← Masters of Scale
Loading summary
Transcript80 lines
- [00:01]
Deal Advertiser
Startups grow in unpredictable ways, but your infrastructure shouldn't. Deal scales with you from Employee 1 to Employee 1000. Deal's real time payroll engine cuts weeks long cycles down to minutes and their global experts keep you compliant as you expand. Plus, startups can tap into perks like up to $25,000 in deal credits and support programs designed specifically for high growth teams. Deal Deal is the fastest way to hire, manage, pay and equip anyone, anywhere. Deal your Forever People platform. Visit d e-l.commos to learn more and book a demo
- [00:40]
PJ Vogt
on Masters of Scale. We talk a lot about inflection points, those moments when a technology stops being optional and starts being fundamental. AI presents one of those moments. And behind every breakthrough, whether it's medical research or education or and cinematic visual effects, there's infrastructure most people never see as the essential cloud for AI. Core Weave is building that foundation. It provides a purpose built AI cloud designed specifically for pioneers at leading AI labs and enterprises. Ready to power the biggest ideas and the boldest ambitions. Ready for Anything. Ready for AI to learn more about how CoreWeave powers the world's best AI, go to CoreWeave.com ReadyFor Anything.
- [01:34]
Matthew Prince
Google is the hero of the last 30 years of the Internet. They financed the entire thing. The problem is that AI breaks that. 18 months ago it was 20 times harder to get traffic from Google than it was 10 years ago. Now it's 50 times harder. It's 3500 times harder to get traffic from OpenAI than the Google of old. 65,000 times harder to get traffic from Anthropic. Why did Sam and Elon start OpenAI? Because they were terrified that Google was going to run away with the whole game. Everything is started effectively as a counterweight to Google. And that's the puzzle.
- [02:17]
Bob Safian
That's Cloudflare CEO Matthew Prince speaking to me live on stage at south by Southwest Austin, Texas in mid March. Cloudflare helps facilitate more than 20% of all online traffic, which gives Matthew unique perspective about how AI is changing the Internet and the motivations of the major tech platforms. He also shares his real time experience on the front lines of cyber war with Iran. This is Matthew's second appearance on the show and he doesn't disappoint. So let's get to it. I'm Bob Safian and this is this is Rapid Response.
- [03:00]
Eve Tro
So please join me in welcoming Matthew Prince, CEO and co founder of cloudflare. Yes,
- [03:13]
Matthew Prince
these look very official.
- [03:15]
Eve Tro
Very official. This is a very official gathering. We're very official folks you were on my show six months ago, and since that time, so much has changed. You talked about how the number of new websites has grown dramatically after plateauing
- [03:31]
Matthew Prince
for a really long time, actually declining for a little bit. We're now seeing the fastest creation of new websites that has happened in the entire history of the web.
- [03:40]
Eve Tro
We have the number of. The amount of bot traffic has gone from about 20% of all traffic to trending to 50% by 2027 and growing.
- [03:50]
Matthew Prince
We think it'll be. It'll be. Over half of Internet traffic is generated by bots.
- [03:54]
Eve Tro
And the business model of the future is unclear. You gave this example about how even Walmart, Amazon and Target have wildly different business models. Right. Wildly different approaches.
- [04:07]
Matthew Prince
It's. Walmart has said, agents are welcome. Come on, come all. Amazon is literally suing companies. And it was just successful in a. In a motion against perplexity, saying, you are not, if you're an agent, allowed to shop on Amazon.com, two of the smartest retailers in the world have wildly divergent strategies. That's so rare that that happens. And I think it just shows how uncertain everyone is about what the future is going to look like.
- [04:33]
Eve Tro
To add to all of this, we now have a conflict in the Middle East. Really? Yeah. And I wanted to. Well, I wanted to ask you about this because I remember when you talked to me, you said, and I'm going to quote this, Cloudflare goes to war every day with Iranian hackers, Russian hackers and North Koreans. Now, what is the status of that now with the conflict going on?
- [04:59]
Matthew Prince
Yeah, so it was interesting. Sometimes we see things, we can't explain them. But on February 27, the attacks coming out of Iran, so from Iran Targeting the west, increased 7x off baseline. So there's always sort of background. And then on February 27th, massive spike. Why? We're not sure. Like, did they get an advance warning that on the 28th the US was going to start bombing? Was it just coincidence what was going on? We're not sure. But we did see that massive spike, usually in kinetic conflicts. So what we saw Russia, Ukraine, we saw Israel, Hamas, cyber proceeds and then stays elevated through the entire time of when there's a physical war going on and then often lasts slightly after whenever sort of peace is declared. In this case, it was different. So we saw a massive Spike on the 27th, and on the 28th, it dropped down to less than 10% of baseline.
- [06:05]
Eve Tro
So less than 10% of what had happened, what had happened before.
- [06:09]
Matthew Prince
We think that what had happened was that the US and Israeli strikes were so effective at disrupting command and control inside of Iran that even the hackers were sort of like, we're not sure what to do. Now that's coming back. We've seen a dramatic uptick in Iranian attacks. Stryker, when a big, big Michigan based corporation got, got completely compromised, looks like by Iranians. I think you're going to see a lot more of that because it is one of the things when you have sort of these asymmetric conflicts where you have a very powerful nation taking on a much weaker nation. Cyber is one of the ways that they can strike back. I also think that we'll see a lot of other nation states that don't necessarily want to get blamed for cyber attacks. Russia is the largest offender of this. Russia is extremely vulnerable to cyberattacks themselves. And so I would have predicted that after the Russia Ukraine conflict starts in February of 2022, that we would have seen a massive uptick in the number of Russian cyber attacks that happened. That has not largely happened nearly as much. There have been some, but it's not to the wave that we thought. But part of the reason we think, and again speculation, but part of the reason we think is because Russia itself is so vulnerable to attack that if they directly attack the U.S. they would be, they would, we would come back at them and would shut down their entire power industrial base. A lot, a lot of different, different things. And so there's sort of a mutually assured destruction concept to cyber. Russian attacks really ticked up when Israel, Hamas took off and Russia was doing a lot to try and disguise their cyber attacks as if they were coming from Hamas or Iran. And so I would imagine that we're going to see a lot of, a lot more Russian attacks that will be trying again to look like Iran and
- [07:58]
Eve Tro
the efforts to sort of degrade Iran's capabilities. That doesn't necessarily extend to cyber in the same way it does to other
- [08:08]
Matthew Prince
things or we don't know. One of the internal debates at Cloudflare right now is the Iranian Internet is shut off. And the question is, did they do that to themselves? Because that's what they have historically done. They want to block it. It's a way to quell protests, a way of things. Or did the US do that to them or Israel? Because typically in a, again, kinetic conflict, the first thing that you do is you try to disable the communication systems of whoever you're about to bomb. And so that would be very normal. But on the other hand, the State Department went through Extensive efforts to try and get things like Starlink and other things into Iran. It's not clear entirely whether Iran shut down the Internet themselves or if the US and Israel shut it down. The evidence that they did it to themselves is that there is still from some cellular networks and specifically from some SIM cards, we see access. And what that traffic is accessing is things that you would imagine elites inside of Iran trying to access. So a lot of cryptocurrency exchanges. Because if you're trying to get assets out, that's one of the ways that you would do it. A lot of sort of social media and news trying to see kind of what, what is going on in the world that leans in the favor of they did it to themselves and they've basically whitelisted a certain set of elites. The. The leans against it is the US is really good and Israel are very good at cyber.
- [09:36]
Eve Tro
One of the discussions about AI is sort of for hackers. Like, you want your AI to be better than their AI because they're using one and you're. Is that a factor in this engagement or not? Not as much yet.
- [09:48]
Matthew Prince
AI is making. AI is making everyone more efficient, and that includes hackers. So that the time from when a hacker gains access to a system to how much damage they can do has compressed massively. So that, you know, if in the past it might have taken days from. You get access to one system, how do you get access to the whole company? Now it's taking, you know, hours or minutes in order to do that. There was a tool made by a company called Salesloft called Drift. It got compromised by some Russian hackers, and it gave the Russian hackers access to the vast majority of Salesforce clients. It was not nearly as harmful as possible because the hackers didn't understand Salesforce. And so they were spending several days just trying to meet. I don't know. Not a lot of people understand Salesforce,
- [10:32]
Eve Tro
but sometimes even able at Salesforce, even
- [10:36]
Matthew Prince
people, yeah, like, how do we have so many Salesforce administrators?
- [10:40]
Bob Safian
But.
- [10:41]
Matthew Prince
And this was an attack that happened six months ago. I think what you see today is that AI would allow them to get
- [10:46]
Eve Tro
much smarter, to get up to speed
- [10:48]
Matthew Prince
that much faster, that much faster, and do that much more harm. That's the bad news. The good news is at the end of the day, who wins in the AI race is whoever has most the most data. Like it's. We can talk about chips, we can talk about researchers. Those things will become commodities more and more over time. The thing that will differentiate is Whoever has the most data, and the good guys have more data than the bad guys. At some level, we would have never described ourselves this way, but at some level, Cloudflare has always been an AI company where we take. I mean, the reason we have a free service is because all of that is information that we can find feed into machine learning algorithms to look for what the new security threats are. In the same way that we all. Three and a half years ago, ChatGPT comes out and we're like, whoa, that's amazing. Internally, we'd had those machine learning systems that were finding things that we already knew about and identifying them more quickly. But about three and a half years ago was the first time where our system started to identify threats that no human had identified before. And so I think that there's going to be a bunch of horror stories around AI. There's going to be families that send their life savings to some gang members.
- [11:50]
Eve Tro
Yeah. I mean, AI phishing data is bad.
- [11:54]
PJ Vogt
Yeah.
- [11:54]
Matthew Prince
But like, I think the macro trend is that actually the Internet is going to get. And cyber is going to get a lot better because of AI, Even though you're going to have a whole bunch of headlines that are going to sound very scary.
- [12:08]
Eve Tro
Since the last time we talked, you've become much more vocal about AI layoffs. I will say that I've had other conversations with folks who are sort of like, some of the layoffs you see, it's really just businesses like using AI as an excuse. You see this as more transformational.
- [12:23]
Matthew Prince
Everybody should just be super honest with themselves, which camp they're in. And there are two camps right now. There is a camp. It's like, if we were in the, like, on this floor, there are a bunch of screws. If we were in the business of screwing screws into floors up until about six months ago, we were doing it by hand and we were really good at it. And we hired some of the best hand screwdrivers in the world. And then again, there was a. There was a. There was talk for the last three years that, oh, this new technology is coming and we're going to have, you know, automatic electric screwdrivers and everyone's going to be so much more productive. And we're like, yeah, we tried it. It doesn't sort of work that well. The. But somewhere around November, and anyone who's paying attention felt it, the screwdrivers got really good and the electric screwdrivers became incredibly effective. And we now have people in one camp that are literally a hundred times more productive than they Were before they put down entire floors in a day, it would used to have taken us a month. That's one camp. We have another camp who's like, yes, yes, but I prefer my manual screwdriver.
- [13:39]
Eve Tro
I'm so used to it. It's always worked well for me.
- [13:41]
Matthew Prince
And you told me that if I came here and I used my manual screwdriver that I would have a job forever. And I believed it when I said that. But I can't have a world where one employee is 100 times as productive as the other people are. Like, are you losing sleep over Iran? I'm like, I probably should be, but I'm not. Are you losing sleep over, you know, other things? What I'm losing sleep over right now is that we have a bunch of our team that's using electric screwdrivers. We still have a bunch of our team that thinks that their job is to use manual screwdrivers. And they're complaining that, oh, the electrical screwdrivers make mistakes or they do, you know, do various things, but they're missing the point. And my job has to be, how do I get those folks, they're really good with manual screwdrivers to come along. And what's scary for them because it tends to be, it's not the super senior people, it's not the super junior people, it's the people in the middle who said, you told me this is the game. I've been playing this game really, really well. I've been doing that. How dare you change the rules on me? And if you do change the rules and I adopt them, how do I have any advantage over the intern?
- [14:49]
Eve Tro
Right. Am I going to be good with the electric screwdriver?
- [14:53]
Matthew Prince
I don't know. Everything that I was good at was, I was really good with the manual screwdriver. And that's a super scary thing for people who are sort of in the middle of their career. And what I am deeply worried about, and this isn't Cloudflare thing, this is an industry wide thing. It's not even a tech thing.
- [15:07]
Eve Tro
Yeah, I was going to ask you, like, is this just. I mean, I know it applies to software and coding, but if you're in
- [15:14]
Matthew Prince
another business, legal, It's a funny thing because everyone expects machines to be perfect. These are not perfect machines. These are fallible machines. But it's effectively like you have 100 times the number of employees. Right. And they make mistakes, but it's great. They can check each other and they take a whole bunch of work. That was pretty drudgerous. Writing comments on code in finance, copying information from one spreadsheet to another in legal, you know, summarizing a bunch of documents. I mean, those are all things that humans did. We don't need that anymore. That doesn't mean we need fewer humans. I want we. Cloudflare will have more employees. We will continue to grow the number of employees about the same rate that we have. But what I'm deeply worried about is how do I get the folks who are just worshiping their manual screwdriver to realize that the world.
- [15:57]
Eve Tro
And if I'm someone who, like, you know, I'm using an electric screwdriver sometimes, but not for everything in my job I don't need.
- [16:06]
Matthew Prince
It's totally fine then there's a time and a place. Be honest. If you think that your job is to convince people that electric screwdrivers are bad, you're a dinosaur. And what I'm deeply afraid of is I think that there's a generation of people who are between the ages of about 25 and 40 who there's a real risk they're going to get completely left behind. It's also terrifying for society because let's imagine you've got a whole unemployment rate spikes for people in that age group. That's the people who are going to start to become more politically active. Like the populism we've seen today could get even worse and even crazier. And again, I think a huge part of my job is how do I convince people the ship has sailed. We are not debating whether electric screwdrivers are better than manufacturers.
- [16:50]
Eve Tro
And it's not a question of. You're saying I'm in favor of it? It is. It's the real.
- [16:55]
Matthew Prince
Not just us, it's every single company. And so what we saw out of Oracle, what we saw out of Atlassian, what we saw at Block, everyone is going to do it. And even the companies that are growing like crazy and they might turn around and hire a whole bunch of new people like we. We're hiring a thou over 1,111 interns. It's a number that means something to us. But why we're pairing each of the interns with a. A more senior member of our team because usually the mentorship goes from senior to junior. We think it's going to be completely different this year where we think that our team needs to learn from the people who are sort of the natives at electric screwdrivers in order to understand it. We had an outage last year was really bad. It was basically Because a configuration file went out that was, that was malformed. So what's the solution to that? Well, we now have an employee, effectively that doesn't need to sleep, doesn't need to eat, doesn't take breaks and who's like information biases are completely uncorrelated with the rest of our team and they can check every single configuration change before it goes out. Not to, not to change it or even but to flag it like this, hey, this might be a problem.
- [18:01]
Eve Tro
And this employee is AI, is an agent.
- [18:04]
Matthew Prince
Right. And again, you're not going to hire someone to do that. You couldn't. And they would inherently have biases and all kinds of things. It's the perfect use case for this and everyone's going to implement it. And by the way, the Internet, the cloud, all these things are going to become massively more reliable because we can have something that just checks everything. And There was a 14 day debate internally about whether or not we should do this. What the like, of course we should do that.
- [18:34]
Bob Safian
Matthew is not one to mince his words, which I love. From the Iran cyber war to electric screwdrivers, he doesn't hold back. So how much does he trust the big AI companies? And how much control do the rest of us have in shaping our tech future? We'll talk about that more after the break.
- [18:52]
Eve Tro
Stay with us.
- [18:58]
PJ Vogt
When you have a creative vision, you want total control from the first frame to the final render. Which is why so many creative teams are using LTX2 from Light Tricks. LTX2 is a next generation open source AI engine for audio and video. Built for teams who need production ready quality and total control. It runs end to end creative workflows directly on your own device. With 4K video, synchronized sound and full creative flexibility. No bandwidth limits, no data leaving your machine. Just the freedom to generate high fidelity video with synchronized Audio at roughly 1/8 the cost and 1/7 the rendering time of comparable tools. Start creating at LTX IO model
- [19:48]
Matthew Prince
AI
- [19:49]
PJ Vogt
breakthroughs are quickly reshaping every industry around the world. Healthcare, education, science, entertainment and beyond. But true global scale only happens when the underlying infrastructure can maintain pace with the rapid speed of innovation. As the essential cloud for AI, coreweave is a purpose built cloud made for the real demands of today's most complex AI models. Bringing the pace, performance and partnership. Leading pioneers need to continue pushing boundaries. Ready for anything Ready for AI to learn more about how Core Weave powers the world's best AI, go to coreweave.com ready for anything hi, I'm PJ Vogt
- [20:31]
Deal Advertiser
my podcast, Search Engine has a new two part series for you. Of all the new technologies coming out of AI, the most transformative one might be driverless cars. They're already on the road in 10American cities and they're quickly coming to more. We tell the story of how we got here, the secret team at Google that spent 15 years building what might be the safest vehicle on the road. And we cover the fights brewing in blue cities where unions and politicians are working to keep those cars off the streets. Listen to Search Engine wherever you get your podcasts.
- [21:01]
Bob Safian
Before the break, Cloudflare's Matthew Prince talked about cyber war in Iran and how AI impacts have shifted in just the last few months. Now we talk about whether the big AI companies are motivated to operate in the larger public interest and how changes in the Internet create both challenges and opportunities for industries from media to small business. Let's jump back in.
- [21:32]
Eve Tro
When we talked last time, the news that we were talking about was a new tool that Cloudflare had out about sort of helping to block AI bots from crawling content sites if they didn't want it. In that conversation, you said you sort of believe that the AI companies were trying to do the right thing. Has that perspective moved at all? I mean, we've seen a lot of differentiation lately between anthropic and OpenAI and how they're dealing with the government. Do you feel like anything is moved in that?
- [22:05]
Matthew Prince
I think I'm more convinced that that's the case. 80% of the AI companies use US and so we know them really well. All the big ones use us and they're not stupid and they understand that they are part of an ecosystem and they have to behave inside that ecosystem and they're going to act rationally. And so in the case of media, if you were just giving media away for free, of course they're going to take it. It's not the right answer for everyone. Like for Cloudflare, we don't want to block AI crawlers from crawling our knowledge base. We want the AI systems to be able to know how to use Cloudflare, so we don't block them, we welcome them. But if your business is selling ads or selling subscriptions, you might say, hey, unless you compensate me, you don't get my stuff. And that's rational. And so we're just providing the tools to do that. What we have seen, having now done that, is that the deals that people are getting, whether it's Dot Dash Meredith or Conde Nast or other big media companies, have Gotten significantly better. Now there's still a lot that's there. And the real challenge is it's sort of like a superhero movie where the hero of the last movie becomes the villain of the next. Google is the hero of the last 30 years of the Internet. They financed the entire thing. They built a search engine and then they needed to create content in order to make the search engine valuable. So they built the monetization engine that power the entire Internet. And the business model of the Internet was generate content, drive traffic and then sell things, subscriptions or ads. That's been the business model of the Internet. The problem is that AI breaks that. It doesn't work that way. And so like at some level there has to be a value exchange. It could be monetary, but it could just be ego. A lot of times you create something, you put Google Analytics on it just because you're just like, wow, look how many people are reading it. But that's gone way, way, way down because they're not reading it directly. They're reading the right, they're reading the summary, the AI summary Cliff Notes version of it.
- [24:04]
Eve Tro
And so even which we've all become addicted to.
- [24:06]
Matthew Prince
Yeah. And so it has gone. 18 months ago it was 20 times harder to get traffic from Google than it was 10 years ago. Now it's in 18 months, now it's 50 times harder. And that's the good news. It's 3,500 times harder to get traffic from OpenAI than the Google of old. It's 65,000 times harder to get traffic from Anthropic than the Google. That's changing. And so the value exchange needs to change now. I think it's going to get fixed. But who's the problem today the problem is Google because Google believes that they have a God given right to be able to take all of your content and then they would send you back traffic, but they stopped sending you back traffic. Why did Sam and Elon start OpenAI? Because they were terrified that Google was going to run away with the whole game. And so everything is started effectively as a counterweight to Google. And so when you go to OpenAI, when you go to Anthropic and you say you should pay for content, they're like we will as soon as Google does. And that's the puzzle, that's the trick. If I'm Sundar and I'm running Google, like the smartest thing they could do is actually go and make a market because they have the most money, like
- [25:13]
Eve Tro
they're at a profitable business, it's a cost that has never been part of their business model.
- [25:17]
Matthew Prince
But in the future, AI companies are going to look more like Netflix than some university research lab. The thing that's going to differentiate them, and you all intuitively know this, we see how the underlying model is a commodity. So what differentiates commodity? At the end of the day, it's going to be who has access to unique data. So if I'm sundar, my strategy is to go out and say, I'm going to do semi exclusive. You can't do exclusive because there'll be antitrust problems. But semi exclusive deals with all of the most important content in the world, and that will lock them into being the winner going forward. What they're doing now is they just have unique access. How much more of the Internet does Google see than Microsoft Bing? For every one page that Bing sees, Google sees five. And when you look at those four pages that Bing isn't seeing, there's some of the most valuable content, especially for AI. Things like health data, academic data, a lot of specific local information. Those are the things that are the Most valuable for AI. OpenAI knows how important this is, so they are the second best. But they. For every 3.5 pages Google sees, OpenAI only sees one. So Google has this massive advantage. And when people were saying like, how did Gemini catch up? And we can debate whether it's better
- [26:32]
Eve Tro
or worse, whatever, but it's back to your point about data. It's about data.
- [26:35]
Matthew Prince
It's all about data. And I think what the problem is, if we don't, if we either don't find a way to bring Google down and say that you, in the new world, you have to compete on the same level as everyone else or bring everybody else up. Which actually is kind of interesting because I actually think that the AI companies, what they would pay to support journalists is actually not anything close to what they would pay to catch Google. And so maybe the answer to this, maybe the first move is to say, hey, OpenAI, you want to see those three pages that Google sees every time you only see one, here's what it costs. And by the way, we're going to distribute that out to the content creators,
- [27:10]
Eve Tro
but you have to create a marketplace for that to be able to happen.
- [27:13]
Matthew Prince
Yeah. But if you game theory that out, what happens? Well, all of a sudden when content providers are like, wait a second, why are we giving this all for free to Google? Then Google starts to get restricted and maybe we actually get to a place, I am super optimistic in the media space that we might be on the golden age of content creation and that what will get rewarded in the future is actually much better. Like the business model of generate content, drive traffic, sell ads, I would argue maybe has led to a lot of the problems the world has today. Because what generates traffic, it's whatever sharpest, spicy. Whoever generates the biggest cortisol response generates the most traffic. AI doesn't care about that because all AI cares about is the facts. They want all the facts. And so there's amazing things that journalists have that are incredible resources that the AI companies would pay a fortune for. Like imagine you don't just get the article, but you get all the journalist notes that were behind it right now. And there are things you have to do around protecting sources and that stuff. But you can imagine a world in which you do that, that all of a sudden is this treasure trove that really does start to advance towards a media future. What the AI companies want to pay for, they have a mathematical model of human knowledge. They also then know what the gaps are in human knowledge. They want to fill the gaps. Turns out I want to read the stories about people filling the gaps. The thing I don't want to read yet another story about what crazy thing happened in Washington D.C. yesterday. I want to read the. Wow, I didn't know that. That kind of amazing thing. That's what we all light up for. And that's the thing that's going to get rewarded. My wife and I bought a small local newspaper in our hometown. It's in Park City, Utah. We bought it because it was dying. We think local news is important. What we did not appreciate was it actually might be that local news is the thing that is the most valuable going forward. We might make more this year off licensing our stories to AI companies than we do off digital media, digital advertising. Why? Because Park City, Utah is a place some of you go on vacation. When you go on vacation, you might want to ask your AI what's the hot new restaurant in Park City, Utah? And if you don't have the park record, then you don't know the answer.
- [29:17]
Eve Tro
So it's more valuable information, but to a smaller group of people.
- [29:21]
Matthew Prince
Maybe to a smaller group of people, but.
- [29:23]
Eve Tro
But it doesn't matter because it's. It's valuable.
- [29:26]
Matthew Prince
It's valuable. And so I think that it's, it's those things. Like, I would love it if the New York Times started reviewing hotel rooms. Is it better to stay in room 1427 or 1429 in the Marriott Marquis? And that would actually be super valuable information. Those sorts of little hyper local things are way more interesting than just telling the same story. A slightly different. Ben, the thing which is terribly unfair to the New York Times that I keep saying, if you don't license the New York Times as an AI company, just license the Wall Street Journal and then have your AI rewrite it as if it's a New York liberal and you got the New York Times right. Unfair. But there's some version of that. And it's the reason why Reddit, the public Numbers Reddit got 7 times for the same corporate, same number of tokens, 7 times the amount that the New York Times did. Why? Because if you don't have Reddit, you don't have Reddit. There's no substitute, whereas there's lots of substitutes for that. I think the media of the future is how do I create information and knowledge that there isn't a substitute for?
- [30:24]
Eve Tro
So I want to ask you a little bit about like, the value of brand for you. How important is the brand of Cloudflare? I mean, I mean, you've been, you know, talked about the most important Internet company that you never heard of, right? It's like backhanded compliment. But like, for you, how do you
- [30:40]
Matthew Prince
think about what the brand outages from time to time just so that people appreciate how much. It's funny, we have an outage, our stock goes up. It's like, it's very weird.
- [30:53]
Eve Tro
But seriously though, how do you think about what the brand of Cloudflare should be and how important is that in the value that you have in the future?
- [31:03]
Matthew Prince
So first of all, let's talk about what brands are. Brands are just shortcuts for humans to understand value and quality, right? You buy Nike shoes and you have an expectation of what the value and quality is like. If I say, what's it like to walk into a Walmart? We all know exactly what that experience is like because we've all had it and the brand stands for that. What's interesting is I'm not sure any of those things matter in a world of agent and commerce. Like what a brand is, is going to be radically different and someone has to invent it. Because if we don't invent what that is, then then the agents are all going to be this massive force of consolidation because.
- [31:39]
Eve Tro
Because all they're going to do is go to whatever they don't care about the brand.
- [31:43]
Matthew Prince
They don't. They're going to go to whatever they think. Are they going to be able to get whatever you want done the most, in the most efficient way that will tend towards bigger consolidated place. And so I like, I'm actually optimistic about media. I am terrified for small business. Why do you shop at the small business you shop at? It is typically because it's either physically convenient for you or because you have some emotional connection to them. Your agent doesn't give a shit about either of those things. I think the most interesting question in the next five years is what's the future business model of the Internet? I think that a sub question of that is what is a brand in a world of agentic commerce? Let's imagine that you could aggregate here are all the things that people bought, here's how many of them were returned, here's the customer service response rate, you know, packaged up into some cryptographically verifiable thing where an agent can come in and say, hey, I've never heard of this brand before, but I can see that customers are super happy shopping from it and then it rewards that. If we don't have that again. I worry that where we're headed towards is a world where we're actually going to crush small businesses. We're trying to work with like the visas and PayPals and Shopify's and everyone else in order to say, how do we get all the signal back in order to be able to identify that for Cloudflare, I mean, I don't know. We don't think about our, we're terrible. Like you've never seen. We don't put billboards up. We don't, we're, we're terrible at all of those, those things. We're throwing a party tonight and then no one even invited me. So that's how bad we are right now. What I tend to think about more is that we're fundamentally in the business of trust. And like what we ask people to do is kind of crazy. It's like route all of your traffic through us and trust us, we'll handle it well. So we have to just be very much aligned on how do we make sure that we live up to our mission. And our mission is to help build a better Internet. There are so many times like someone will come in and say, if you just did X, we'll give you $100 million. Why didn't we ever get in the advertising business? It could have been huge. Just sync Cookie. It's, we could have sold it, but at the end of the day it
- [33:52]
Eve Tro
doesn't make a better Internet.
- [33:53]
Matthew Prince
It doesn't help build a better Internet.
- [33:54]
Eve Tro
We're just about out of time, Matthew. But we've talked about a lot of things that maybe make people a little nervous about the future, but the future's
- [34:04]
Matthew Prince
gonna be so much better.
- [34:05]
Eve Tro
That's what I'm saying. You are optimistic. So what is the opportunity that people aren't seeing? What makes you optimistic?
- [34:12]
Matthew Prince
You, as an individual, on your own, can build something that will completely change the world and generate generational wealth for you. The ability for a small team to do just absolutely heroic things. It's never been possible before. There is a world in which, you know, we only have one or two AI companies where journalists, academics, researchers get crushed or go work for one of those big AI companies, which is sort of the black mirror version of it. But there's another version where we create a world where anyone can start an AI company, where we make the resources as easy as possible, where everyone can start to create content. And what wins isn't who pisses people off the most, but who wins is who furthers human knowledge. Like, if I were the AI companies, I'd be starting the Academy Awards of content. Like, they should give a reward every single year for who advances knowledge the most. And they can measure it mathematically in different fields. Right? And just celebrate that. And we can create a world where we actually then make it so that if you are a great entrepreneur and you build a better mousetrap that everyone can find you and you can be successful, and that you don't get crushed because somebody's got a bigger marketing budget than you are, because you can actually demonstrate, truthfully, not by some advertising shortcut, but truthfully, that your product is better. That's the world we should be building towards. And I don't know how to do it, but I think just articulating that, we want lots of AI companies, we want lots of content creators, we want lots of businesses large and small competing in a fair market. And that's what we should be playing for and asking ourselves what we are doing technologically, what we're doing from a policy perspective, what we're doing from a regulatory perspective, what we're doing as consumers is are we moving towards that future or away from that future? The next five years, we're going to figure out what the future business model of the Internet is. And all of you have the opportunity to create that. Imagine how amazing that is. It's going to completely change. And you don't ever see systems as big and complicated as that have these massive disruptive changes. And so, again, if you think your job is to argue for manual screwdrivers, you're not going to have fun over the next 20 years of your life. It's going to be hard. If, on the other hand, you're like, listen, I'm going to be open, I'm going to be curious, I'm going to try these things and I'm going to say that's the direction of the future. But let's make it as optimistic and positive as possible. Again, I am 100% confident that tomorrow is going to be better than yesterday.
- [36:47]
Eve Tro
Well, Matthew, thank you so much.
- [36:54]
Bob Safian
I learned something new every time I talk to Matthew about what's going on behind the SC scenes in the tech world. I don't always agree with all of his predictions. The future of journalism that he sees doesn't quite fit my ideal, but he's always thought provoking. One clear takeaway, echoed by my chats with other CEOs, is the changing nature of the workplace. Executives believe that electric screwdrivers are taking over, and they're making decisions with that assumption in mind. Even for those of us who never use a manual screwdriver and who aren't even trying to screw in screws, the expectations are shifting. Will that be good or bad? As Matthew acknowledges, that in part depends on how we respond. One thing seems pretty certain to me, though. It's going to be a bumpy ride for everyone. I'm Bob Safian. Thanks for listening. Rapid Response is a Wait.
- [37:53]
Eve Tro
What?
- [37:53]
Bob Safian
Original I'm Bob Safian. Our executive producer is Eve Tro. Our producer is Alex Morris. Associate producer is Mashumaku Tonina. Mixing and mastering by Aaron Bastinelli. Our theme music is by Ryan Holiday. Our head of podcasts is Lital Milad. For more, visit rapidresponseshow.com.