Safeguarding the Internet's Future

[upbeat music]

Hi everyone.

Matthew, thanks you so much for joining us today.

Thank you for having me.

Very excited to talk to you

and a about a really important topic.

We, I think the title of this discussion

is Safeguarding the Internet's Future.

Yeah. Which feels

appropriately dramatic and extreme,

but it is, it's a dramatic and extreme situation.

I wanna start by introducing you a little bit

and your company, if that's all right.

Because CloudFlare is so critical,

it's critical infrastructure to the internet,

but it's maybe not the most public facing,

people might not be as familiar with it.

So you are the co-founder and CEO of CloudFlare.

Yep. Which is a company that,

as I tried to think up an appropriate metaphor.

Here's the best I came up with. Tell me if I did okay.

I feel like CloudFlare is like the Internet's highway patrol

in some ways because you make sure traffic is getting

where it needs to go.

You take care of any pile ups,

but you also pull over any bad actors

and stop them from going wherever they're headed.

Is that a fair? Yeah.

How do I do?

I tend to say that Cloudflare's job is

to help make the internet faster and protect from bad guys.

But highway patrol, that's-

Just like a highway patrol does!

Just like highway patrol does, right?

You can use-

I don't actually usually think of highway patrol

is trying to make the traffic go faster but it's okay.

You know?

Well, we all, we- Maybe in the future,

We all drive our own way.

So because of that,

because CloudFlare has such broad visibility

into what's going on on the internet,

you have an very interesting perch.

You saw at some point over the last year or two,

you tell me when something that started

to ring alarm bells for you.

And like so many things that we're talking about today,

and in tech in general, it has to do with ai.

Do you mind going back to when you first started

saying, Hey, something's not right here.

Yeah. So CloudFlare sits in front of

more than 20% of the internet,

and we have millions of clients.

And about two years ago we started getting calls from

a bunch of our media clients

and they said, you know, We have a new threat.

And I said, Oh, is it North Korea? Is it Iran, is it-

And they said, No, no, no, it's ai.

And I remember being like, ah, media people,

like, why are they such Luddites?

We are dramatic. They are.

And you're overly dramatic.

And it's like every technology is gonna destroy the world

and what are we gonna possibly do?

And the sky is falling.

And so it actually took quite a while to convince me

that this was a challenge.

And finally one of our customers said, Just pull the data.

Just go look at how things change.

And we have really reliable data for the last 10 years

on how sort of trends on the internet look.

And so we took 10 years ago as sort of the baseline,

and we said, okay, 10 years ago,

for every two pages that Google scraped from your website,

in order to build their search index,

they would on average across the entire internet,

send you one visitor back.

And the business model of the internet has been, you know,

always generate content that drives traffic

and then sell either things, subscriptions or ads.

And that's been the business model

of the internet for the last 30 years.

And so 10 years ago we saw the data,

it was for every two pages they scraped,

they sent one visitor.

If you look at Google today,

that's up to 20 pages scraped for every one visitor.

And what's changed is instead

of Google giving you 10 blue links that you click on,

and you know that when, when it's a search engine,

the search isn't done, when the results come back,

you still have to go on the treasure hunt

to go find the thing.

But that's less and less true.

If you do a search on Google today at the top of the page,

there's an AI overview.

And if you all just think about

how many times are you relying on that ai overview

as opposed to clicking on a link,

you can see all of us are saying,

well, that oftentimes gives me a great answer.

That's the good news for publishers.

The bad news is if you look at open ai, it's 1500 to one.

If you look at Anthropic, it's 40,000 to one.

And that started to say, wow, no matter what,

as more and more of the world looks to AI

and more of the, we stop having search engines,

we have more answer engines,

that the business model of media,

and I'd actually argue the business model

of the entire internet is fundamentally going

to change over this, over this period of time.

And that's if our mission is

to help build a better internet,

then we thought it was really just important for us

to start thinking about this

and trying to figure out what is

that future business model of the internet gonna look like.

Well, yeah, because you, as you say,

it's not 10 blue links anymore.

It hasn't been for a while, right,

'cause Google's done featured snippets

or whatever for a long time,

but this is fundamentally different

as we shift to answer engines.

And I guess as you saw it, you said,

All right, we need to do something about this.

Do you mind talking through July 1st was a big day for you.

Yeah. And CloudFlare,

do you mind just telling people

what you did on July 1st to try to switch that calculus?

Well, what the, the first thing we said is what could

that business model look like in the future?

And the reality is that for a lot of people

putting things online, so like CloudFlare,

we have developer documents,

we want those to be in AI engines.

But on the other hand, if you were running a business

where your business is putting up ads

or selling subscriptions, if all of a sudden AI is taking

that data and basically republishing it

without you getting any traffic for it,

that's actually a real threat

to your fundamental business model.

And this is not a hypothetical threat.

No, no, no.

It's for me, It's an enormous challenge

for Wired and Conde Nast

and DocDash and all of these media companies

that are seeing less and less traffic coming to their sites.

And so what we said was, listen, our primary job has been

how do we stop threats like again, North Korea or Iran

or whatever it is.

But we can use those same tools

to give folks like media companies the ability

to control whether or not AI is able to come to them.

So on July 1st we declared Content Independence Day,

which gave for free tools to anyone who wanted to say,

listen, I don't necessarily want AI

to be taking my data without compensating me in some way,

and those tools then got enabled.

And what's been amazing has been

that then we've stopped over 400 billion AI requests

from taking content data.

And what we've seen is that organizations like Conde Nast,

which is, you know, Wired's parent, have been able

to do much better deals

with the AI companies in order to license their data.

So 400 billion stopping 400 billion AI requests.

It sounds like a lot. It feels very impressive and it is.

I do though wanna get a sense of

what that means in the broader scale,

'cause CloudFlare itself just said 20% of the internet.

So you, 20% of the internet, you're helping do this.

How much is still happening, right?

It's not like this is a solved problem all of a sudden.

So can you help put that 400 billion number in context?

Yeah, I, you know, I think it's,

we've tried to be very cautious here

and we've tried to say, we wanna make sure

that this is is done in a way which is right,

and you know, if you look at the history of the internet,

the great patron of the internet

for the last 27 years has been Google.

Google invented not only the search engine

to sort of index the internet.

It was in their interest for more webpages to exist

because that actually makes search more valuable.

So they then built all of the tools to be able

to like monetize the traffic that it was, was going there.

And they've really funded the creation of the internet

over the last 27 years.

Today though, they actually have created

what is the biggest challenge

that keeps us from totally locking things down,

which is Google has combined their search crawler

with their AI crawler.

You can't opt out of one without opting out of both,

which is a real challenge.

And it's crazy.

Here's the stat that sort of blew me away.

That we've not talked about ever publicly before.

How much more of the internet does Google see

than say OpenAI or Anthropic or Meta or Microsoft

or anyone who is doing this?

The answer is Google sees today 3.2 times more pages

on the internet than OpenAI does,

they see 4.6 times more than Microsoft does.

They see 4.8 times more than Anthropic or Meta does.

They have this incredibly privileged access.

And so when we ask the question of like, why did Gemini get

so much better than OpenAI over the last little bit,

I think that the answer is because they have more data.

It's not about the chips, it's not about the research,

it's not about the technology.

And the problem is that

we can't fully shut down the ability for AI crawlers

to access information on people who want

that control without you dropping out of search.

And if we block it for everyone else,

we've just given the game to Google.

And so I think that the most interesting thing,

it's almost like a Marvel movie.

The hero of the last film becomes

the sort of villain of the next one.

Like Google is the problem here.

It is the company that is keeping us from

going forward on the internet.

And until we force them

or hopefully convince them that they should play

by the same rules as everyone else

and split their crawlers up between search and ai,

I think we're gonna have a hard time, you know,

completely locking all the content down.

Now you say until we force them,

I am not unfortunately in a position to do so.

I don't know that CloudFlare is either, but, or?

Maybe. I wanna hear

more about that.

I mean there, I think we have

very constructive conversations with them

and they are nervous that,

that we are in a position to do things.

I think the most likely outcome, unfortunately,

is that it's going to require

some sort of a regulatory effort.

And I think that what the UK is looking at right now,

where they have, they have sort of designated Google as,

you know, a strategic market participant,

which is British for Monopoly.

[crowd laughing]

It's true, that they are gonna look at one of the remedies

of potentially splitting crawl apart.

And if that happens,

I think that's actually gonna be the best thing,

not only for media publishers, but for AI in general.

It shouldn't be that you can use your monopoly position

of yesterday in order to leverage in

and have a monopoly position in the market of tomorrow.

And so if you are competing with Google

and we all forget, but the reason OpenAI started was

because Elon and Sam were so afraid

that Google was gonna run away

with the game that they thought they needed something

that was a counterbalancing force.

That's still the case.

And I think that in order for this, to be fair,

we have to say everyone has to participate

on a level playing field.

And you can't leverage your unique relationships

that you've built because of search

in order to have unique relationships

and access to content because of ai.

I wanna go back just like 90 seconds to,

when you think about ways, when you strategize, you say,

okay, if regulators don't get it done,

we have some mechanisms that we could use to get this done.

What do you think of, what are some things you think

about some levers that CloudFlare can pull to get Google

to enter this sort of-

Well, we could just block them across, you know,

again, we had a little bit of an outage a couple weeks ago.

You saw how much of the internet went down. Apologies.

But, that's, we're in a position

to be able to influence what they do be.

And, and if you think about the very nature of

how PageRank works is it follows the links that are around,

if a huge chunk of the internet

just drops out of Google search,

PageRank breaks for the entire internet.

So, so again, I I don't think that we're powerless

to push back, and I am hopeful

because every time we have meetings with Google,

well, they'll say, well, we're never gonna do this.

Quietly we'll get anonymous emails which are like,

listen, we still, there are a lot of us here at Google

who still believe in the internet,

still believe that we have to support it.

And the reality is that if we don't find ways for people

who are creating content to be able to get compensated,

then you're gonna stop creating content.

And that's bad for all of us,

but that's actually bad for the AI companies as well.

There has to be some viable business model that exists.

And again, I don't even know exactly

what that's gonna look like,

but something's got to be able to replace

what has been ads and subscriptions,

because in a world of ai, it's not gonna look like that.

What I think people don't realize is AI is a platform shift

the same way that we went from the web to social to mobile.

AI is the next platform shift.

I keep watching reruns of the old cartoon show,

the Jetsons, and there's a lot of things

that are anachronistic about it,

but I think asking the question,

where does George get his information from?

It's a really interesting one.

And the answer is Rosie the robot, right?

Where he says, Hey Rosie, I want a recipe

for chocolate chip cookies.

Rosie doesn't say, here are 10 blue links.

Go, go find one yourself.

Rosie says, Here's a recipe for chocolate chip cookies.

And when, and then when George says, you know,

I go order eggs like it shows up on pneumatic tubes,

that seems probably not like how the future's gonna look,

but Rosie, we now are all carrying around in our pockets.

And that's going to disintermediate the relationship

between readers and content creators,

but also small businesses.

Like we've talked a lot about media.

What's gonna happen to small businesses going forward

all of a sudden, like what is a brand?

A brand has historically been a shortcut for human beings

to understand value and quality.

In the future,

if Rosie is disintermediating that relationship,

what's the role for the small business?

Where is Rosie gonna choose to get eggs from?

It might not be the bodega down the street,

it might be something very different.

And these are questions like the business,

the fundamental business model of the internet

is about to change massively.

And I don't know exactly what it's going to be,

but the most interesting question in the world today is

what is that business model gonna be?

And if you're not thinking about that for yourself

and your own business, that's a much bigger change.

And it could be great, but it also could

pose some real challenges.

Now, I believe that you believe in a better internet.

And that is why CloudFlare is doing,

that's something I believe,

what I know is that you have a fiduciary responsibility

to CloudFlare shareholders.

So this is not an altruistic mission necessarily. Right.

What is Cloudflare's skin in the game here

that's at stake for your company?

[Matthew] I have voting controlled company.

So well then for yourself.

A fiduciary responsibility,

I know I like, I want us to be a successful company,

but I, we don't exist if the internet doesn't exist.

And I think that this, if we don't think about

what the future looks like

and design the incentives for a good future,

like we might be in a place where content doesn't create it,

where media companies can't exist,

where you can't eat

because there's not some business model for you to exist.

So what I'm thinking about is

what is a positive future going to look like?

It shouldn't be one where there are five AI companies,

there should be 500,000 AI companies.

We need to make it as easy as possible for anyone

to create an AI company.

It shouldn't be one where there's just a, you know,

handful of media companies.

We should make it as possible for everyone

to be creating content,

sharing knowledge, making that available.

It shouldn't be one where there's this massive consolidation

of business behind a few major conglomerates,

which left to its own devices,

AI has this massive consolidating influence that's there.

We need to be able to make sure that businesses large

and small flourish on a fair playing ground.

That is the future that we're trying to play for.

That's the best thing for our business

because that's more people to be customers of ours.

That's more internet for us to be able to protect.

And if we don't have that future, if we do have a world

where, you know, it's a very small set of AI companies,

you know, talking to a very small set of people,

creating media, talking to a very small set of companies

that are out there,

I think that's an existential threat not only to us,

but to basically everyone

who's sitting in this room today.

And I think that's an interesting,

there's an interesting distinction here too,

because your position is not AI bad, right?

No. And which,

and I think there is some tools that you've introduced too

that speaks to, that there's a pay per per bot click

a trial that you're right now a beta program.

Yeah. Is that correct?

Do you mind talking about how that's going so far?

Are publishers or other content creators actually

seeing some real benefits from that?

Our AI company was on board with that.

What, how is that going?

To be clear,

AI is a, a massively good thing for 95% of use cases

for 95% of people.

You know, an answer engine is better than a search engine.

And so like that's going to be the better business model.

And what's amazing is, if you think about it,

there's a lot of things that were wrong

with the business model of generate traffic

and then sell ads or subscriptions against it.

It's what's led to I think, a lot of polarization in society

because the way that you generate

that traffic often is rage baiting

or click baiting people into going to your sites.

And that's like, none of us feel good about that.

We don't want that to be the case.

What is happening though is

for the first time in human history,

we actually have a mathematical model of knowledge.

It's not perfect, but that's what all of these LLMs are.

And I picture it like a giant block of Swiss cheese.

There's a lot of cheese,

but there's a lot of holes in in the cheese

that are there.

And what's interesting is

what the AI companies want is information

that fills in the holes in the cheese.

It's actually what you all want too.

You want that original, interesting creative content.

Not the thing that just makes you mad,

but the thing that actually furthers human knowledge.

And so there's an alignment between what I think we want

as a society and what the AI companies want.

And the best example of this is if you look at two,

what I think are great media organizations,

the New York Times and Reddit, right?

If you look at their corpuses,

they actually have about the same amount of words,

the same number of tokens in the two corpuses.

And yet who got more when they licensed their content?

The answer is Reddit

by about seven times what the New York Times got.

Why because even the New York Times

is an amazing media organization,

and even though the New York Times has played

the sort of traditional media game better

than almost anyone else,

if you don't have the New York Times,

you just can train on the Wall Street Journal

and tell AI to rewrite it as if it's a New York liberal

and you have the New York Times.

Now that's not fair to the New York Times entirely,

but it's kind of fair.

Whereas if you don't have Reddit, you don't have Reddit.

And so what I'm super optimistic about

and what we're seeing is that information

which is more local,

information which is more unique,

information which is more original,

that's what's getting compensated.

My wife and I bought our local newspaper in Park City, Utah

we have a circulation of 4,000,

and I've been astonished.

And again, I run CloudFlare,

so we've locked it down and AI can't, you know, crawl it.

I've been astonished how many AI companies are reaching out

saying, we wanna license your data,

'cause the answer is that if you don't have the Park Record

in Park City, Utah, then you don't know

what the hot restaurant is that just opened

or what's playing at the local theater

or you know, what the new runs are

at at the ski resort.

We have that information

and the AI companies need it.

So what we've already seen is that

as you lock down information,

that's the first step in order to create a market.

We talk about markets need supply and demand.

That's not exactly right. They need demand.

And the AI companies have an insatiable amount of demand.

Why, again, is go, why is Gemini better than everyone else,

'cause they have 3.2 times more pages

they see on the internet than OpenAI does.

That, again, that's not fair,

but that's just the reality today.

They want to get as much information as possible

and they're willing to pay for it

if it's something that advances real human knowledge,

if it fills in those holes in the cheese.

And what I think is optimistic about that is, again,

that aligns with what we would all prefer

the media to look like going forward.

So what I'm hearing is that people should subscribe

for things they value, like in-depth tech reporting

and analysis.

And click on that!

That you can't find anywhere else.

And click on the ads and subscribe. Perfect.

Matthew, I wanna say in, in the few minutes we have left,

AI is also evolving, right?

Agentic AI is coming, the landscape is changing.

How does CloudFlare change with it?

And I'm gonna bundle maybe an impossible two parter

that what does this all look like in five years?

Yeah, so, so I feel like we're sort of this like,

we're not a media company,

but we believe in the internet, we're not an AI company,

but 80% of the leading AI companies use CloudFlare.

And so I feel like we, you know, we would not,

if you'd asked me this three years ago,

I would not have predicted we were in this position,

but we're in a position to sort of be that traffic cop

to say, okay, what are the rails that we need to build

to make sure that agents can go?

But then what are also the guardrails that have

to exist in order to make sure

this is going to be secure and healthy going forward.

But I think even at a higher level, how do we think about,

you know, how do we make sure that media companies continue

to exist and evolve

and that you can continue to eat?

How do we make sure that small businesses are going

to be able to thrive and survive

and partner with great companies like Shopify or PayPal

or others to say, here's the toolkit that you need

to serve in an agentic world,

no matter what the business model of the internet

is about to change dramatically.

I don't know what it's going to change to,

but it's what I'm spending almost every waking hour

thinking about.

And I think it's the most interesting question in the world.

And if you're not thinking about it for yourself

and your business, then again the world might change

underneath you massively.

And I think that that creates amazing opportunities

and we should be all trying to figure out

what they're gonna be.

Well, I'm sure we will sooner than later.

Hopefully we'll see. Matthew Prince, thank you so much.

Thank you. For being here.

Really appreciate it.

[upbeat music]