0:00

different but you were sort of like

0:02

you've kind of been turned into some

0:04

sort of censorship boogie woman. I

0:06

never heard boogie, boogie man turned into

0:08

boogie woman before but it makes it

0:10

sound like someone pretty cool actually.

0:43

Welcome back to Posting

0:45

Through It. I'm Jared Holt.

0:48

And I'm Mike Hayden. Thanks

0:50

for joining us another week.

0:52

I really enjoyed our guest

0:54

today, and I learned a

0:56

lot about social media. You

0:58

know, one person who is on

1:00

social media is a fellow named

1:03

Tim Poole, and I just

1:05

wanted to just let our

1:07

listeners know that we are hard

1:09

at work on a who is

1:12

Tim Poole. podcast episode, which I

1:14

believe will be out next week

1:16

with a very special guest who

1:19

happens to be a New York

1:21

Mets fan, like myself. And so

1:23

everybody is going to enjoy eating

1:26

their vegetables with today's podcast

1:28

and learn a lot. And then

1:30

we are going to consume a

1:32

large stuffed crust pizza next week

1:35

on Timpool. Yeah, we are hard

1:37

at work. you know, down in

1:39

the shafts, if you can conjure

1:41

the image in your mind, listeners

1:44

close your eyes unless you're driving,

1:46

but imagine, you know, the

1:48

mechanic crawling out of the pit

1:50

when you get your oil changed,

1:52

just covered in grease and oil.

1:54

That is us, hard at work,

1:56

preparing the Timpool episode. Imagine

1:58

if you will. a man

2:00

who wears his winter cap in

2:02

the middle of August in the

2:04

hot West Virginia sun? A

2:06

man who I once watched

2:08

a video of this like

2:11

weird alt-right delist influencer

2:13

pull his hat off his

2:15

head and I legitimately thought

2:17

Tim was gonna like punch him

2:19

in the face. I thought. I

2:21

believe that's millennial Matt. Yeah. You

2:24

will be learning about all that

2:26

and more. next week, but

2:28

for now a very, very

2:30

in-depth discussion about social media

2:32

and how it impacts the

2:35

brain and our culture. Joining

2:59

posting through it now, we have

3:01

Renee duressa. She's a associate research

3:04

professor at the McCourt School of

3:06

Public Policy at Georgetown. Before that,

3:09

she's technical research manager at the

3:11

Stanford Internet Observatory. She is the

3:13

author of a book I've really

3:16

been enjoying over the last couple

3:18

weeks called Invisible rulers.

3:20

The people who turn lies into reality

3:22

and I'm really excited to have her

3:24

on the show this week. Renee, how's

3:27

it going? It's good. Thanks for

3:29

having me on. First off, how

3:31

are you holding up? I mean, the

3:33

world is kind of, you know, 10

3:35

kinds of fucked up and crazy right

3:37

now. How are you doing? It is so

3:39

busy. I mean, between my job,

3:42

which keeps me busy, and then

3:44

now there's a whole pile of

3:46

congressional hearings that are starting again,

3:49

that are keeping me busy, and

3:51

then I've got three kids, which

3:53

also keeps me busy, and travel,

3:55

which is busy. I feel like I'm

3:58

I feel like I'm getting things done I

4:00

think helps me cope with the

4:02

chaos. Yeah, you're like me. It's

4:04

like when the world is burning,

4:06

there is almost a weird sense

4:08

of maybe it's just because I'm

4:10

distracted, but like a mental piece.

4:13

Really? Agency. Like you can get

4:15

something done. Yeah. Which feels wrong

4:17

and backwards, but Mike, do you

4:19

feel that at all? Or I

4:21

don't know. I don't know. Like I've

4:24

been doing nothing but my book edits.

4:26

And so I kind of just feel

4:28

nothing. The book doesn't come out until

4:30

2026, so there's a feeling of just

4:33

being kind of mired in something that

4:35

no one can see yet, right? So

4:37

I don't know, I'm not feeling great

4:39

about it, but next month. That was

4:41

a hard process, like just how long

4:44

it takes between when you feel like

4:46

you're done and then turning in the

4:48

edits and then waiting like almost a

4:50

year until everybody else sees it.

4:53

I remember being surprised by just how

4:55

long that gap was. The gap is

4:57

over. Like I said, I've

4:59

really been enjoying it. It's

5:01

connected a lot of sort of

5:03

loose threads in my head.

5:06

But the synopsis of

5:08

Invisible rulers, it's just

5:10

this excruciatingly in-depth,

5:12

but somehow still

5:14

digestible, look at

5:16

the way that social media

5:18

is becoming sort of the

5:21

new invisible ruler of society.

5:23

And I think... Borrowing

5:25

the idea of invisible rulers

5:27

is a really interesting way

5:30

to approach it and maybe we should

5:32

start there like the idea of the

5:34

invisible ruler, the ghost in the

5:36

machine, whatever you want to call

5:38

it. Where does that come from?

5:40

And why did you pick that

5:42

to try to explain how social

5:44

media works? Yeah, it comes

5:46

from Edward Bernese. It comes from

5:49

the book propaganda from the late

5:51

1920s, 1929 or so. I remember

5:53

when I was, so my

5:56

background is in computer science

5:58

and political science. you know

6:00

I rather infamously as it kind of the

6:02

way it came out was sort of infamous

6:04

but I worked for CIA when I was

6:06

an undergrad you know and so I had

6:08

been in in my very kind of first

6:11

job was was in intelligence but not in

6:13

propaganda at all and in kind of technical

6:15

work but I'd always been interested in this

6:17

question of you know how people came to

6:19

hold opinions and I got very interested in

6:21

specifically how people came to form opinions about

6:23

vaccines when I was a new mom being

6:25

suddenly delused with the kind of stuff that

6:28

Facebook throws at you when you're a new

6:30

parent, which is, you know, all of a

6:32

sudden your entire feed changes, your entire experience

6:34

on social media changes the minute that you

6:36

have a kid. And you don't realize it

6:39

until you're there. The minute you post your

6:41

first baby picture, boom, like everything that is

6:43

recommended to you is different. And so I

6:45

got very interested in that question of

6:47

why, right? Why are we shown certain

6:50

things? How do we form opinions? How

6:52

do we form new friend groups? Because

6:54

I really did, right, around that time.

6:56

How do we, how does our identity

6:58

change? How does what is appealing to

7:00

us in our kind of new identity?

7:02

How does that shift happen? Around that

7:04

time that I was paying attention

7:06

to all of a sudden being

7:09

tossed into these, you know, getting

7:11

this constant recommended stream of content,

7:13

anti-vaccine As that was happening to me, the

7:16

conversation about social media was

7:18

also very much focused on ISIS.

7:20

These two things were happening concurrently.

7:22

So as I was studying anti-vaccine

7:24

networks on social media, other people

7:26

were studying ISIS networks on social

7:28

media, and I was really struck

7:30

by the similarities between these two

7:32

completely different groups, but just how

7:34

like the structure of social media

7:36

had fundamentally transformed how they could

7:38

communicate, who they could reach, the

7:40

ways in which networks were forming,

7:42

the ways in which messages were

7:44

being pushed out to people. Unexpected

7:46

things, like if you followed one, you

7:49

know, the recommender system would push you

7:51

more. Just the complete lack of awareness

7:53

from the recommender system of what it

7:55

was suggesting to you, just the absolute

7:57

kind of amorality of the entire thing.

8:00

And I started kind of thinking,

8:02

all right, what is new here?

8:04

And it seemed like the technology

8:06

was very, very new. So I

8:08

was focusing a lot on that.

8:10

But I thought, OK, we have

8:12

a tendency to overreact to new

8:14

technologies as being, OK, this time

8:16

is so different. This time is

8:18

so different. So I got very

8:20

interested in reading about propaganda literature

8:22

from the Olden days, literally from

8:24

a century back. And that was

8:26

what took me to Bernese and

8:28

Lippman and Jacques Alul and these

8:31

sort of foundational writers that I

8:33

had never read in college actually,

8:35

because I had not really gone

8:37

through that training. And I remembered

8:39

reading the book propaganda and getting

8:41

to the passage, you know, there

8:43

are invisible rulers who control the

8:45

destinies of millions. And he's talking

8:47

in this passage about the role

8:49

of appealing to people. in this

8:51

capacity as members of a group

8:53

identity. And it just kind of

8:55

like really clicked in my head.

8:57

It really stuck with me as

8:59

this, it resonated with what I

9:02

felt like I was seeing, this

9:04

idea that all of a sudden

9:06

we were engaging so foundationally as

9:08

members of groups and thinking, oh,

9:10

this is so interesting. Here is

9:12

somebody a century ago talking about

9:14

the power of being unaware of

9:16

where a message is coming from,

9:18

but feeling it resonates so deeply

9:20

with who we are as people

9:22

and that desire to belong to

9:24

a group that I just got

9:26

very interested in in that particular

9:28

body of literature and I really

9:30

started kind of going down this

9:33

rabbit hole. This was maybe eight

9:35

years ago now into trying to

9:37

understand, you know, What was new

9:39

and different, which was the technology,

9:41

but what was foundationally the same,

9:43

which was the psychology of belonging

9:45

and the way that propaganda worked,

9:47

the way that it functioned, and

9:49

why we wanted to pay attention

9:51

to these kinds of messages. So

9:53

the concept of the invisible ruler

9:55

that Bernese lays out and that

9:57

you explain in the book, is,

9:59

you know, there are visible

10:01

rulers. These are our

10:04

politicians, prominent members of

10:06

society, you know, governments,

10:08

major media outlets, and then there

10:10

are invisible rulers, which are the

10:12

kind of force, you know, it

10:14

could be like a force, like

10:17

advertisers. If advertisers pull your funding

10:19

on your cable network after a

10:21

segment, well, they're a kind of

10:24

an invisible ruler, sort of, you

10:26

know, but or the press person.

10:28

that generates like that gets all

10:31

the media to show up at

10:33

a conference or a ribbon cutting

10:35

like that that is a type

10:37

of invisible ruler the person leaking

10:39

information to an outlet is an

10:41

invisible ruler and I just thought

10:44

that was a really interesting

10:46

way to think about today you

10:48

know looking at the Trump

10:50

administration but really just the

10:52

last I mean like decade or

10:54

two of politics as social media

10:57

has become more important,

10:59

more integrated into

11:01

political strategy, it makes

11:04

me wonder like, is the

11:06

lives of TikTok account, and

11:08

then like a new kind

11:10

of invisible ruler, you know,

11:12

if outrage generated on

11:14

that account can shake up

11:17

the deliberations of the press

11:19

shop at the White House. What say

11:21

you? Well, that's the argument that

11:23

I make is that the

11:25

influencers are profoundly important, right?

11:27

That content creators and influencers

11:29

are the most effective propaganda

11:32

that we see today in a lot

11:34

of ways because people don't think

11:36

about them in that context. They

11:38

still think about political propaganda in

11:40

particular as being something that comes from

11:42

the top down the sort of old

11:45

Chomsky model of You know, the mass media

11:47

being in cahoots with the government and,

11:49

you know, taking its cues from advertisers

11:51

and taking its cues, what it can

11:54

and can't say from a, you know,

11:56

from some government leader. When what's happening

11:58

now, I would argue. is that the

12:00

information is increasingly, you know, public

12:02

opinion is increasingly shaped from the

12:05

bottom up. And you have these

12:07

accounts that are very, very good

12:09

at understanding how to tap into

12:11

psychology, how to tap into group

12:13

identity, how to tap into a

12:15

desire to belong, to participate, to

12:17

be a part of a movement.

12:19

That's what Libs of TikTok is

12:21

actually good at, right? They take

12:23

something, they turn it into a,

12:26

you know, they take a, we

12:28

use the term main character, but

12:30

they take sort of an avatar

12:32

of a, an extreme opinion held

12:34

by an alternative faction, right? Some

12:36

enemy, some Lib, they make that

12:38

an object of ridicule and that

12:40

act of participating in the ridicule

12:42

really binds these people together. Like

12:44

that's what the other faction is

12:46

getting out of it. It's really

12:49

solidifying that in group cohesion by

12:51

hating these other people, that sort

12:53

of outgroup animosity. That's what really

12:55

kind of brings this, this group

12:57

of people together. And what you

12:59

see is the realization that that

13:01

can then be used as not

13:03

just an online main character moment

13:05

where we're all getting together to,

13:07

you know, to hate this person

13:10

or yell at this person or

13:12

participate in this online trend. It

13:14

actually then is used, it's, it's,

13:16

it's kind of like transmuted into

13:18

a bigger story. Do you remember

13:20

the, the eating the pets thing

13:22

that happened during Oh, of course.

13:24

Yeah, of course. No one will

13:26

ever forget eating the pets. Of

13:28

course. You're with two people who

13:30

might as well have been eating

13:33

the pets. But

13:35

that, that's another example

13:37

of this, right? You have

13:39

an online rumor, you

13:42

have something that galvanizes participation.

13:44

People are having fun

13:46

with it, right? They're making

13:48

these memes or using

13:50

grok. They're coming together as

13:52

a community. It's again,

13:54

it's bringing that, that cohesion,

13:57

that movement building. But

13:59

then you see JD Vance

14:01

pick it up, right?

14:03

And all of a sudden

14:05

it's a bigger story.

14:07

Now it's not just about

14:09

this one stupid meme

14:11

or this one weird accusation or

14:14

this one person who regrettably is photographed, you know,

14:16

holding a goose or whatever, that that poor guy

14:18

who winds up in the photo, the actual human

14:20

being, who is who becomes kind of like the

14:22

target of this of this online mob, J.D. Vance

14:24

then turns it into like, well, this is really

14:26

a story about who should be in America, who

14:28

should be in that community, who that community is

14:31

for. And that's where you see that handoff, right?

14:33

rumor and community and you know socialization and mob

14:35

dynamics to the propaganda campaign around the bigger story

14:37

which is who is America for what is the

14:39

deep story about immigration that we're telling here and

14:41

that's why I think this handoff that happens there

14:43

it increasingly comes from the bottom and is picked

14:45

up by the political elite as opposed to moving

14:48

from the top down where some machine picks

14:50

it up and sells it. Why does it feel

14:52

that feel like reactionaries in

14:54

general or people who

14:56

may even have fascist

14:58

ambitions are able to take

15:01

advantage of this technology

15:03

so much better than

15:05

left liberal whomever, right?

15:07

Even centrist. I've kind

15:09

of wondered about that too. I

15:11

always believed that the

15:14

algorithm sort of kind of

15:16

pulls people to the right, that

15:18

it benefits moneyed interests. more than

15:20

anything else and maybe left right

15:22

isn't the right way to even

15:25

look at this I don't know

15:27

but it just seems like those

15:29

things that would be that that would

15:31

float to the top of a left

15:33

liberal agenda don't get the same

15:35

kind of juice as something lengthy

15:37

the they're eating the cats

15:39

they're eating the dogs there've

15:42

definitely been you know you

15:44

can definitely point to mob

15:46

dynamics where we've had the you

15:48

know, the weird main character of

15:50

the day situations that are not

15:52

explicitly right coated, right? You remember

15:54

the, I love eating, I love drinking

15:56

coffee in the morning with my husband

15:59

and the weird... Oh yeah, the insane

16:01

mob that's brought up on that one.

16:03

The replies would be like, this is

16:05

insensitive to people who can't drink coffee

16:07

in the morning with their husbands. Don't

16:10

have gardens, don't have gardens. There have

16:12

been, or Covington Catholic was another one

16:14

that came to mind that I wrote

16:16

about in the book. The example where

16:18

there was like the kid in the

16:21

magga hat and you know that sort

16:23

of situation where people really prejudged that

16:25

situation in very unfortunate ways before the

16:27

actual facts emerged. So you definitely do

16:30

see it. What is curated? That's been

16:32

one of these things where there have

16:34

been some papers. There was one that

16:36

Twitter itself was a, members of, researchers

16:38

at Twitter were authors on the paper,

16:41

I'm blanking on the name right now,

16:43

but that did make this argument that

16:45

actually right leaning content does tend to

16:47

perform better. It was sort of put

16:49

out almost in opposition to the argument

16:52

that right leaning content is censored or

16:54

downrank or deprecated on Twitter. But as

16:56

far as the, you know, one of

16:58

the things that happens I think is

17:01

also just this investment, right? Who is

17:03

really making these, you know, that handoff

17:05

piece, that question of like, when are

17:07

the political elites really grabbing this stuff

17:09

and pulling it up? We saw Donald

17:12

Trump doing this in 2016. I mean,

17:14

you guys remember like Maga 3X, right?

17:16

Or when Donald Trump was running this

17:18

very insurgent style campaign, Hillary Clinton wasn't

17:20

doing that. Lord Clinton wasn't on there

17:23

saying like, what are my meme lords

17:25

making and how can I boost it?

17:27

No, it was, you know, it was

17:29

Pokemon, go to the polls. Yeah, I

17:32

am, chilling and see your rapids. So

17:34

it was like that investment, I think,

17:36

wasn't really there. There wasn't that that

17:38

same appreciation for it. Then there was,

17:40

you know, you know, Yoki Bankler and

17:43

the, the books that looked at the

17:45

sort of investment in the right-wing media

17:47

ecosystem. even that was not influencer driven,

17:49

but just the networked propaganda book that

17:51

looked at the link, you know, the

17:54

sort of link driven. ecosystem of just

17:56

the alt-media apparatus that emerged both on

17:58

Facebook and in the you know in

18:00

the broader web they really invested much

18:03

more heavily there I think also because

18:05

the sense was that mainstream media was

18:07

owned by the Libs so why would

18:09

you invest in this ecosystem when you

18:11

already had this apparatus over here that

18:14

honestly that was the sense I got

18:16

looking at the anti-vaxers too right like

18:18

That's why they invested because they didn't

18:20

have access to that other ecosystem. So

18:22

they went and built their own using

18:25

the tools that were available to them,

18:27

even as the CDC and the institutionalists

18:29

were like, whatever, we don't need that

18:31

because we have this other thing. So

18:34

part of it was investment, part of

18:36

it was like a conscious decision to

18:38

do this, right, to say like, we

18:40

can make our people feel. empowered by

18:42

boosting them. Jared, you must have seen

18:45

this also, right? There would be these

18:47

people who would, their entire Twitter bio

18:49

was like, retweeted by Donald Trump four

18:51

times, Donald Trump Jr. six times, like

18:53

Dan Bonjino four times, like, it was

18:56

such a point of pride. It's like

18:58

a trophy, because, yeah, exactly, like, there

19:00

was, like, that sense that, like, the

19:02

elites would boost you. I never, I

19:04

don't think I've ever seen anybody on

19:07

the left do that. Maybe because I

19:09

don't know if that's just not a

19:11

cultural thing or if it's not happening.

19:13

It is to be very cringe. It

19:16

is extremely, I mean it is. I

19:18

think there's credit to the left liberal

19:20

side of social media. It's a pretty

19:22

cringe thing to do. But at the

19:24

same time, you know, you do get

19:27

people who you'll see them say like,

19:29

oh my god AOC followed me, right,

19:31

or you know, so and so followed

19:33

me like there is still that sense

19:35

of like. My hero who is good

19:38

at social media is paying attention to

19:40

me. I had that good tweet and

19:42

somebody liked it. People want to be

19:44

recognized as being part of that community

19:47

as having input and insight and I

19:49

think that it's just it's treated very

19:51

very differently and so starting in 2016

19:53

you really saw that culture begin to

19:55

emerge and it was you know they

19:58

really doubled down in 2020 and. And

20:00

it's kind of continued

20:02

ever since. So I would be

20:04

curious your thoughts on why this

20:06

has been so successful. In the

20:08

book you bring up a couple

20:11

things that I think are interesting

20:13

to get into around that question.

20:15

One is the 99.1 problem, which

20:17

I have to say slowly, so

20:20

it doesn't sound like I'm saying

20:22

like one giant messed up number.

20:24

Being that like on a lot

20:27

of social media platforms. Like

20:29

something like 90% of the

20:31

users fairly consistently across platforms Just

20:33

lurk they like don't post or

20:35

anything they just check it out

20:38

and leave Nine percent of

20:40

users will post like sometimes

20:42

which an occasional user quote-unquote

20:45

of social media platforms tends

20:47

to be defined fairly liberally.

20:49

It's like did you post this

20:52

week? Okay, you're an occasional user

20:54

and then one percent of social

20:56

media users are just like Well,

20:59

me, just incessant, incessant,

21:01

several posts per day. We

21:03

will not stop, we will

21:06

not stop posting kind of

21:08

thing. Then there's also the

21:10

algorithmic sorting of the

21:12

recommendation systems. You know,

21:14

the algorithm has become

21:16

kind of like this like sloppy

21:18

kind of bookie man in the

21:21

wrong hands when people talk about

21:23

it. But. The fact of the

21:25

matter is these platforms recommend you

21:27

content and it seems like in

21:29

recent years are doing it more

21:31

and more. I can't get on

21:33

Instagram without like every third post

21:35

being an advertisement or some kind

21:38

of recommended post that I didn't

21:40

ask for. And the systems that

21:42

recommend that content shape not only

21:44

what we see, but also the kind

21:46

of content that gets created because

21:49

the... influencers want to appeal

21:51

to them and like make a

21:53

living or get notoriety or whatever

21:55

goal they might have. So all of

21:57

that's to say what we see online

22:00

is like so far removed oftentimes

22:02

from reality. We interact with

22:04

people in ways we would

22:06

never interact with them in

22:08

person. We see content that

22:10

is just like borderline unnatural.

22:12

You use an example of like guitar

22:14

guy getting erratic, you know, audience

22:17

captured and stuff, which spoke very

22:19

directly to me as somebody with

22:21

like a pedal board full of

22:23

like. gizmos and funny boxes

22:26

that make silly sounds. But

22:28

it's just like the content

22:30

almost, it's like it is

22:33

an extension of reality, but

22:35

it is also just so

22:37

like deeply unnatural too. But

22:39

yet it has this vice

22:41

grip on society. I'm curious

22:43

like how you square that in

22:45

your head. What you're describing now and

22:48

the things that you're seeing in your

22:50

feed that you didn't ask for that's

22:52

called Unconnected content and Facebook actually discloses

22:54

what percentage of Unconnected content it shows

22:57

you you can see that in their

22:59

transparency reports quarterly Which is actually kind

23:01

of nice. I mean most most platforms

23:03

won't tell you that And it's risen

23:06

from about 8% to last time I

23:08

did a report on this was like

23:10

March of 2024 so it was around

23:12

2425% by that point Right. And the

23:14

reason for it is to compete with

23:17

Tiktok. That's that is why they're doing

23:19

it, right? Because Tiktok, their big innovation

23:21

was the realization that people just wanted

23:24

to be entertained. And it actually didn't

23:26

matter if they had a social graph

23:28

that was attached to those, you know,

23:30

to the people whose content they were

23:32

seeing. Whereas Facebook had started in the

23:34

quaint early days of social platforms where

23:36

you followed your friends and you looked

23:38

at like party photos, right? That was

23:40

what we did with it. Same thing

23:42

with Instagram to some extent. As they

23:44

realize that you could build an entire,

23:46

like what they called interest graphs originally,

23:49

and then as they realize that interests

23:51

were actually in some ways quite transient,

23:53

that it could be entertainment and memes,

23:55

which didn't have to be tied to

23:57

some deep interest you had, but could

23:59

just be like... the thing that you

24:01

were obsessed with that week. I

24:03

remember, like, I started, I got

24:05

recommended the Red Bowl danceier, like dance

24:07

your style, like competition for 2024,

24:09

which actually came out like four

24:11

or five months ago, I think, so

24:14

these are not even new videos,

24:16

but I watched one and I

24:18

was like, hey, this is actually really

24:20

good. These dancers are like amazing.

24:22

I used to dance back in

24:24

the day and so I do watch

24:26

dance content. And I started watching

24:28

more of them and like actually

24:30

then going and looking up some of

24:33

the dancers I liked, and this

24:35

was maybe like two weeks ago,

24:37

my entire YouTube feed is nothing but

24:39

dance videos now. That is it.

24:41

Like that is all I'm getting,

24:43

you know. And I think for the

24:45

probably for like the next couple

24:47

weeks until I hit on something

24:49

else or like go proactively looking for

24:52

something else, like that's what I'm

24:54

going to get. And it's just

24:56

going to be like a deluge. And

24:58

Instagram is like this too, like

25:00

this too, like this too, like,

25:02

like, like, like, like, Like, you do

25:04

one thing and then it thinks

25:06

like, okay, boom, this is what

25:08

they want and in order to keep

25:11

them on site, we're going to

25:13

give them just like an absolute

25:15

shit ton of it and like that's

25:17

what you're going to get. And

25:19

it's very hard to break out

25:21

of it. And I remember when I

25:24

was looking at, this unconnected content

25:26

project, we were looking at AI

25:28

Slop, which didn't even have a name

25:30

then. We were just doing this

25:32

project studying this project studying like

25:34

these garbage AI pages. And that was

25:36

all I was getting and it

25:38

actually rendered Facebook like completely unusable

25:40

for me because once you engaged with

25:43

one of the pages, literally that

25:45

was all you were all that

25:47

you were going to get because the

25:49

like the poll from the recommender

25:51

system like it people really were

25:53

curious about AI-generated content and it just

25:55

pushed it and pushed it and

25:57

pushed it. And one of the

25:59

things that used to be able to

26:02

see on crowd tangle was the

26:04

engagement with these pages after they

26:06

kind of. But what you would see

26:08

is the view counts on this

26:10

content. was like tens and hundreds

26:12

of millions of views on some of

26:14

the videos or some of the

26:16

images and then all of a

26:18

sudden they changed the recommender system and

26:21

you just see the content engagement

26:23

just flatline just like boom just

26:25

drops off a cliff and it's because

26:27

it's entirely determined by what the

26:29

platform decides to show people it's

26:31

not like organic searches where people are

26:33

going and looking for the page

26:35

so that power of suggestion and

26:38

that nudge is really incredibly powerful and

26:40

it's also completely So what you're

26:42

seeing and what your neighbor is

26:44

seeing are completely different, every now and

26:46

then there'll be some cultural moment

26:48

that everybody gets, right? Something that

26:50

is like very, very major, but otherwise

26:53

it is so random and so

26:55

determined just by the things that

26:57

you do that it makes it I

26:59

think challenging to feel like we're

27:01

having some kind of shared conversation

27:03

or shared, you know, like national... discourse

27:05

in a way? Well, we have,

27:07

we have, we have like we're

27:09

completely divided, right? Like that's, I mean,

27:12

it used to be people that

27:14

talk about, oh, we're, we're a

27:16

divided country. And now it's like, there's

27:18

no discussion about it anymore. We

27:20

just know that it's like we

27:22

just know that we're just completely divided.

27:24

And I was just going to

27:26

ask about that, that the way

27:28

those sort of algorithms kind of that

27:31

this kind of widespread sort of

27:33

fascism, fascist streak inside magga or

27:35

maggots help being a fascist movement, is

27:37

that the result of these algorithms?

27:39

I mean, because it didn't seem

27:41

like there was an agenda to turn

27:43

the country this far to the

27:45

right or this authoritarian, right? It

27:47

was, there was a pushback against Trump

27:50

from the Republican Party, the minute

27:52

he started running. And a lot

27:54

of people trying to push against it

27:56

and he became a social media

27:58

phenomenon. on and it seems like

28:00

now like a decade later I don't

28:02

want to oversimplify but it does feel

28:05

like the algorithm really has taken the

28:07

entire party or half the country on

28:09

this ride that they wouldn't have gone

28:11

on necessarily a decade ago. I think

28:13

it helps things become very normalized. It's

28:15

not only the algorithm that's the one

28:17

thing that I think it's like I

28:19

try to make the point even in

28:21

the book it's like the reason I

28:23

describe it as influencer algorithm crowd is

28:25

like you can't have one of those

28:27

things without the other. The algorithm serves

28:29

what it has available. So whether, you

28:31

know, it's showing me these dance videos

28:33

because they're there and I like it,

28:35

right? And it is that because I

28:37

like it piece that actually is, you

28:39

know, one of the things that's there, right?

28:41

There's like curiosity, there's interest, there's something

28:44

that keys off of it. This is

28:46

the same thing with, you know, I'll

28:48

use anti-vaccine content as an example. When

28:50

I was getting pushed it, I would

28:53

click. Now I was like kind of

28:55

hate clicking I was curiosity clicking but

28:57

the reason I kept getting more of

28:59

it was because I engaged and so

29:02

you know when you mentioned a couple

29:04

minutes ago the influencer is incentive or

29:06

the guitar guy example that I that

29:08

I created just to not point to

29:11

one particular person as the audience capture

29:13

avatar. The influencer is creating content for

29:15

the algorithm and the audience and they

29:17

have to always be doing it for

29:19

both. Otherwise, their content isn't going to

29:21

get promoted, which means that nobody's going

29:23

to see it, and they're not going

29:25

to get that initial burst of engagement,

29:27

like the likes and the shares or

29:29

views that are then going to push

29:31

it out to more people. So they

29:33

have to be doing it for the

29:35

algorithm and the crowd. And these things

29:37

always, like, you have to think about

29:39

it basically as like a triangle. And

29:41

so what that means with the kind

29:43

of normalization piece, the thing that's very

29:45

interesting about the influencer is that

29:47

They become that like engine

29:49

of normalization. And when

29:52

they realize that the audience

29:54

is receptive to something that

29:56

feedback loop between the influencer

29:59

and. the audience means that they're

30:01

essentially getting permission to say the more

30:03

extreme thing because the audience is like,

30:06

okay, cool, right? If the audience is

30:08

like, no, no, no, no, no, what

30:10

are you doing? Then they won't do

30:12

it. But if the audience is like,

30:15

yes, totally I'm with you, let's go.

30:17

Then they do. And that's where you

30:19

see that, like, you know, you don't,

30:22

you can't just blame the tech, right?

30:24

The tech recognizes that that that that

30:26

that that that that cycle is happening

30:28

is happening, that cycle is happening, But

30:31

there is that piece of, yeah, I'm

30:33

curious, like, yeah, I'm going along for

30:35

the ride with you. Yes, I'm actually

30:38

gonna kind of like, in some cases,

30:40

nudge you in a particular direction. There's

30:42

this, some of the research that like

30:44

Damon Santola does, he's a professor at,

30:47

I think you pen, if I'm not

30:49

mistaken, I hope I didn't get that

30:51

wrong, where what he talks about is

30:54

this thing called complex contagion, where the

30:56

influencer is almost acting as like the

30:58

gatekeeper in a sense where they're deciding,

31:00

you know, they're seeing their kind of

31:03

phantom talking about all this stuff, they

31:05

can decide whether to pick it up

31:07

and move it along, right, by boosting

31:10

that random thing that they're, you know,

31:12

that their fan is saying, you know,

31:14

like you can see this in the

31:16

election rumor mill, if you pick it

31:19

up and boost it and say, like,

31:21

big if true, you're still kind of

31:23

giving it your credibility, you're giving it,

31:26

you're pushing it out to more people,

31:28

you're pushing it out to more people,

31:30

is that that is a fully normal

31:32

reasonable, completely accepted position to hold in

31:35

that community now in a way where

31:37

even four years ago, they would still

31:39

try to couch it a little bit.

31:42

You know, hey, some people should, you

31:44

know, people should really look into this,

31:46

oh, I don't know, big if true,

31:48

somebody should go check that out. Now

31:51

they're like, look at this, they're doing

31:53

it again, you know, they're stealing it,

31:55

here they are stealing it, here they

31:58

are stealing it, gatekeeping it anymore. They're

32:00

just like moving it right along and

32:02

that's this this shift. from they're no

32:04

longer risking their credibility and kind of

32:07

gating it and serving as a limit

32:09

to it, they're essentially just like moving

32:11

it right on through. And I

32:14

think as that audience gets bigger,

32:16

it also has like a, it

32:18

disincentifies is the opposite, right? I

32:20

mean, Mitt Romney and Liz Cheney

32:22

are political has been at this

32:25

point. Why? Because they had the

32:27

nerve. to stick to their point

32:29

that the capital riot was not

32:31

ideal. And they were punished, right?

32:34

So it's like a punish mechanism

32:36

too. But I, you know, sort

32:38

of last question on the theory

32:40

of understanding part of this before

32:42

we shift gears a little bit

32:45

is, you know, I want to

32:47

get your thoughts on just how

32:49

impactful is this stuff really? Because

32:51

sometimes, you know, like in my

32:54

day job, it can be. I

32:56

sometimes, a post can get like

32:58

10,000 retweets or something like this,

33:00

right? It's like somebody who manages

33:02

an organization who spends like an

33:05

hour a week on the internet

33:07

because they have like a real

33:09

job doing a real thing. But

33:11

you know, sometimes a post can

33:14

get like 10,000 retweets or something.

33:16

But like what's it mean? Nothing,

33:18

right? Or sometimes supposed to get

33:20

like a hundred retweets and then

33:22

holy shit. There's like a hate

33:25

group that just showed up at

33:27

your book launch event or something.

33:29

You know, so how powerful is

33:31

social media like really is there,

33:34

you know, any sort of consistent

33:36

identifiers in terms of like once

33:38

this happens or like if it

33:40

does this or you know, it's

33:42

Like all media it's I've never

33:45

been under the illusion that we

33:47

would like record a podcast or

33:49

I would write a news article

33:51

or a paper and it would

33:54

change the world forever. It's like,

33:56

it's just material and people use

33:58

it and pick it up for

34:00

whatever they might want to do

34:02

or whatever they might want to

34:05

take away from it. But is

34:07

social media in the way it's

34:09

integrated in our lives, like inherently

34:11

impactful in a way that other

34:14

media maybe isn't or are there?

34:16

things about it or like ways

34:18

of interpreting it that can make

34:20

it more or less impactful in

34:22

any given circumstance? The thing that,

34:25

so first of all, I agree

34:27

that oftentimes it's overstated where someone's

34:29

like, look at this one post,

34:31

it got 10,000 likes, oh my

34:34

God, that's such a big deal,

34:36

right? And we've known for a

34:38

long time now, you know, dating

34:40

back to media theory studies and

34:42

from the 1940s that, you know,

34:45

what they called the hypodermic, the

34:47

hypodermic model of media changing people's

34:49

minds, you know, the idea that

34:51

if you saw something on the

34:54

nightly news, your opinion would magically

34:56

be changed. Like, no, that's not

34:58

how it works. But one of

35:00

the things that's interesting about it,

35:02

is that what studies did find,

35:05

even back then, was that it

35:07

was, you know, your opinion would

35:09

be shaped based on the conversations

35:11

that you had with people, right.

35:14

there'd be members of the community

35:16

that were called opinion leaders who

35:18

paid a lot of attention to

35:20

media, and that you would talk

35:22

to those people who paid a

35:25

lot of attention to media, you

35:27

would kind of decide amongst yourselves

35:29

what the facts of the matter

35:31

were. And that was sort of

35:34

how you would over time shift

35:36

your opinion and come to believe

35:38

something or come to like a

35:40

political candidate or whatever it was.

35:42

And what I think is interesting

35:45

about social media and particularly influencers,

35:47

and why I spend so much

35:49

time paying attention to them. is

35:51

that those two things are now

35:54

like merged into one thing, right?

35:56

So they're both media in that

35:58

they have that like. elevated

36:00

reach, that mass reach, and they're kind

36:03

of just like you and they're talking

36:05

to you. And the community of people

36:07

who kind of grow up around them,

36:09

you know, the sort of friends that

36:11

you make online, are the people that

36:13

you're talking to. So you no longer

36:15

have that maybe more diverse set of

36:17

people that you meet in the real world.

36:20

You're no longer geographically bounded with all the

36:22

different constraints that come with that. Instead, you

36:24

have this very kind of homogenous group that

36:26

you've been sort of shunted into as you've,

36:28

you know, your people, you may know, algorithm

36:30

is like, you're just like these people. You

36:32

should all be just like each other together.

36:34

You know, oh look, you all, you know,

36:36

you all share weird anti-government beliefs. You should

36:38

join this Q-anon group over here, right? You

36:40

remember this from the old days of like

36:42

the Facebook, like the Facebook, like the Facebook,

36:44

like the Facebook, like, That's sort of dynamic

36:47

of, hey, we're going to put you

36:49

all together. You're all going to spend

36:51

a whole lot of time talking to

36:53

each other. You're going to follow influential

36:55

figures who are kind of just like

36:57

you. They're going to talk very directly

36:59

with you. They're not going to feel

37:01

like media. They're also going to have

37:03

massive reach and all of the sort

37:05

of trappings that go with what we

37:07

used to think of as media. So

37:09

they're this weird hybrid figure. And that's

37:11

where you're going to spend less and

37:13

less and less time. with friends in

37:16

the real world, this is going to

37:18

kind of replace your socialization

37:20

too. So I think that it

37:23

makes it an incredibly effective persuasion

37:25

engine just in terms of like

37:27

these different dynamics all coming together

37:30

in one place. That's why I

37:32

think it's actually a really

37:34

interesting thing to be paying attention

37:36

to. I think there's this dynamic

37:39

of like the rumor into propaganda

37:41

cycle, right? Like eating the

37:43

pets. was actually a thing. Is

37:45

everything a thing? No. But some

37:47

of them hit. You know, and

37:50

it's understanding how these random moments

37:52

feed into these bigger stories and

37:55

then also the ways in

37:57

which we as individuals actively

37:59

have. the power to participate in

38:01

shaping those things now, which makes

38:03

people feel, I think, much more

38:05

invested as opposed to when they

38:07

were kind of sitting on their

38:09

couch just watching it go by.

38:11

So that's what I think is

38:13

different. I have an unscientific theory,

38:15

and I'm curious if you would

38:17

weigh in just to tell me

38:19

whether it has any basis or

38:21

not. So I always thought like

38:23

something about influenza or social media

38:25

was more effective at radicalizing people

38:27

or at least getting people to

38:29

be more you know to be

38:31

more laser focused on one particular

38:33

thing or the other in part

38:35

because they you they people access

38:37

you influencers access you usually completely

38:39

alone and trapped with your phone

38:42

or your device or whatever whereas

38:44

you know the newspaper is something

38:46

that it's like you know it's

38:48

out in the open everybody knows

38:50

that you're reading it and what

38:52

you're reading when you open a

38:54

newspaper and when you like watch

38:56

Better Call Saul with your husband

38:58

or whatever, there's a kind of

39:00

like sit on the couch and

39:02

watch TV together. Go to the

39:04

movie theater and watch the movie

39:06

together. Go to the concert and

39:08

you know, and watch pavement play

39:10

or something like that. I don't

39:12

know. Yeah. So, so these things

39:14

that I just mentioned, but but

39:16

when you're, you're kind of like

39:18

just online. You're kind of in

39:20

this, you know, this sort of

39:22

stereotype, cliche, whatever, of like you're

39:24

just in a dark room and

39:26

Alex Jones is ranting at you

39:28

about, you know, about, you know,

39:30

what Ukraine is secretly trying to

39:32

do. It has a different, it

39:34

has a different effect, right? And

39:36

then the social media, maybe I'm

39:38

wrong, but like that's just the

39:40

way I've always thought about it.

39:42

It's just people. you know, and

39:44

those kind of misogynistic places and

39:46

stuff like that, where you just

39:48

imagine. these guys in these really

39:50

dark rooms, like just their brains

39:52

cooking on propaganda, they're not socializing.

39:54

They are, they're just trapped with

39:56

their influencers. I think they are

39:58

actually, like one, one thing, people

40:00

will take that stuff and they'll

40:02

chuck it into the group chat,

40:04

right, or the discord server, or

40:06

the message board. It's actually like

40:08

really, or the, you know, subreddit

40:11

back in the day when that

40:13

stuff was on Reddit more. I

40:15

think that there's actually a fair

40:17

bit of I saw this thing

40:19

and now, and now I'm sharing

40:21

it. and I think that it

40:23

is actually very much like who

40:25

you share it with. I mean,

40:27

you look at the number of

40:29

random links that just land on

40:31

even the Chan boards, right? Like

40:33

people are, I see this thing,

40:35

I share with my friends, I

40:37

see this thing, I share with

40:39

my friends, like that's the, like

40:41

that's kind of the function in

40:43

the group chat, right? It's to,

40:45

it's to, like we're like jointly

40:47

processing this weird thing that we've

40:49

just seen. You can do it

40:51

in public on Twitter, right. and

40:53

you can do it with strangers

40:55

on those places, but you also

40:57

have these places where you have

40:59

that more persistent standing community. That's

41:01

I think the difference between the

41:03

group spaces, like the spaces that

41:05

are group oriented, whether that's Facebook

41:07

groups, which are a little more

41:09

obscured, or the ones where the

41:11

message boards where you can actually

41:13

kind of see it all go

41:15

by, versus the ones where you're

41:17

just talking in public to whatever

41:19

random person happens to be online

41:21

at the same time as you.

41:23

You do see, there was like

41:25

the, like the, I used to

41:27

say it is a joke, but

41:29

I think it's actually kind of

41:31

true, right? Like Twitter makes these

41:33

mobs that kind of come together

41:35

really spontaneously, but Facebook was where

41:37

you would have like the cults

41:40

that would really come out of

41:42

there, right? Like the people who

41:44

were like, we are here, man.

41:46

And like Alex Jones stuff would

41:48

get tossed in there all the

41:50

time. It was just this is

41:52

where we are here to get

41:54

tossed in there all the time.

41:56

the standing communities will be a

41:58

little bit more like that. So

42:01

the government has taken some

42:03

interest in this, our good

42:05

friends, the US government, avid listeners

42:07

to the podcast. I learned in

42:10

a very strange way that I

42:12

may retail on a later episode

42:15

of the show, but the Republican

42:17

Party particularly has

42:19

taken a pretty heavy interest in

42:21

social media. Like you said, in

42:24

the early days it was all

42:26

ISIS. It was like, how is

42:29

ISIS using Twitter, using Twitter, to

42:31

get towns to empty out before

42:33

anybody even sets a foot there,

42:36

right? Or like, how are they

42:38

using it to recruit? Which, like,

42:40

even if it was a small

42:42

number of people, that's a huge,

42:44

you know, it doesn't take very

42:47

many people to make a huge

42:49

impact on something, right? So it

42:51

started on that, and then,

42:53

like, some steps happened, and

42:55

now it's like Jim Jordan.

42:57

Really and Mark Zuckerberg about the

43:00

diamond and silk Facebook page and

43:02

why it's not getting as many

43:04

likes as it should or like

43:07

Donald Trump Jr. on Instagram being

43:09

like my post only got 10,000

43:11

faves. Obviously I'm being censored

43:14

here and like that became

43:16

that sort of partisan calling

43:18

card on the flip side,

43:20

especially after the election of

43:22

Trump and you know as

43:24

Democrats tried to hit their head

43:26

against the brick wall figuring out

43:28

what happened. You know, one of

43:30

the early theories was like, well,

43:33

it's the internet. It's Russian trolls

43:35

on the internet and bots on

43:37

the internet, which had a grain

43:39

of truth to it. It's not

43:41

the whole story, but like, even

43:43

a little bit, but in the

43:45

course of all of that, people,

43:48

including yourself very personally, have

43:50

been turned into, like, you

43:52

said it earlier, like artificial.

43:55

Boogie women, you know,

43:57

which which I agree Mike

43:59

sounds incredibly cool. But

44:01

like I don't imagine that experience

44:03

felt incredibly cool for you Renee.

44:05

The first time I ever heard

44:08

of you was in the context

44:10

of hearing how horrible a person

44:12

you are from a person from

44:14

a person who was genuinely horrible

44:16

to be clear. You know I

44:18

can't remember what what Nazi was

44:20

throwing that out. So the government

44:22

went from being like kind of

44:25

interested vaguely like supportive on quote,

44:27

unquote national security grounds of like

44:29

this sort of research into how

44:31

does the internet influence the world,

44:33

misinformation, disinformation, blah, blah, blah. Me

44:35

too. Me too. Me too. That's,

44:37

if it, I realize I speak

44:39

monotone, but like, there's contempt in

44:42

my voice. So they went from

44:44

being vaguely interested to then just

44:46

like going on the offensive against

44:48

researchers and institutions that look at

44:50

this stuff. You had a very

44:52

unfortunate front row seat to this

44:54

kind of through the whole life

44:56

cycle. What changed? Why did they

44:59

get so fucking mad at everybody

45:01

who tries to understand how the

45:03

internet works? Yeah, because some of

45:05

us said the 2020 election was

45:07

free and fair and that was

45:09

an inconvenience to them. Ah, that

45:11

sucks. Yeah, that's your first mistake.

45:13

Yeah, I know, I know. Troublesome

45:16

facts. No, so in, first of

45:18

all, there were some really clear,

45:20

bright lines. Obviously, ISIS was extraordinary,

45:22

you know, unambiguously clear, like we

45:24

don't like terrorists on our social

45:26

platforms. Though, actually, it's funny enough.

45:28

If you go back in time

45:30

and you pull up the old,

45:33

you know, you know, you know,

45:35

you know, you know, you know,

45:37

media articles about that Twitter is

45:39

for a while there making the

45:41

one man's terrorist is another man's

45:43

freedom fighter kind of argument wondering

45:45

like is it a slippery slope

45:47

if we start taking these accounts

45:50

down and like where will it

45:52

end So for a while, they

45:54

would be called, you know, like

45:56

unintentionally based. It was a, but,

45:58

you know, what, what, it was

46:00

like the botical on attack and

46:02

a couple of other things that

46:04

really shifted the narrative as people

46:07

began to realize that for our

46:09

earlier chat, propaganda does work in

46:11

fact, and at some point, some

46:13

susceptible people commit atrocities and, and

46:15

that is, you know, the, the

46:17

trade off of that, you know,

46:19

they began to, you know, to,

46:21

to take down some of that

46:24

content. What happened with Russia then,

46:26

you know, this was one of

46:28

these things that was a little

46:30

bit frustrating for me because I

46:32

was asked to run one of

46:34

the research teams that analyzed the

46:36

data sets that were turned over

46:38

to the Senate Intelligence Committee. So

46:41

there were data sets that the

46:43

platforms attributed to the internet research

46:45

agency, just to be clear, I

46:47

did not attribute. I was just

46:49

given the data that they attributed.

46:51

And the internet research agency is

46:53

like the firm that It's kind

46:55

of widely thought of as like

46:58

executing a lot of like online

47:00

information information operation kind of things.

47:02

Yeah, they would use these these

47:04

trolls when they wanted to you

47:06

know, like kind of change the

47:08

information space ahead of invading Crimea

47:10

or you know, manipulating the American

47:12

discourse. They didn't do it solely

47:15

to interfere in the election. They

47:17

did it far more broadly than

47:19

that. But they did it through

47:21

the 2016 election, which was how

47:23

it came to be very caught

47:25

up in the Hillary Clinton Donald

47:27

Trump discourse. My report on that

47:29

was actually, it's funny because I

47:32

re-read it again recently as I

47:34

was fighting with Matt Tivey about

47:36

it. It's fairly boring to be

47:38

honest. It is extremely descriptive because

47:40

I wanted it to be as

47:42

neutral a description of what happened

47:44

as possible. Nothing that we were

47:46

given would have enabled us to

47:49

answer the question, did this swing

47:51

the election? Right? So I just

47:53

wanted to say, like, here is

47:55

what it is, here is how

47:57

it worked, here is what it

47:59

did, and you know, here's how

48:01

it intersected with these various communities

48:03

and here's why we should be

48:05

paying attention to it. However, because

48:08

of the broader conversation about Donald

48:10

Trump and collusion, which was wholly

48:13

outside of what I looked at,

48:15

it got caught up in that,

48:17

you know, that thing that came

48:19

to be called rushgate, which no

48:22

one can define, but you understand

48:24

it, you know, I too became

48:26

part of the quote unquote rushigate

48:28

collusion hoax. Right, because anybody who

48:31

did any research, even vaguely touching

48:33

Russia, was rebranded as this Democrat

48:35

operative trying to, you know, like

48:38

allege that Donald Trump's election was

48:40

illegitimate. So that was

48:42

a very frustrating thing. But again,

48:44

it was, you know, it's not

48:46

really the end of the world.

48:49

And to their credit, the Trump

48:51

administration did actually institute a series

48:53

of counter foreign disinformation efforts within

48:55

the FBI, within DHS, within

48:57

the intelligence community, within OD&I.

49:00

I feel like I've just

49:02

left one now. Oh, and

49:04

within the State Department, the

49:06

Senate expanded the mandate of

49:08

the Global Engagement Center to

49:10

counter foreign propaganda. That had

49:12

actually been established to counter

49:14

ISIS propaganda. terrorist propaganda, it

49:16

was established to counter Russian

49:18

and Chinese propaganda as well.

49:20

So around this time, China

49:23

had also begun interfering, Iran

49:25

had also begun interfering, and you

49:27

know, and as we found out,

49:29

like the US Pentagon was running

49:32

influence operations, it became table stakes,

49:34

but there was this like really

49:36

clear line again still around foreign

49:38

versus domestic. The problem

49:40

was what happened in 2020 was

49:43

that a lot of the

49:45

same kinds of dynamics of

49:47

wildly viral false

49:49

accusations intended to

49:52

manipulate and mislead

49:55

people began to emerge

49:58

from domestic American

50:00

influencers. And what we started to

50:02

see was that those narratives, false

50:05

and misleading claims, intended to manipulate

50:07

audiences, which do, in some capacities,

50:09

some specific ones, did meet the

50:11

definition of disinformation campaigns, were being

50:13

run by domestic American influencers. And

50:15

this created a real problem. So

50:17

those of us who were studying

50:20

the 2020 election, like Also at

50:22

Stanford Internet Observatory, Kate Starboard's team

50:24

at University of Washington Center for

50:26

an informed public, DFR lab. I

50:28

think you might have been a

50:30

DFR lab at the time, right?

50:32

I was. I didn't. I didn't

50:34

work on EIP. I was like

50:37

deep in militia land at the

50:39

time. Right. Which, you know, looking

50:41

back was maybe a blessing. Yeah,

50:43

you escaped. But as this, you

50:45

know, as EIP, this is what

50:47

we, for those, for the listeners

50:49

like that, was what we called

50:51

the sort of inter-institutional consortium that

50:54

came together. Grafica was the fourth

50:56

org. We came together again largely

50:58

because we thought we would see

51:00

a lot of foreign interference. And,

51:02

you know, Russia, Iran, China did

51:04

mess around, but the most impactful

51:06

claims, both rumors and, you know,

51:09

these allegations that like dominion voting

51:11

machines were changing votes and that

51:13

the election wasn't free and fair

51:15

that had been stolen. all the

51:17

things that we chronicled, the really

51:19

impactful stuff was not coming from

51:21

foreign trolls, it was coming from

51:23

Donald Trump and his inner circle.

51:26

And as we chronicled that, as

51:28

we documented it, as we occasionally

51:30

engaged with tech platforms and said,

51:32

hey, you know, some of these

51:34

posts appeared to violate your terms

51:36

of service, as we engaged with

51:38

state and local election officials and

51:40

said, hey, like this rumor about

51:43

Sharpie markers, this kind of a

51:45

big deal, you should probably respond.

51:47

That work that we did, which

51:49

we did in full view of

51:51

the public, we wrote a 250

51:53

page final report on it, was

51:55

recast two years later when in

51:58

2022 the house flipped and Jim

52:00

Jordan got his gavel, was recast.

52:02

passed by the weaponization committee as

52:04

a vast effort not to track

52:06

what was happening in the 2020

52:08

election, but to somehow mass censor

52:10

the 2020 election in real time.

52:12

And those narratives that I've described,

52:15

dominion, sharpie gate, these like massive

52:17

viral things that we watched go

52:19

by and that we chronicled as

52:21

they went by, they allege that

52:23

we somehow censored as they went

52:25

by, which is complete bullshit because

52:27

anybody listening. Like, Fox News paid

52:29

out $700 and something million dollars

52:32

over that dominion claim, but the

52:34

allegation that Jim Jordan wants you

52:36

to believe is that we somehow

52:38

censored it, right? Nonetheless, that was

52:40

really the tipping point whereby they

52:42

began to make this argument, particularly

52:44

because Donald Trump was also deplatform

52:47

after January 6th, that academic research

52:49

into rumors and propaganda was really

52:51

a vast plot to censor conservatives.

52:53

And that's because... the vast majority

52:55

of people who were moderated, you

52:57

know, or who had their accounts

52:59

taken away after January 6th, were,

53:01

you know, the Q&N on accounts

53:04

and some of the conservative election

53:06

deniers. I remember these, you know,

53:08

some of the, I believe, your

53:10

reports and, or the reports that

53:12

you were working on and other,

53:14

and other folks at the time,

53:16

and, and they were like, the

53:18

dominion pushers were, of course, Trump

53:21

and, and Don Junior and people

53:23

like that, but also like Tim

53:25

Poo And these are people who

53:27

all have some connections to Russia

53:29

as well, right? And one thing

53:31

where like Timpul was recently found

53:33

of that scandal, like that's 10

53:36

media. Right, yeah. So I mean,

53:38

is it like when we say

53:40

that they're like US influencers, I'm

53:42

not I am not saying that

53:44

it's like another Russia gate or

53:46

whatever buried inside the US thing,

53:48

but like have the lines broken

53:50

down so much between our American

53:53

influencers and like. Are the goals

53:55

so closely aligned between magga and

53:57

the Kremlin that it almost doesn't

53:59

matter anymore? There's no need for

54:01

a conspiracy anymore because they're

54:03

just kind of, I mean, look

54:05

what's happening, right? You see what's

54:08

happening with Trump and the Trump

54:10

administration and getting rid of everything

54:12

that the Kremlin would want them

54:14

to get rid of. Bottom text.

54:16

We wouldn't make that attribution at

54:18

Stanford and Observatory. We kept those

54:21

lines very bright because in our

54:23

opinion, we would talk about. the narrative

54:25

overlap is like one very very very

54:27

small part and that's because even when

54:29

I did the work for the Senate

54:32

on the you know the the data

54:34

set that covered from 2015 to 2017

54:36

Russian like the Russian trolls that were

54:38

targeting the right would just go pick

54:40

up turning point USA memes and they

54:43

would literally cover the turning point logo

54:45

with their own logo and they were

54:47

just plagiarized. And that's because it's always

54:49

been true. Well, I mean, they would

54:51

do this with like a whole lot

54:53

of different communities. You know, they would

54:55

target the black community and they would

54:57

pull from like black media and plagiarize

54:59

from there. They would plagiarize from news

55:01

media. They had this really terrible, you

55:03

know, like the live laugh love, like

55:05

that horrible shit, the Pinterest, Pinterest lady

55:07

stuff. Like they had a being liberal

55:09

one that targeted women that used that

55:11

kind of stuff, which I'm sure they

55:13

just pulled from Etsy. They didn't have

55:15

to come up with new stuff because

55:17

Americans do it to ourselves. And if

55:19

you want to create identity and divisive

55:21

content, you just go and take what's

55:23

already available to you. And so when

55:26

we would talk about foreign interference,

55:28

it was very important to us that

55:30

we weren't relying just on content that

55:33

we were always looking at, either some

55:35

sort of evidence of like a network,

55:37

you know, a network in play or...

55:39

explicit accounts where we could say this

55:41

is demonstrably tied to some sort of

55:43

foreign network, some sort of foreign actor.

55:45

So we did in our, you know,

55:47

in our interest to kind of like

55:49

keep those two things separate, like we

55:51

wouldn't have made that argument, but you

55:54

certainly can make the point that there

55:56

is a lot of ideological overlap and,

55:58

you know, I read the tenant. media

56:00

indictments also, there is clearly a realization

56:02

that that kind of stuff can happen.

56:04

I mean, they would also, you know,

56:07

they ran some, they call it peace

56:09

data, I think was the name of

56:11

it, if I recall correctly, like a

56:14

burning left kind of garbage site that

56:16

they tried for a while. So, you

56:18

know, any time they want to appeal

56:21

to some sort of, you know, niche

56:23

identitarian politics, they're going to go and

56:25

find it and find someone who can

56:28

boost it. You know, I've no doubt

56:30

that they're going to continue to do

56:32

it, particularly now that, again, I mentioned

56:35

kind of to its credit, most of

56:37

that election defense infrastructure was built and

56:39

established during the Trump won administration. In

56:42

the first six weeks of the Trump

56:44

to administration, it's all been destroyed, just

56:46

to be clear, all of it is

56:49

gone. Barn Influenced Task Force was dismantled.

56:51

All those employees at Sissa who... had

56:53

the audacity to say that the 2020

56:56

election was free and fair have been

56:58

placed on administrative leave, right? The Odi

57:00

and I, a lot of those people

57:03

have been, the political appointees are gone.

57:05

It's not totally clear what they're going

57:07

to staff there. That's always hard to

57:10

follow because it's, you know, within the

57:12

intelligence community. And then the global engagement

57:14

center was defunded back in December. It

57:17

seems like they're quite well set up

57:19

to just. you know, shoehorn in whatever,

57:21

whatever foreign influence things are going to

57:24

help them and just allow it to

57:26

happen. Well, you know, well, I mean,

57:28

it's, it's, it's that the justification for

57:31

it is like rushigate, like that's the,

57:33

rushigate, like, that's the, rushigate, like, that's

57:35

the, which again, I can't define rushigate

57:38

for you if I tried at this

57:40

point. It's like, you know, you know,

57:42

The other piece of it that's kind

57:45

of fascinating though is like where the

57:47

social media platforms are going to come

57:49

down because it has never been any

57:52

kind of regulatory requirement for them to

57:54

do those investigations and the So as

57:56

I mentioned, you know, Jim Jordan got

57:59

his gavel and he decided that that

58:01

the election integrity partnership and all of

58:03

the work that we had done was

58:06

part of this vast cabal. And the

58:08

argument that they made was that we

58:10

did our work at the direction of

58:13

the deep state because remember, Trump ran

58:15

the government, Trump appointees ran the government

58:17

during the 2020 election. So somehow, even

58:20

though this letter that we get from

58:22

Jim Jordan alleges that he's investigating the

58:24

Biden censorship regime, the people that we

58:26

were talking to worked for the Trump

58:29

administration. So we find ourselves in this

58:31

like the sort of surreal Kafkaesque, you

58:33

know, universe in which we're being accused

58:36

of being told by the government to

58:38

tell the tech companies to take down

58:40

conservative tweets. And they come up with

58:43

this number like 22 million tweets. And

58:45

they get that number 22 million by

58:47

going to our report, which again has

58:50

been on the internet for two years

58:52

by this point, and in our table

58:54

of the most viral narratives, we had

58:57

like the top 10 most viral narratives

58:59

that we had followed, which we calculated

59:01

after the election, right? We added up

59:04

all of the numbers, like 9 million

59:06

for dominion and, I don't know, 700,000

59:08

for Sharpie Gate, whatever it was. And

59:11

at the bottom of the column, it

59:13

sums up to 22 million. And so

59:15

the number that was the top, you

59:18

know, the sum total of the most

59:20

viral stories was 22 million, and they

59:22

just refrained that they literally just took

59:25

that number. And Matt Taaby and Michael

59:27

Schellenberger sat under oath in a congressional

59:29

hearing and said, Stanford Internet Observatory in

59:32

the Election Integrity Partnership, like censored 22

59:34

million tweets. Just a fucking insane allegation,

59:36

you know, and I was like, this

59:39

is so real. If I was, if

59:41

I was that sloppy and it was

59:43

like clear that I was that sloppy,

59:46

it's so crazy because like when I

59:48

was in journalism school, I was like

59:50

reading Tyvee, enrolling Stone. and you know

59:53

with my like long form investigative professor

59:55

just being like this guy is amazing

59:57

I want to be him and if

1:00:00

I had made the kind of blunders

1:00:02

he seems to make every week or

1:00:04

every time he tries to talk about

1:00:07

this I would have like Just like

1:00:09

shot off to an island somewhere and

1:00:11

like never shown my face again. You

1:00:14

wouldn't have been able to work No,

1:00:16

I wouldn't have a job You would

1:00:18

have faced repercussions because your ideology is

1:00:21

and is you can do almost anything

1:00:23

on the right and get away with

1:00:25

it In terms of this is really

1:00:28

important. I think as as a point

1:00:30

like I edited that report just to

1:00:32

be clear like one of the things

1:00:35

that and it was a it was

1:00:37

a massive team effort, you know But

1:00:39

I was the one who edited, you

1:00:42

know, like, aggregated. It pulled it all

1:00:44

together. Like, Kate's team wrote one chapter.

1:00:46

People at Grafka wrote the foreign influence

1:00:49

section. You know, different people wrote different

1:00:51

subsections. And I was the person who

1:00:53

was doing the, you know, the big

1:00:55

final aggregation. And I mean, I cannot

1:00:58

tell you how many rounds of editing

1:01:00

and legal review and just like. Like,

1:01:02

the anxiety I had for every fucking

1:01:05

sentence in that report, checking the citation,

1:01:07

going back, checking the citation, did we

1:01:09

get it right, right? Like, I think

1:01:12

I found one mistake at one point

1:01:14

where, you know, we issued a correction

1:01:16

and, you know, and put something out

1:01:19

after the fact, if I recall correctly,

1:01:21

it'd be up in the Stanford library,

1:01:23

you know, the stacks, just like a

1:01:26

URL that was not correct or something

1:01:28

along those lines. But the, um... You

1:01:30

know, I feel that way about everything,

1:01:33

even when I did the book, I

1:01:35

mean the sheer number of like hours

1:01:37

I spent, like both hiring fact checkers

1:01:40

and fact checking it myself and just

1:01:42

the, literally like the anxiety over not

1:01:44

wanting to be wrong because like your

1:01:47

credibility is on the line with this

1:01:49

stuff. Even the Senate report, man, just

1:01:51

knowing how many people were going to

1:01:54

go through that thing with a fine

1:01:56

tooth comb. And the one thing I

1:01:58

will say. is even after with all

1:02:01

the bullshit smear campaigns that Congress came

1:02:03

at us with they were never actually

1:02:05

able to find anything in it that

1:02:08

we didn't stand behind. They went with

1:02:10

smears and lies where they took these

1:02:12

numbers and turned them into bullshit because

1:02:15

they couldn't actually find anything that was

1:02:17

like wrong to use to discredit us

1:02:19

so they went with lies instead and

1:02:22

or with like we don't like your

1:02:24

speech so we're going to complain about

1:02:26

your speech because we can't actually find

1:02:29

evidence. you know and all the emails and

1:02:31

things so what happens is you know the subpoenas

1:02:33

begin to come in and of course we have

1:02:35

to turn over all of our documents all of

1:02:37

our work product all of the things to you

1:02:39

know all the emails that we're being asked to

1:02:41

turn over both with the government and with the

1:02:44

tech companies and of course there's no

1:02:46

emails in which the government is telling

1:02:48

us to tell the platforms to do

1:02:50

anything but what happens is like you

1:02:52

can't exonerate yourself this is the thing

1:02:55

I think that people really need to

1:02:57

understand right When you're hauled in front

1:02:59

of an investigation, this stupid, you actually

1:03:01

can't exonerate yourself. You have to think

1:03:03

about it as like House and

1:03:06

American Activities Committee. Like that is

1:03:08

the parallel here. You will never be

1:03:10

able to actually prove that you didn't

1:03:12

do the thing. Because when there were

1:03:14

not 22 million tweets, when you know,

1:03:16

I was like, let Elon turn over

1:03:18

the 22 million tweets, like they claim

1:03:20

that they're the Twitter files, boys, like

1:03:22

where are the 22 million tweets? You

1:03:25

know, when they don't have them, they

1:03:27

just move the goalposts. Oh, well, yeah,

1:03:29

maybe the government didn't actually

1:03:31

send you anything, but some analyst

1:03:33

flagged this tweet and we don't

1:03:35

like that you flagged it. And I'm

1:03:38

like, okay, but that's a complaint

1:03:40

about my speech. That's a complaint

1:03:42

about my First Amendment protected speech.

1:03:44

That's not censorship. That's my speech.

1:03:46

You know, and you just wind

1:03:49

up in this surreal universe where...

1:03:51

Nobody remembers the accusation that was

1:03:53

actually made about you and instead

1:03:55

they just complained about something else

1:03:58

and that is the like that is the process

1:04:00

here and for the point about

1:04:02

the sloppiness of the Twitter files

1:04:04

and the reporters, they're never actually

1:04:06

held to account for it because

1:04:08

they were actually asked by the

1:04:10

minority on the weaponization committee to

1:04:12

refile their testimony because Meti Hasan

1:04:14

actually did, funny enough, go and

1:04:16

read the report and he had

1:04:18

Matt Taebi on for an interview

1:04:21

at the Twitter files and he

1:04:23

pointed out that he was wrong

1:04:25

by, you know, that we had

1:04:27

flagged 4, 700 tweets. 4,700 URLs

1:04:29

in total. And he said, you

1:04:31

know, you were off by like

1:04:33

21,997,000, you know, 900 and something.

1:04:35

And why aren't you correcting this?

1:04:37

And they just never bothered to

1:04:39

correct it. So like that too

1:04:41

is, you know, sort of a

1:04:43

surprise, right? You can perjure yourself

1:04:45

if you're, you know, on the

1:04:47

right. If I can suggest a

1:04:49

title for the next report, if

1:04:51

that media ecosystem can offer us

1:04:54

any tips, it's maybe you should

1:04:56

call it big if true. And

1:04:58

that should help you wiggle. The

1:05:00

art of the shameless, right? Yeah.

1:05:02

The tech platforms have also sort

1:05:04

of changed their stance on this.

1:05:06

You know, a platform like Facebook

1:05:08

still has some teams. Not as

1:05:10

big. much resources they used to,

1:05:12

you know, hunting for potential for

1:05:14

an influence campaigns. But their external

1:05:16

funding of that kind of research

1:05:18

is pretty much gone. The rolling

1:05:20

back, a ton of content moderation.

1:05:22

Mark Zuckerberg wrote a letter to

1:05:25

Jim Jordan and was like, yes,

1:05:27

it's true, mean, mean, Joe Biden,

1:05:29

you know, threatened us if we

1:05:31

didn't pull content. And a lot

1:05:33

of these platforms are dealing with

1:05:35

European regulators. who actually have a

1:05:37

fucking backbone and like want to

1:05:39

regulate it or is the approach

1:05:41

perfect is it correct I don't

1:05:43

I don't know if to take

1:05:45

a moral position but they are

1:05:47

doing something where the US government

1:05:49

has done like historically not a

1:05:51

lot. And now it's like after

1:05:53

they've spent multiple years pummeling these

1:05:55

platforms, pummeling researchers, as Facebook, I

1:05:58

would expect behind the scenes potentially,

1:06:00

you know, Tik-talk as it tries

1:06:02

to escape being banned in the

1:06:04

US, has these platforms sort of

1:06:06

changed their attitude and sort of

1:06:08

not give up the fight but

1:06:10

really just deprioritize it a lot?

1:06:12

The tone of Jim Jordan has

1:06:14

been like, oh, come here, sweet,

1:06:16

sweet, sweet, Mark Zuckerberg, my boy.

1:06:18

What are the Europeans doing to

1:06:20

you? Oh my God. And it's

1:06:22

like now that they don't have

1:06:24

Biden to blame. They're planning to

1:06:26

go on the offensive. So I'm

1:06:29

just curious, how much are these

1:06:31

platforms just capitulating? Yeah, well, there's

1:06:33

a, I mean, there's so much

1:06:35

there. The first, content moderation is

1:06:37

both hard and not always well

1:06:39

done, right. And that's the truth

1:06:41

of it. Content moderation in most

1:06:43

of what we've talked about has

1:06:45

been, you know, state actors and

1:06:47

political propaganda and what is the

1:06:49

correct response to political propaganda. So

1:06:51

one of the things that we

1:06:53

argued, I don't even say we,

1:06:55

like that I argued actually, is

1:06:57

just that takedowns don't really work

1:07:00

a lot of the time. I

1:07:02

think people have different opinions on,

1:07:04

depending on what you study, right?

1:07:06

For people who study, maybe militia

1:07:08

groups, maybe violent speech, I think

1:07:10

people have different arguments about. platforming.

1:07:12

For a lot of the stuff

1:07:14

that I look at, the vaccine

1:07:16

speech, political speech, it just doesn't

1:07:18

work. It just goes somewhere else.

1:07:20

It doesn't actually make a difference.

1:07:22

And so a lot of what

1:07:24

I argued for over the years

1:07:26

was the labeling, was the community

1:07:28

notes, actually the community notes are

1:07:30

good, was the giving users more

1:07:33

control, was making recommendation engines better

1:07:35

on a variety of different axes.

1:07:37

And this is where, you know,

1:07:39

some of my current work actually

1:07:41

looking at blue sky as an

1:07:43

architecture that you can build. on

1:07:45

top of is actually trying to

1:07:47

get into that question of when

1:07:49

we say better what do we

1:07:51

mean and how and trying to

1:07:53

actually do those experiments. When we

1:07:55

talk about the shifts that Facebook

1:07:57

has made and the capitulations like

1:07:59

with that letter to Jordan, the

1:08:01

thing that was so interesting about

1:08:04

that letter to me was that

1:08:06

it was a response to job

1:08:08

owning. So we went through this

1:08:10

entire series of like, you know,

1:08:12

the arguments in which Big Neine,

1:08:14

Joe Biden, and the Biden censorship

1:08:16

regime, even during the time it

1:08:18

was run by the Trump government,

1:08:20

was, you know, was demanding all

1:08:22

of these things. He made that

1:08:24

capitulation. Rumor was for a while

1:08:26

to try to get out of

1:08:28

a couple hearings and subpoenas himself.

1:08:30

I don't know how much it's

1:08:32

really going to pay off for

1:08:35

him. The FTC is still kind

1:08:37

of saber rattling over this. So

1:08:39

is the FCC. There are, you

1:08:41

know, Josh Hawley, I don't know

1:08:43

how much people pay attention to

1:08:45

hearings in the way that I

1:08:47

do, but Senator Schmidt, who actually,

1:08:49

you know, launched the Murphy v.

1:08:51

Missouri court case that was one

1:08:53

of these kind of canonical cases

1:08:55

that SCOTUS actually wound up tossing

1:08:57

that made the allegations of the

1:08:59

censorship industrial complex and, you know,

1:09:01

that the Biden censorship regime was

1:09:03

trying to silence all of the

1:09:05

speech. SCOTUS found no evidence of

1:09:08

it. Toss, toss the case. Anyway,

1:09:10

Schmidt had a hearing on Tuesday

1:09:12

of Tuesday of this week. performative

1:09:14

nonsense. The most interesting point of

1:09:16

the hearing was actually Josh Hawley

1:09:18

saying even, you know, they had

1:09:20

these journalists, these right wing journalists,

1:09:22

talking about being censored and then

1:09:24

some free speech focused professors on

1:09:26

the left for the minority talking

1:09:28

about how the real censorship was,

1:09:30

you know, some of the things

1:09:32

that the administration is currently doing.

1:09:34

Anyway, Hawley came out and just

1:09:36

said like, look, for all of

1:09:39

what we're talking about, Mark Zuckerberg

1:09:41

made to be held accountable. So

1:09:43

it's interesting to hear some of

1:09:45

the senators still saying, like, for

1:09:47

all of the capitulation and groveling,

1:09:49

you have attempted to do, like

1:09:51

we are still going. to regulate

1:09:53

you. And it is interesting to

1:09:55

see that that dynamic continuing to

1:09:57

come from some of the lead

1:09:59

figures in the administration. So I

1:10:01

think Zuckerberg is trying to buy

1:10:03

himself friends, but it's not clear

1:10:05

that that has been successful. On

1:10:07

the European front, you mentioned like

1:10:10

regulators with spines. This is going

1:10:12

to really be a test of

1:10:14

those spines, because if you followed

1:10:16

what JD Vance has gone over

1:10:18

to your to do recently, A

1:10:20

lot the rhetoric has been that

1:10:22

content moderation of speech, even in

1:10:24

the European market, where they do

1:10:26

not have the First Amendment, where

1:10:28

they have different views of free

1:10:30

speech, right? Nazi speech is prohibited,

1:10:32

for example, that advance has made

1:10:34

the point that European content moderation

1:10:36

rules are censorship and that American

1:10:38

companies shouldn't have to abide by

1:10:40

them. And so the question is...

1:10:43

to what extent are the European

1:10:45

regulators really going to take a

1:10:47

stand and actually demand that the

1:10:49

American tech companies comply with their

1:10:51

laws, given that this could trigger

1:10:53

one of these stupid tariff wars

1:10:55

or the ways in which the

1:10:57

administration has been rather provocatively picking

1:10:59

fights with purported allies. It's quite

1:11:01

clear that Vance and others in

1:11:03

the administration don't have a particularly

1:11:05

high regard for Europe, quite a

1:11:07

lot of contempt there. And so

1:11:09

it's going to be very interesting

1:11:11

to see whether these laws are

1:11:14

actually, like whether Europe fights on

1:11:16

that sovereignty principle. I think you're

1:11:18

going to see Brazil encounter the

1:11:20

same dynamic as Musk and Zuckerberg

1:11:22

seemingly expect the administration to fight

1:11:24

this fight for them. I've been

1:11:26

talking a lot. Mike, do you

1:11:28

want to wind us down a

1:11:30

bit? Maybe we can talk about

1:11:32

like what's the answer to it.

1:11:34

Oh, no, I want to ask,

1:11:36

I want to ask a question,

1:11:38

which is, what's it like interacting

1:11:40

with Matt Taaby, which is because

1:11:42

I've seen your substance. And I

1:11:45

found it very entertaining in a

1:11:47

kind of gossipy media way because

1:11:49

just it's fun to see people's

1:11:51

private emails and like the attitude

1:11:53

they take. And I just found

1:11:55

him very petulant and ridiculous is

1:11:57

the way I would describe. Seems

1:11:59

like it's just a ridiculous person

1:12:01

who's kind of clinging on to

1:12:03

some idea that he's still a

1:12:05

journalist, but not really. How would

1:12:07

you describe these interactions where he's,

1:12:09

you know, back and forth reaching

1:12:11

out for comment and getting, you

1:12:13

know, and kind of getting fact

1:12:15

checked by you? and vice versa.

1:12:18

You know I think it's the

1:12:20

dumbfounded is the word that comes

1:12:22

to mind every time I get

1:12:24

an email from him but it's

1:12:26

it's been like this since the

1:12:28

very first email I got which

1:12:30

a long time ago he wrote

1:12:32

me a note after I'd written

1:12:34

an article talking about Elon Musk

1:12:36

acquiring Twitter and he wrote me

1:12:38

this random email asking like How

1:12:40

dare I write an article critical

1:12:42

of Elon Musk? And I thought

1:12:44

it was just the most bizarre

1:12:46

thing I'd ever seen. And he

1:12:49

said, you know, why didn't you

1:12:51

reference and then made an allegation

1:12:53

about a past company I'd worked

1:12:55

at and something that they had

1:12:57

done wrong and why didn't I

1:12:59

disclose it or something? And I

1:13:01

said, you know, can you give

1:13:03

me the rules like? You write

1:13:05

about Title IX issues and you

1:13:07

got tagged for sexual harassment. I

1:13:09

don't see you disclosing that at

1:13:11

the top of every article you

1:13:13

write about it. So like, can

1:13:15

you lay out the rules? Like,

1:13:17

what are they? That's my very

1:13:20

first interaction and then it only

1:13:22

got better from there, I guess.

1:13:24

The, you know, when he started

1:13:26

the Twitter files. He just sent

1:13:28

me a note that said, I

1:13:30

need you to comment on the

1:13:32

fact that you worked for the

1:13:34

CIA. And that was the only,

1:13:36

there was nothing like, there was

1:13:38

no question in this. Just comment

1:13:40

on the, just comment on the

1:13:42

fact. Yeah, that was literally it.

1:13:44

It was like, comment on this

1:13:46

publicly available information. Yeah, and I

1:13:48

thought like, what the hell does

1:13:50

that have to do with anything?

1:13:53

Like, yeah, when I was 20,

1:13:55

the smear started and. The way

1:13:57

that I have chosen And I

1:13:59

was like, maybe new to bad

1:14:01

faith attacks, like I'd gotten a

1:14:03

couple, but now having, now that

1:14:05

I am an old pro, bad

1:14:07

faith attacks after the last two

1:14:09

years, my response is just to,

1:14:11

is actually just to publish them,

1:14:13

right? And just to publish the

1:14:15

full, the full, unvarnished interaction, including

1:14:17

my responses so that people can

1:14:19

see what I said, what they

1:14:21

said, and that other people can

1:14:24

make up their mind. And my

1:14:26

frustration with Matt is that I

1:14:28

had requested. corrections on many things

1:14:30

that he has gotten wrong. He

1:14:32

has taken quotes that I have

1:14:34

given and cut them in half,

1:14:36

just foundationally changing the meaning of

1:14:38

them. He has, again, like, just

1:14:40

been factually wrong. Not opinion, like,

1:14:42

I don't dispute his characterizations. I

1:14:44

get it. He wants to, you

1:14:46

know, smear me, spin things. Okay,

1:14:48

this is the way that shitty

1:14:50

media works. Okay, I get it.

1:14:52

But just on the factual factually

1:14:54

wrong things like getting the dates

1:14:57

of my employment wrong who I

1:14:59

worked with wrong Projects he you

1:15:01

know he kept trying to act

1:15:03

as if I worked on Hamilton

1:15:05

68 I did not I had

1:15:07

nothing to do with it. These

1:15:09

sorts of things were I'm like

1:15:11

can you just be like just

1:15:13

just tell the truth you know

1:15:15

what's interesting also the Elon Musk

1:15:17

aspect of it because this is

1:15:19

this guy like really kind of

1:15:21

presents himself as like you know

1:15:23

Jimmy journalists speaking truth to power

1:15:25

and it's like, is there any,

1:15:28

is there any, is there, have

1:15:30

we ever encountered anyone during our

1:15:32

lifetimes as powerful as Elon Musk?

1:15:34

I mean, the guy literally just

1:15:36

bought his way into being president,

1:15:38

you know, and, you know, in

1:15:40

so many words, I can't imagine

1:15:42

anyone as powerful as Elon Musk

1:15:44

and yet. Is the richest man

1:15:46

on the planet or Elon Musk?

1:15:48

Your little press cap and being

1:15:50

like, yeah, I'm stuck, is sticking

1:15:52

up for the little guy and

1:15:54

doing my, you know, you know,

1:15:56

I'm inflicting the power like pain

1:15:59

on the powerful and like taking

1:16:01

taking orders from this like cosmic

1:16:03

freak. How dare the richest man

1:16:05

on the planet face criticism from

1:16:07

the towering behemoth Renee duressa?

1:16:09

It's amazing. I did see

1:16:11

that a lot. Pick on

1:16:13

someone your own size Renee

1:16:15

that's why. It wasn't even

1:16:17

a main article in the

1:16:20

Atlantic. It was like fairly

1:16:22

fair I thought. When you

1:16:24

report on the kind of the

1:16:26

sort of the magga right or

1:16:29

even just white nationalists or

1:16:31

whatever. There's a lot of

1:16:33

effort to sort of portray people

1:16:35

in our position as being

1:16:37

just overwhelmingly powerful. It's like

1:16:40

it blows my mind, you

1:16:42

know, considering the type of

1:16:44

very powerful people that they fluff

1:16:46

up all the time. I just really

1:16:48

felt like, look, if you want to,

1:16:50

if you want to criticize the

1:16:52

work that I did, I would just.

1:16:55

like you to actually criticize the work

1:16:57

that I did, not make up 22

1:16:59

million tweets or a, you know, or

1:17:01

a, in this particular case, in my

1:17:03

most recent engagement with him, he was

1:17:05

trying to discredit the idea that Russia

1:17:07

interfered in the election by insinuating

1:17:10

that the number that Facebook had

1:17:12

put out that 126 million people

1:17:14

had seen and engaged with the

1:17:16

Russian content was really a number

1:17:18

that could be dismissed because it

1:17:20

came from me and my report.

1:17:22

And I kept saying it doesn't come

1:17:24

from me, it doesn't come from

1:17:26

my report, it comes from Facebook.

1:17:28

And I pointed that out with

1:17:30

links, with dates, with like, you

1:17:33

know, it's unambiguously true. There is

1:17:35

no universe in which it's not true,

1:17:37

and I just could not get over

1:17:39

the fact that I had to have

1:17:41

six back and forths about this, and

1:17:43

that rather than just correcting it,

1:17:45

he then turned it into like, well, duressed

1:17:47

a quibbled with the number. And I

1:17:50

didn't I didn't quibble with the number

1:17:52

at all actually I took it at

1:17:54

face value because it was a reasonable

1:17:56

assessment but I quibbled with their Instagram

1:17:59

number because it wasn't and you know

1:18:01

it was just like it's just

1:18:03

an extraordinary thing to pick a

1:18:05

fight over and I you know

1:18:07

it like ate up a Saturday

1:18:09

and then I thought okay I

1:18:11

can't do anything more with this

1:18:13

and I'm just gonna go on

1:18:15

with my life. Well that's what

1:18:17

I wanted to know I just

1:18:20

wanted to know what it's like

1:18:22

to to interact with with Matt

1:18:24

Taaby particularly this new sort of

1:18:26

mutated version of him that is

1:18:28

like now just you know completely

1:18:30

beholden to... billionaire money. I think

1:18:32

all you can do is is

1:18:34

put out the the full exchange

1:18:36

when you deal with bad faith

1:18:38

media. There are plenty and I

1:18:40

don't I don't even I deal

1:18:42

with conservative media and I get

1:18:44

plenty of reasonable inquiries and if

1:18:46

I don't feel like it's an

1:18:48

immediate bad faith inquiry I don't

1:18:50

screenshot it and toss it on

1:18:52

the internet but at this point

1:18:55

whenever something comes in from him

1:18:57

it's a screenshot and toss it

1:18:59

on the internet there's no way

1:19:01

it's going to go well. I'm

1:19:03

excited to see if this podcast

1:19:05

episode makes it into Taiy's lore.

1:19:07

Anyway, Renee, thanks so much for

1:19:09

spending some time with us today.

1:19:11

Listeners, go check out Renee's book.

1:19:13

It's called Invisible rulers, the people

1:19:15

who turn lies into reality. I

1:19:17

can't say enough good things about

1:19:19

the book. It really is great,

1:19:21

Renee. Do you have anything else?

1:19:23

People should check out. Loose Guy

1:19:25

mostly and haphazardly write newsletters sometimes.

1:19:27

Oh yeah, where can people see

1:19:30

the Matt Taaby email exchanges? Oh,

1:19:32

that's on sub stack because that

1:19:34

actually gets good SEO as opposed

1:19:36

to newsletters. But yeah, so just

1:19:38

sub stack and Renade arrested. Maybe,

1:19:40

Jared, can we link to it?

1:19:42

Yeah, yeah, we'll put a link

1:19:44

down in the episode description. Yeah.

1:20:06

You