

some... I don't know. The

1:13:47

eyebrow looks kind of weird to me. I'm going to

1:13:49

say AI. I'm going to say real. Okay. Oh, you were

1:13:51

right. All right. So

1:13:54

here's a person who I would say respectfully has a

1:13:56

head shaped like a trapezoid. In a way that makes

1:13:58

me feel like he's fake. I'm

1:14:01

gonna say fake too just because the hairline

1:14:03

is like like either This

1:14:05

is a fake image or this person's barber

1:14:07

owes them an apology Quiz

1:14:14

has turned into us just like insulting the real

1:14:16

faces and appearances of real people I want to

1:14:18

say this man while I did say you have

1:14:20

a head shaped like a trapezoid I think I

1:14:22

said it was in a hot way. It's like

1:14:24

a hot it's a hot shape for

1:14:27

your head and Look

1:14:30

that's a beautiful man.

1:14:32

Okay. Okay. We're trying

1:14:35

another was a woman wearing glasses

1:14:37

and I mean

1:14:40

at this point I have completely lost any sense

1:14:42

of what is real and fake and I truly

1:14:44

am just flipping the coin Yeah, I I'd say

1:14:46

the real to me. I'm not really real. Oh

1:14:48

my god. It's AI. Oh Jesus,

1:14:52

okay Um,

1:14:54

there's a man. This is gotta be real. I Know

1:14:58

I like don't trust my

1:15:00

own. I don't need anymore. So I'm gonna say

1:15:02

fake even though I think it looks realistic I'm

1:15:04

saying real. All right Okay

1:15:10

AI Nope

1:15:15

that was real person wonderful. Okay, so we're

1:15:17

four out of ten forty percent I'm not

1:15:19

great at math, but that's not a passing

1:15:22

grade. No, we have failed out of AI

1:15:24

facial recognition school So I've taken a few

1:15:26

of these kinds of like is this thing

1:15:28

AI generated or not tests and

1:15:31

quizzes over the years and This

1:15:33

is definitely the hardest one you know look if

1:15:35

you just want to do some like shenanigans in

1:15:37

the world and like and that those shenanigans are

1:15:39

gonna start with you creating like a fake LinkedIn

1:15:41

profile like you're in heaven because That's

1:15:44

one of these images would get you anywhere.

1:15:47

You needed to go. So I think this is like just

1:15:49

a sign that we just cannot we

1:15:52

just cannot reliably tell like if you and I Students

1:15:55

of AI and cannot score

1:15:57

higher than a four out of ten on this quiz

1:16:00

No one can. We're host. All

1:16:02

right. Stop generating. All right.

1:16:05

That's it for this week on Hat GPT. We'll close up the

1:16:07

old hat, put it back on to pedestal where it

1:16:09

belongs. Heart

1:16:21

Fork is produced by Rachel Cohn and Davis Land. We

1:16:23

had help this week from Kate LaPreste. We're

1:16:26

edited by Jen Fount. This episode

1:16:28

was fact checked by Caitlin Love. Today's

1:16:30

show was engineered by Corey Schripple. Original

1:16:33

music by Alicia Baiteub, Marion

1:16:35

Lozano, and Dan Powell. Our

1:16:38

audience editor is Nell Gologli. Video

1:16:40

production by Ryan Manning and Dylan Bergison.

1:16:43

If you haven't already, you can check

1:16:45

us out on YouTube at youtube.com/Heart Fork.

1:16:48

Special thanks to Paula Schumann, Pui Wing

1:16:51

Tam, Kate LaPreste, and Jeffrey Miranda. You

1:16:53

can email us at heartfork at nytimes.com.

1:17:11

Thank you.