Chris Hayes on How Your Attention Gets Monetized

The Trump administration has undertaken

a series of strikes on civilian boats.

These are not military boats.

They're civilian, they say drug traffickers,

although in some cases it seems like they're fishermen.

In some cases it seems like maybe they're both.

They're like fishermen who are paid,

no, really- Yeah.

Like they're fishermen who are paid some money

to run a product somewhere.

Yeah, right? People who are trying

to make ends meet. Yeah.

Our forces have killed over 100 people this way.

And what's been so striking about it

other than how I think both legally and morally

and defensible it is to just essentially murder people

on the high seas, is that from the beginning

it has been produced as content.

Like Tom, very Tom Clancy,

the sort of under, the unclassified.

It looks like an '80s movie, which I think

is exactly the kind of genre touchstone for Donald Trump.

Yes.

[Katie chuckles]

So, the first cut at that answer would be,

yes, they perform aggression war,

imperialism foreign policy all as content,

all as means of gaining attention,

holding attention. Yeah.

I mean, there's the iconic shot of they've got Twitter up

during the Venezuela raid. Oh, yes, yes.

Seeing who's tweeting about it.

[Katie] Wild.

But then underneath that,

there's also the fact that this is real bombs and real guns

and real missiles and real people die,

and there's real children numbering

maybe as much as 150, 180 who are dead in Iran

because our missiles or Israel's missiles

were still not clear, killed them in a strike.

They're doing it for attentional reasons, right?

Because the president likes to keep everybody's attention.

He has to be at the center of attention,

he has to be doing all the time.

He has to have you thinking about him.

And also they have very old school,

I mean, pure 19th century,

straight up, no chaser imperialist ambitions.

Yeah.

So, it's sort of, it's imperialist ambitions

in a vertical video rapper, in a social media, always on-

Yes. Content machine.

And I think there's actually kind of an interesting

and profound point to that,

which is that you could make the argument

that these have always been intertwined.

I mean, if you look at the history of American imperialism

and the Spanish American War

and the famous Hearst Papers and the Yellow Press,

that was both about conquest and producing content.

Mm-hmm, mm-hmm, sure. I mean, quite famously.

And in some ways is sort of the dawn

of the American newspaper era.

So, I think these two things have always been twinned.

The sort of history of imperialism is also a history

about kind of the propagandistic uses of it

to capture and hold the attention of the masses.

But I think yes, their version of it is a very kind

of 21st century post-modern vertical video scroll,

doomed scroll version of it. Right, right, right.

On tech steroids.

You wrote in a piece for The New York Times,

unrelated to Iran, I should say, quote,

President Trump has a feral,

almost pathological genius

for getting people to talk about him.

You called it suffocating to his opponents.

When you think about your role,

our role as media, we're both journalists,

to not be the foot soldiers helping to fuel that dynamic,

when you go tape an episode of your show, let's say tonight,

what decisions do you make about how to approach, let's say,

what's happening in the Middle East

to avoid sort of playing into that hand,

to avoid playing into that imperialism

on social media steroids?

Well, the thing we can't do is ignore him,

or what he's doing, right?

So, like we, the US actually is at war with Iran.

There are real human lives.

The latest account is 1,000 plus Iranian civilians,

not to mention we don't know how many combatants

or members of the regime.

You can decide whether political figures

in a regime counted civilians or not.

Human lives are human lives.

Human lives are human lives.

So, in that sense, it's like,

he's the president of the United States.

He has the nuclear codes.

He's now launched multiple forms

of extra territorial killing, let's call it.

Yeah.

So, the way that I think we do it is to try to,

A, not do war porn- Sure.

With our B-roll.

Like there is a subtle,

but unmistakable ideological substrate

to certain forms of depictions of war.

Try to avoid that.

Also don't let him set the terms of things,

which is like, we're not gonna play huge chunks

of whatever his nonsense is

except to sort of set them up to show why they're lacking.

But like there's no really avoiding it,

I guess is what I would say in the end.

Donald Trump being the President of the United States,

which is the most powerful nation on Earth,

having access to nuclear codes and also the full force

of the American military,

and also attempting to replace the constitutional order

with essentially a presidentialist personalist dictatorship

is the top story of our time.

[Katie] Yeah.

And I cover that story every night and the question is,

on whose terms do you give attention

and what you give attention to?

So, here's a great example of that.

They made an amazing miscalculation in Minnesota.

There was this viral right-wing video

that was alleging to uncover fraud in Minnesota daycares

run by either Somali immigrants or Somali Americans.

There actually was,

there has been this huge fraud network there.

It's been prosecuted and investigated,

in fact, by the US Attorney's Office,

and by the very prosecutors who would later resign

because they didn't like

what the Trump administration was doing.

Trump administration saw this and they were like,

We want to bring more attention to this.

We're sending Bovino and CBP and ICE there.

What ended up happening was that they kidnapped people

and they killed two Americans in broad daylight on camera.

Yeah.

And that was where all the attention went.

And you could see there was this Trump backpedaling

where he was furiously posting on Truth Social,

like, You should be talking about the fraud.

It's like, you just killed two people.

We saw you all.

And then you called them, you slandered them,

call them domestic terrorists.

You're kidnapping people's neighbors.

You're teargassing high school students, this is-

Yeah, yeah, yeah, and we can all see it, it's everywhere.

So, that was a great example to me of like, right?

You're not, like were we paying attention to Donald Trump?

Yes, at some level.

Was it on his terms?

No, right?

I mean, I think that's basically

the question we ask ourselves.

Now, your book, Sirens' Call, is out in paperback now,

it's a great book. Thanks.

And it's all about this, it's about attention.

And you argue that attention has become a commodity

in the same way that labor was made a commodity

in the early years of industrial capitalism.

I'm curious about when you start the clock

on that process, right?

Like, when did the process

of commodifying attention really start?

How do you track that through history?

You know, it probably starts

with two technologies commercial billboards,

and the Penny Press The New York Sun,

you know, it's 19th century basically,

early to mid, depending on how sophisticated it gets.

In both of those technologies,

the idea was that you were selling an audience

and you had to come up with metrics

that you could measure that audience

and then sell that audience to advertisers.

So, early technology in billboards,

there would be people who would stand

by the corner the billboard was on

with clickers- Oh, wow.

And they would go like this and they'd say,

you know, we get 600 people an hour, you know,

if you're talking about Times Square or something

or we get 25 people an hour or whatever it is.

And then you can go to your advertiser and say,

This is how many people are gonna see it.

The Penny Press, you know, the big innovation there

is that you sell the paper

for less than it costs to make the paper.

You lose money on every paper,

but then you sell the advertising.

[Katie] Right.

And so that's the initial process.

It gets, there are multiple iterations,

magazines come in and then you get radio

and then you get television

and then you get, you know, social media and the internet,

all of which are getting more sophisticated

at measuring that audience and selling that audience.

There's always a degree,

I mean, hilariously about this technology,

like there's always a little bit of a,

maybe a bait and switch by the people who tell you their,

like when you sell a magazine,

it's like, does everyone look at that page?

Right.

Sure, sure, sure- Do they open it up

and read it, right?

And that's been a frustration

by the people buying the ads for years, right?

Like the old thing about,

I'm wasting half my marketing budget,

I don't know which half. Right.

You know, like this thing about,

how well are you measuring?

Do people stay in the room

during the TV commercial during ER

or is that when they get up to get a sandwich?

And if they, you know, if they're getting up

and get a sandwich, like you can't really measure that.

Nielsen Box just tells you the TV's on.

So, what has happened though is the global scale

that you can sell it at is new, right?

So, no media companies ever had billions of users before,

like these attention companies do now.

The amount of data you have about your viewer

is orders of magnitude more. Yeah.

The micro second auctions that you could run

in each second about how you're going to serve that viewer

and specifically them.

So, you now have this sort of auction

for eyeballs happening in nanoseconds, you know, constantly.

And the thing about the algorithm

is you don't have to have people making programming choices.

There's just a ton of stuff up

and you just see what people start to look

and then they, you serve them that over and over.

So, you know, the old model, which was,

well, what are we putting on the front page

of our Penny Press, you know, The New York Sun,

or what are we putting the AM/PM time slot on NBC or?

What content will get you to spend eight or nine hours

on TikTok is basically the calculus now,

and it's one that-

Except, but they don't have to make the choice-

People aren't making it.

Exactly. Exactly.

And that is actually a huge difference.

I'm curious very much so,

in the book, you acknowledge your own sort of role

in this economy, right?

You're an attention merchant yourself, you're a TV anchor,

you're also on social media, right?

You film clips for your Instagram account-

Yep. For MS NOW.

How do you navigate your own role

in that sort of algorithmic attention landscape?

Like how do you, I think part of it is sort of

how do you grapple with it as a human being?

And then how do you think about it strategically?

I think there's a few different ways

depending on the platform or the medium.

I mean, I think with my television show,

I have been doing it long enough and established,

I think, enough of a rhythm and a relationship with viewers

that I am not in a position where I am,

I have a sense of where attention is flowing

and that's an imperative for me.

I mean, I say in the book, it's necessary,

but never sufficient. Yeah.

Like if no one watches my show,

then I haven't done my job. Yeah, yeah.

No, you have not.

So, I have, the first level

is that I have to get people's attention

and then the second level

is I have to do something worthwhile with it.

And sometimes those are intention.

Because sometimes the best thing to get someone's attention

is not that worthwhile to me.

That hasn't really been a,

like on a week when we just went to war,

that has not been really a problem.

This is one of those weeks

where I'm not real tortured about it.

Like the audience's attention is slowing

towards the fact we just started a war with Iran.

I think that's the most important story.

I'm not like internally torn- No, sure.

And they come to you because you can help them cut through

and understand what's going on and-

There was the period where the plane went missing

and I was on air for that.

And, you know, that was a three-month story.

And at a certain point it's like,

yes, it's an interesting story.

It's certainly a newsworthy story.

It's tragic, but-

The plane is still gone.

The plane's still gone

and the audience still wants you to talk about it.

I mean, that was really a difficult thing.

It really did, like the signal and the noise was,

just keep giving us a plane.

I still feel the pressures every day.

And on my podcast, I feel the same way too.

Like I just do what I'm interested in the podcast

and I let the chips fall where they may.

Social media's interesting.

We've been doing more and more vertical video

because everyone does,

and it's- Everyone does.

I think it's such a weird slot machine effect.

Like I did this thing the other day about the,

there's a pretty little notice set of tariff votes

in the House that Donald Trump lost

that were House votes to overturn some of the tariffs,

including the Canadian ones.

It actually happened

before the Supreme Court struck him down.

I did a little thing about like,

this is kind of interesting,

like he's lost Republican votes on this.

It blew up. Oh, there you go.

And sometimes you're like,

This one's gonna blow up and then it doesn't.

And you're like, Well, what did I do wrong?

And it's unclear to me,

I guess if I put more time into this,

if this my whole life was like playing the slot machine,

remember I get better at playing the slot machine?

[Katie] Yeah.

And obviously there are people like Mr. Beast and-

Sure. I think people lose-

Who are very good at it. Yeah, but-

Well, and I think that's sort of one of the challenges

as journalists or sort of purveyors in my view

of accurate and newsworthy information

is you are competing now not with a couple

of other cable news shows, you're competing with Mr. Beast

and with cooking videos and, I mean, with everything-

Every piece of content- Every single thing

in that feed.

Every piece of content is at every moment pitted

against every other piece of content ever created.

Yeah.

So, in many ways that tariff video doing well

is like a little miracle. Yes-

It's a miracle in the internet.

Right, but then I was like,

I was really like feeling myself about it and I was like,

Oh, it's awesome.

And then my next one I was like.

Yeah, but the thing is,

you essentially have to participate, right?

You know, you know what I mean?

Like you can't opt out of shooting host to camera,

vertical video anymore?

I mean, right, you can [crosstalk]

but not if you are trying to-

You're trying to reach- Reach people.

A mass audience with news about tariffs.

Yeah, I mean, this is the problem.

First of all, I think vertical video ends up

being a kind of terminal point [laughing]

in the development of attentional technologies

just because it's such a slot machine apex predator,

you know, it's very difficult to compete with.

I wanna ask you a little bit about the midterms,

because we are getting to be at that season.

You wrote in that piece

for The New York Times I mentioned.

You argued that the Democrat's main problem

isn't their message.

You're reflecting on the Harris campaign.

You said her core problem was her inability

to get people to hear her message,

it wasn't the message itself.

So, basically an attention deficit,

which I would argue still a problem

for the Democrats heading into the midterms.

I'm curious about your view there

and sort of your view on the Democrat's ability

to galvanize an electorate online

in the way Trump and the GOP were

going into the 2024 election.

Sort of, has anything meaningfully changed there

in your view?

Yeah, so the reason that I said that,

I just think it's important,

one of the most important pieces of data

that we had from 2024 is that amongst voters

who said they paid a lot of attention to the news,

Harris won by five or six points.

And as you moved further down,

like sometimes to literally never,

Trump's margin increased. Yeah.

So, and I say this for two reasons.

One is a lot of people like to blame the media

for Trump's victory and it's like,

well, the people that consume the most like journalism

and news media were the most Harris inclined.

So, it complicates that story quite a bit.

The thing that comes after that

is for a very long period of time,

basically I would say from the 1980s until recently,

there was a very straightforward

kind of theory of attention in campaigns,

which was you raised money and then you spent it on TV ads.

[Katie] Yeah.

One of the points I was making there is like,

that's clearly broken down.

You can't just say, we're gonna raise a lot of money

and then we're gonna run a lot of ads on what?

The local news, who's gonna reach who exactly?

[Katie] Right, right, right.

Some of the voters you need,

you know, but a lot of voters that you need are not there.

So, you need to have a theory of like,

how do you reach the people that don't consume media,

which is like, we used to call like earned media.

Oh, yes, earned media. Right?

So, it's like earned media is like,

you're interviewing me. Yeah.

And then there's paid media,

which is like you're running ads on TV.

Chris did not pay for this interview.

I did not pay for this interview.

So, it's like, if they're not gonna see your earned media

because they don't consume that

and they're not gonna have paid me,

like you gotta come up with some theory.

Do the Democrats have a theory?

Well, I think they've gotten better at it.

I mean, I think that the idea

that Donald Trump kinda went everywhere in 2024

and talked to all kinds of different podcasters,

and made all sorts of content,

including him like driving around that truck

and serving McDonald's. Oh, yeah.

Dude-

Sorry, none of this is actually funny, but-

No, I mean it was- Absurd.

Absurd and kind of comical

and actually pretty effective.

[Katie] Yeah.

The kind of thing that like clearly reverberated out

through the world past people that consumed the news,

past paid advertising. Yeah.

So, you know, I think Zohran Mamdani obviously

was like a huge innovator in this.

You know, his-

Yeah, he did an incredible job.

Yeah, the vertical videos.

Now, it may be the case, like Roy Cooper in North Carolina

is incredibly well known.

He just won the nomination to be the Democratic nominee.

He's gonna go against Michael Whatley,

the Republican nominee for that Senate seat.

Now, Roy Cooper's super well known.

He's been elected statewide,

I wanna say three or four times.

He was the governor for two terms.

Also, he's gonna raise a ton of money

and it may be the case for Roy Cooper.

He can just, he's got a theory,

he's gonna raise money to run ads [crosstalk]

and people know who Roy Cooper is.

Yeah. But like James Talarico-

I was gonna ask you about James Talarico, yeah, yeah.

Yeah, so he's a, you know, he's State Rep.

Jasmine Crockett,

he defeated Jasmine Crockett

in that contested Democratic primary in Texas.

Now, Jasmine Crockett obviously has a theory of attention,

which is-

Well, I was gonna say, I sort of,

I think that that is interesting to me

because I feel like they both have theories

of attention. Exactly.

They're just very different.

They're very different,

but they both- Very different.

But they both had a sense of,

how do I become known?

How do I make sure that people

whose votes I'm gonna want or need know even who I am?

Yeah.

My point is that you better have a theory of this, right?

Like that's based on who you are that cannot be.

What you cannot do is you cannot default

to what had been the paint by numbers approach

for literally decades.

I'm gonna raise a bunch of money

and I'm gonna run a bunch of local TV ads.

That is not going to work.

I wanna ask you about where politics and technology

sort of collide and come together.

We talked a little bit about this at the top,

but, you know, when I got to Wired,

it was very obvious to me and I think to the team here

that covering politics more closely

was not an optional decision, right?

It was an imperative.

There was no space between Silicon Valley leaders

and the government,

particularly true after Trump took office.

You've been a political journalist for a very long time.

You spent your career,

I think, observing and documenting

how power shifts in government.

How have you sort of seen that merging of power

between those two spheres,

between sort of the Silicon Valley elite, the tech industry,

and politics and politicians?

What does that look like from where you sit?

I mean, I, yeah, I thought the inauguration

was such a shocking moment.

Was that shocking to you

when you saw them all sitting there?

I'm genuinely curious. It was.

Yeah, I mean, the fact that the support wasn't shocking,

but the, we're all gonna stand up here with him

and, you know?

Because there are downsides to those calculations

and usually they're thinking about those downsides.

I think a few things happened.

I think as the industry matured

from an insurgent industry to an incumbent one,

its politics got more right wings.

You, yeah, yeah.

This is not a very surprising trajectory.

[all laughing]

It does happen.

It's like, yes,

like, sometimes if you interview someone who's 23

trying to break into something

and you interview when they're 63

and they make six figures or seven figures

like they have some different politics.

[Katie] For sure, yeah.

And so I think part of it is that.

I think part of it is they absolutely all cook their brains

on the internet and Twitter and with each other.

I mean, I think they just-

Like spun each other up.

I think they like just pickled their brains

in a brine of reaction.

Huh. I really do.

I just think they're in like the same way

that people talk about, people talk about this all the time.

They talk about, and this is a very common documented

almost trope, which is relatives

who were lost to Fox News. Oh, of course.

You know, people that just,

you know, they used to have a certain set of views,

maybe they're kind of Right-leading and a Republican

and they just started watching Fox all the time

and it almost did something at like a chemical level.

They're just angrier and more irascible and so,

and wound up.

And I just think there's a kind of right,

like a online version of that

that happened to the tech elite.

And then I think there's just a political economy of it,

which is that these are the most powerful

and profitable corporations-

Yeah. In the world.

And then, of course, the big part of it

is the AI bet, right?

So, that's like the final component.

I think they were kind of cooking their brains,

they were personally getting radicalized and kind of,

I think there's a lot of like backlash reaction politics.

I think they're mad at their Woke workers,

I'm putting that in air quotes.

They were a mature industry that wanted

to cozy up to power in the government,

and then they had this technology

that they think is the kind of make or break technology

and their relationship to the state is existential.

Yes, yes. As we're seeing

with the fight over Claude and Anthropic.

Yes, which I was gonna ask you about

and that's also interesting.

I mean, I think I'm always curious to hear different people

and sort of smart people who look at this in different ways

about whether they see what has happened as more ideological

that these sort of tech elites, tech leaders,

like genuinely move to the right.

And this is legitimately how they feel,

that this is the right way to run a country,

the right way to run a business,

the right way to work with government

or whether it is simply like they are biting their tongues

and holding on for four years and presenting the president

with tchotchkes in the Oval Office when they need to

because they have a fiduciary obligation

to their shareholders, they have a huge base of employees

that they need to support and pay and uphold,

and that is just sort of this fundamental like,

I hate this guy, but I have a business imperative.

And which of those two sort of feels weightier

to different people has been interesting?

Yeah, And I think there's different individuals

on different side of that.

I mean, I think- Sure, yeah.

They're also not all the same person.

I think Bezos has gotten very right-wing

in his personal politics.

I mean, I don't think he was ever a liberal

by any stretch of the imagination, but I think he's just,

I think, I mean, obviously Musk has got, you know-

Elon be Elon, yeah.

Yeah, the terminal brainworms.

Yeah.

[Katie Laughing]

But like, yeah, Tim Cook, I don't know, you know?

I think that's an example of-

He looks deeply uncomfortable,

but he won't say anything- He's giving him the babbles.

[Katie] Yes.

So, yeah, I don't know.

And I think the folks at Anthropic are interesting

because I think they're sort of,

they seem like they're the last gasp

of this older version of sort of soft progressive tech-

Yes. You know, and again,

I don't want, these are not like,

you know, social justice warriors or-

[Katie] No.

We're not talking about the labor boom.

No.

No one needs to give them any sort of peace prizes

or major awards for this.

But I think Dario Amodei is not a person

who is like consumed by seething

and toxic right-wing reaction.

I do not think that at all.

No, I think- He clearly is not that.

In Fact, he-

I think he's consumed by other things-

Yeah, but- The sort of existential,

the existential thinking around artificial intelligence

I think feels very, very prominent for him.

Totally. Which I think

is one of the reasons that we saw

what happened with that company and the Department of War,

if I may call it that,

is that to him, this very much is a life or death situation.

Yeah, you know, it's now a Joe Klein that the,

you know, Sergey Brin and Larry Pages

don't be evil model for Google,

but again, like, that was,

you know, it looks ridiculous now,

but I think it was, A, genuinely felt at the time.

And I also think they thought, you know, again,

the sort of the trajectory of how this happens,

when Google was created, it was the perfect example

of like genuinely building a better mousetrap.

I was searching the internet at the time,

all the searches were bad.

Yeah.

Google came along and- It was very good.

So much better.

And it was better in a way that transformed

the usability of the internet.

And I think they thought,

well, this is a, we're providing a good service

and we're gonna sell advertising

and like a pretty ethical business, which like,

yeah, this works pretty well, you sell it out,

you know, and then slowly over time, right?

Yeah, things escalated from there.

Things change. And here we are.

I mean, when you think about looking at the next,

I mean, we've got three more years of this administration

and you think about the level of proximity,

of collaboration, of I'll say collusion, you don't have to,

but the sort of very close relationships

that we see with some,

between someone like Sam Altman and the administration,

does that scare you?

Yeah, I honestly was chilled to my core

when there was, you know,

there's a meeting between the head of Anthropic

and the head of the Pentagon

and which they can't come to terms on,

basically it's a terms of service agreement

of implementing Anthropic's claude model

in Pentagon situations,

and then the Pentagon throws this temper tantrum

that sounds completely deranged.

Like bond villain kind of thing,

and we're gonna try to cut them off

from all their supply chain risk like-

[Katie] Yep.

You can sell NVIDIA chips to the Chinese government-

[Katie] Yes, but God forbid-

But you can't use,

I mean, come on. Yeah.

And then for Sam Altman opposed this thing being like,

Hey, we've swooped in and we've made a deal.

Look, these companies, particularly those two,

which are the two that are startups, right?

OpenAI and Anthropic. Yeah.

They're not the legacy incumbents

that, you know, have their own AI models

like, you know, Gemini or whatever.

You know, they're on a treadmill.

They gotta run fast,

and they gotta raise money

and they are, you know, their revenues are increasing a lot,

but their costs are increasing arguably faster.

[Katie] Yeah.

And there's a sort of sense of desperation.

And I think you, people,

the combination of people with a very powerful technology

who are banking on making a world-changing fortune,

but also have a kind of like Ghosts in Pac-Man

like financial burden trailing behind them-

Yes. Is-

To the tune of like many, many, many, many, many,

many, many, many billions of dollars.

Yes, that's a- Shocking amounts of money.

That is not, I would not say that that is the best setup

for like ethical and responsible decision making

or decision making that takes into account

the stakeholders involved,

and I think that's incredibly terrifying.

It is really scary.

It's scary and I wanna ask you a little bit more

about how you're thinking about AI.

I think you've called yourself a lame centrist.

I would say I probably fall into a similar camp,

at least in the context-

You think on the AI debate?

On the debate, sorry. Yes, yes.

Not when it comes to what OpenAI should

or should not be doing

with the Department of War. But yes, on, yes.

But the conversation around it,

you know, very polarized.

You've got the Doomers, you've got the Boomers,

very sort of overly simplified arguments.

I would say in general, I think we spend too much time

like ping ponging between those two extreme views

and maybe not enough time talking

about the practical implications

or like the potential future scenarios

that we really should be taking seriously.

You posted on Bluesky that the Left needs to quote,

Start thinking seriously about the AI hype being true.

Tell me more about that

because I remember seeing you post that

and/or something to that effect and many strong reactions.

Thing I like about Bluesky is that I think it's a place

that is, you know,

the kind of general vibe around AI is like,

This sucks, I hate it.

And- Yes.

And I think in a world

in which it's being kind of shoved down in our throats

and there's like- Oh, it's exhausting.

You know, billions being spent to make us love it,

like that's a use, that's a very useful perspective.

Yeah. It's also the case that,

you know, the thing that I'm most worried about is the very,

the job replacement issue.

You know, I think a world in which all of these jobs

that people have right now,

from coders to first year law associates

to the administrators

who work at large health insurance companies,

work at hospitals,

you know, the world in which those are automated

in a relatively short period of time,

it's gonna cause some pretty profound dislocation.

And do you, is your general sense that not enough people,

maybe people on the left

are taking that seriously [crosstalk]

that there's like a head in the sand kind of falling off-

I think there's an idea that if you take that seriously,

you're seeding to their own propaganda

about how useful their product is.

Do you think that's true?

Well, no, I don't.

I mean, I think there's a huge question about how quickly

this is all gonna happen. Yeah.

But I also just like, I can see it.

I mean, I use several different LLMs

for different things and-

I was gonna ask you a little bit

about like your personal use cases.

I have been using them more

because I think I wanna understand what they do.

Yeah.

For instance, like Notebook LLM

where you can upload sources and then you can use it

to navigate from those sources [crosstalk]

is extremely useful.

Yeah.

So, that you're not getting hallucinations

and it's also citing back to something.

So, if I say like,

Oh, what date did this thing happen?

The obscure historical detail that only you,

that won't be Googleable,

A, because Google no longer really works,

but B, because it's embedded in a PDF

of a scholarly article I've uploaded to you.

So, there's lots of useful ways of using it to me

when you're bounding sources,

particularly for research purposes,

but it's just manifestly getting better.

Like obviously- Yeah, 100%, yeah.

And it's just like this idea that it's not as insane,

and the idea that it's not gonna start

to touch like jobs people do is also seems insane.

Yeah, I have a few friends

who are very senior level software engineers

who until very recently, maybe December,

thought this was just so ridiculous.

Yes. They were like,

our CEOs won't stop talking about it.

They're insufferable.

This will never, this is just a ridiculous toy.

It's the new metaverse, it's the latest tech fad.

And then I think it was the Claude Coach release

that all of a sudden they are writing book proposals,

they're trying to use AI to like launch businesses.

Like they're trying to figure out like,

what's my next thing?

Because I'm in my 50s

and I wanna work another 10 or 15 years, I need to work,

and I don't think I'm gonna have a job in a few years.

And that for me, just sitting and having a drink with them

and listening to them talk about that

was pretty jarring actually.

Yeah, and I think, again, you know,

part of the problem is,

I think that we should have some clarity

about the business case proposition

of this technology is exactly this.

Like, the reason that they think

they can sell this to people

is that if you're paying a first year associated law firm

to do something and you're paying them $120,000,

and you can sell them a tailored version of ChatGPT

or Claude or Gemini that's law specific

that can basically replicate that work for 60, 80 or 90.

I mean, think about how much money revenue

that is for you, right?

If you can sell that, and also it's a savings for them.

And the only problem is that someone's out of a job, right?

That's the business case.

So, like, part of the problem is,

if you start to talk about that,

it does feel like you're seeding-

Well, I was about to say-

Their propaganda business case-

When you get that reaction from the left,

it's like, But I don't like this,

and I don't want this to be the case.

It cannot be the case that all of these elites

and oligarchs in Silicon Valley are telling me

that I'm not gonna have a job.

I don't accept that. Yeah.

And so what would it look like to you

for quote, unquote, The Left, to start to have,

to start to take that more seriously?

How does that manifest?

I mean, I think you gotta start thinking

about like job protections, like do,

like how do we wanna deal with that?

[Katie] You mean like meaningful regulation?

Yeah, meaningful regulation.

I mean- Well, first of all-

[Katie] I'm laughing a little bit because I-

First of all, we should be regulating AI

and I don't know how,

I think the idea that it's like-

Terrifying.

Totally unregulated is insane.

Yeah.

So, I don't know the answer.

The big thing I think is it does seem possible to me.

I think I'm radically uncertain about the future trajectory.

If you take a step back

and you think about automation, right?

Like agriculture is wildly automated

and we use far, far fewer people than we used to per,

you know, acre and per productivity.

Manufacturing has gotten more and more automated.

It's both been outsourced and also automated.

What we call knowledge work

is gonna maybe get automated in the same way.

That's not like a crazy thing to think.

It's not crazy.

No, I think it's uncomfortable.

I think it's uncomfortable- In fact that's the point

of the technology is to do that.

Yeah.

There's use cases that already seem

like it's pretty good at it.

It's going to develop future.

And so the question of like, well,

what does that mean and look like?

I mean, and I think partly that's going back

to real Bluesky thinking.

I mean, go back to, you know,

John Maynard Keynes wrote this great essay called the,

on the Economic Prospects of Our Grandchildren

that he talks about I think it's in the 1920s of like,

you know, what happens when we just solve

the economic problems and,

because we have enough and no one has to work anymore?

It didn't go that direction,

and I doubt it'll go this direction too.

[Katie] Right.

But I do think the lesson of both that

and the kind of 1990s economic consensus

that I think really was destructive

and fundamentally was a,

you know, what they have in common

is a redistribution from labor to capital.

I mean, we just last year,

this year, hit the highest share of national income

for capital versus labor of all time.

[Katie] Hmm.

To be thinking in a broad sense of like,

well, if all these jobs were automatable,

like if you didn't need people to do all these things,

what do we want people doing?

What does society do?

Like these are,

we're so locked into,

I mean, the reason I like the Keynes's essay

is we're just locked into like,

where are the good jobs gonna be?

Yeah. And, you know,

work has changed over time and productivity has increased

and the big thing is like, well, what do you want,

what should a person have a shot at?

What should they be guaranteed?

What should they have a shot at

in a wealthy society like ours?

And how do we order the society fairly to do that?

And that's like real first principle stuff,

but I do think in some ways this calls

for some real first principles thinking.

And I don't know that I want Donald Trump

to be the person, you know, making those calls.

No, I mean- So, the timing

is very unfortunate. No.

And I think, but I mean,

the same in grace he's not gonna make those calls

because he's just gonna let the AI companies run rampant

and do whatever they want. Sure.

But I also think that like small acts of resistance,

I think like people at the grassroots level

fighting data centers,

is that like the solution?

No.

But is it a way to operationalize this sentiment?

It's like, wait a second,

you're telling me this is going to replace all of us.

This thing is driving up,

it's now 7% of total US electricity consumption.

[Katie] It's wild, yeah.

It's driving up local electricity prices.

It's intentionally being created as a technology

that will move the distribution of national income

from labor to capital,

and you want to build one in my town?

No.

Like that's a totally good, legitimate act,

like actionable way to start.

Well, I think a small act of resistance

in the context of data centers

is a bit of a hopeful place for us to end.

But before we do,

I want to play a very quick game that we invented.

I don't know if you were briefed on this ahead of time,

but it's too late now.

It's called Control, Alt, Delete.

Okay. So, I want to know,

what piece of technology would you love to control?

What piece would you love to alt?

So, alter or change,

and what would you love to delete?

And people have been very generous

in their interpretations of that question.

Someone tried to control the weather,

and I didn't have the heart to tell them

that that is not technology.

I mean, I guess I want to control AI.

Yes, well- Because, you know,

I guess if I trust myself maybe more than Sam Altman's-

[Katie] Arguably, I would trust you more

than most of the people involved in AI.

So, that sounds good to me.

I mean, yeah, I guess if I could control it,

I'd figure out a humane ethical, I'll be the one.

It's Chris Hayes. I'm the one.

He's in.

Alt, well,

Alt I've got, which is I would love to alter internet search

so that it works again.

Yeah, what's your beef there?

I just- What's going on

with your searches?

I just think that it's gotten so bad.

It's not great. It was,

I think the quality of the product

of just Google Search particularly.

There's other ones that people, you know suggest

that I've used as well,

but essentially it's all, you know,

it's been displaced now by AI

partly because Search got so bad.

[Katie] Yeah.

And so, but I,

it's nice to search things and be able to find them,

and that's become more and more difficult.

Do you just get that big like AI box that people-

Well, first you get the big AI box.

And you have to X out of it and then you-

You also get the big AI box,

but you also get overwhelmed by ads,

and you also just, the search does not surface things

that you're looking for as well as it used to.

Well, if Google is listening.

It's broken. Yeah.

I mean, it's really gotten bad.

Here's my delete. Okay.

Well, I just, I wanna get rid of cell phone calls

and replace them with landline quality calls.

Is it the device being used that you have an issue with

or is it the caliber of the voice on the other side?

I think cell, I find cell phone,

cell talking on the cell technology

is the highest level of failure that we tolerate

from any technology in our lives.

[Katie] Just in terms of like the service

being patchy and-

If your oven just shut off as often as a call dropped,

or you couldn't hear someone,

or your alarm didn't go off as many times,

or your computer just didn't turn on

as many times as you can't hear someone or a call drops,

it's insane-

[Katie] We would be in the streets.

No one would- Yeah.

No one would tolerate it. No.

And it's the reason people text all the time-

Sure. And don't talk.

And the other thing that I hate about cell phone calls,

you know, FaceTime audio can fix this,

WhatsApp audio a little bit,

is that they don't have what's called side tone.

[Katie] What is that?

Okay, when you were in junior high,

you're, I think we're roughly same cohort,

when you were in 12 or 13 or 14,

and you would go home after school and talk on the phone.

Yep, oh, yeah. For hours.

Oh, yeah.

You would be, when you were talking on that phone,

that landline, you would be hearing your own voice

through the receiver in what's called side tone,

because the way that a landline works is that,

in the same way that when you have cans on,

you're doing a podcast- Yeah.

And you're getting your own voice in your ears-

Yep, okay. A landline does that,

and it is such a better, more pleasurable way

to talk to someone because you can calibrate

your own volume.

Cell phones don't have side tone,

which is why people shout when they're on phones,

why people sound weird, why you sound weird,

why you can't actually have good

and intimate conversations on cell phones.

It's why people always wanna put their headphones in,

even though you,

the headphone doesn't even give you side tone.

So, this is my big, I guess that's an alt

because I like to bring side tone into cell,

but I basically just wanna like delete cell vocal technology

and start over.

That is so specific, and so like well studied.

I really applaud that, I love that one.

Chris Hayes, thank you so much.

Thank you. This was fabulous.

Thanks.

[logo whooshing]