AI doesn’t just have an ethics problem—it has an incentives problem. In this episode of Ship Happens, Ruben Verborgh joins Per Krogslund to unpack why today’s AI ecosystem is economically unsustainable, how decentralized web standards could reshape ownership of personal data, and why smaller, controllable AI systems may outperform massive black-box models in the long run.
In this Ship Happens episode, host Per Krogslund interviews Ruben Verborgh, a computer science professor at Ghent University and consultant about building a more responsible digital ecosystem where incentives align with ethics rather than exploitation.
Ruben argues that today’s AI ecosystem is fundamentally unstable: companies often lack control over the models they deploy, sustainable business models remain unclear, and the industry continues relying on advertising-driven incentives that reward data extraction over trust. Instead of focusing purely on “ethical AI” messaging, Ruben says the real challenge is designing better economic systems and incentives.
The conversation explores an alternative future for AI and personal data—one where raw user data stays under individual control and algorithms move to the data instead of endlessly collecting and centralizing it. Ruben explains how decentralized web standards like Solid aim to separate identity and storage from applications, giving users more portability, ownership, and freedom online.
Per and Ruben also discuss mistrust in AI agents, why smaller task-specific models may ultimately be safer and more controllable than giant general-purpose systems, and why society must work alongside younger generations to shape technology rather than attempting to ban it outright.
(00:00) Welcome and Guest Intro
(00:32) Mission for Better Data
(01:24) Incentives and AI Ethics
(02:54) Packaging AI Like Utilities
(05:14) Economic Model Will Implode
(07:10) Beyond Ads and Big Data
(08:36) Bringing the Algorithm to the Data
(10:36) Consent Beats Tracking
(11:45) The Myth of Retargeting Value
(15:10) Trust and AI Agents
(16:55) Guarantees and Smaller AI Models
(18:03) Solid and the Decentralized Web
(20:51) Why Identity Came Later
(23:04) The Web as an Amplifier
(26:22) Fix Incentives, Not Algorithms
(27:53) There Is No “Responsible Tech,” Only Systems
(29:06) Policy Shifts Beyond Finger Wagging
(32:04) Believing in the Next Generation
(33:11) Wrap Up and Outro
Ruben Verborgh is a computer science professor at Ghent University and a researcher focused on decentralized web technologies, linked data, and data ownership. He works with Tim Berners-Lee’s Inrupt on the Solid project, which aims to separate applications from personal data storage to give users greater control, privacy, and interoperability across the web.
[00:00:00] Ruben Verborgh: I use Gemini every week, and at least once a week it will leak its own prompt to me, and I didn't even ask for it. So what you clearly see is that we've got a technology that is so strong that companies themselves are not in control of it. And it's not a solution to say like, "Oh, let's just, you know, do it a bit better."
[00:00:18] Ruben Verborgh: No, no, like fundamentally the way we package AI, the way we package personal data, that's what's wrong. And I mean, just to say it very simply, like electricity is also not packaged as like, um, uranium that we just hand out to everybody, right? I cannot believe for one moment that any returns you get are higher than all the investments in all the infrastructure, sucking up the data, running the algorithms.
[00:00:40] Ruben Verborgh: I will tell you how to make money. And the fact... And I know this for a fact because my supermarket still thinks I have a dog, only because my kid once wanted a squeaky toy, which was an animal's toy, not a kid toy. Ruben, you're a computer scientist. Like, tell us, tell the world, what do we need to do to change the algorithm such that instead of promoting garbage, it promote good things?
[00:01:01] Ruben Verborgh: Like, what is the magic sauce? And I'm just like, "But we know the answer. Like, actually, like we can optimize the algorithm for anything you want. What we don't know the answer to is how do you still make money if we show things that make people happy?"
[00:01:12] Per Krogslund: Welcome to Ship Happens. It's a podcast where I sit down with really smart people and talking about how you can ship great software.
[00:01:18] Per Krogslund: Uh, today is a little bit of a, an exception because today I am meeting with Ruben Verboord. He's a professor, and he knows a lot about responsible AI. He has strong opinions about decentralized data. Uh, you work together with Tim Berners-Lee and, um, as said before, you were in academia as well. So welcome to the show.
[00:01:48] Per Krogslund: We kind of met together, like, 20 years ago in an open source project, so it's just, it's lovely to catch up and know what you're up to. So I think, first of all, give me just, like, a little bit of an introduction to the people who don't know you. I'm
[00:01:59] Ruben Verborgh: a professor of computer science at, at Ghent University.
[00:02:01] Ruben Verborgh: Uh, so I'm, I'm a, I'm an academic, you know, the ivory tower type. At the same time, I try to do things that I like, that are meaningful, but also things that, um, connect to the real world. Yeah, so I'm also a consultant to the industry. In particular, I've, um, I'm working with Inrupt, which is Tim Berners-Lee's company that thinks on how can we, um, develop a new data ecosystem where, um, ethics, um, economics don't have to be contradiction.
[00:02:29] Ruben Verborgh: Um, and my overall mission is to make data flow better and better, as in, in more responsible ways, in ways such that everybody benefit. Not just people, not just companies, but, like, everybody. There's enough data to have a win-win, win-win, whatever, and I like to support making that happen.
[00:02:47] Per Krogslund: Do you think that's possible, really, when, when you have...
[00:02:50] Per Krogslund: You can say you have... Let's, let's talk about AI a little bit and about this, like, win-win. Like, you have these companies that's funded by billions and billions and billions and billions. Do you think they have any interest in making a win-win-win? Don't they just have an interest in making a win for themselves?
[00:03:06] Ruben Verborgh: Of course, everybody should think about themselves, but that's the cornerstone of the economy. Uh, like if you go back all the way to, uh, old Adam Smith, like in the 1700s, he was like, "Look, the baker doesn't give me bread because he likes me. Um, like it benefits him to give me bread, uh, because he gets money in return, and so and I'd rather have the bread than the money," right?
[00:03:25] Ruben Verborgh: So- Mm ... the whole economy is driven by self-interest, and that's a good thing, uh, if the economy's healthy. There's some other indicators. Uh, but the economy doesn't depend on benevolence, and I think that's a problem today. Um, in much of the communication, we look at the big companies, uh, wagging our finger like, "You're not supposed to do this and, and please behave ethically."
[00:03:44] Ruben Verborgh: But the fact that it is even a question, that in a position to be asked this and to decide on the answer is wrong. Not the answer that they picked, because every company has a duty towards their shareholders to, to turn a profit. That, that's what they have to do. It's just that some of the rules of the game apparently are set in a wrong way such that the, um, self-centered behavior doesn't result in the optimum that we see in other places of other parts of the economy.
[00:04:09] Per Krogslund: So you think like the AI companies are basically structurally incapable of building a responsible AI if they don't have an incentive to do so?
[00:04:16] Ruben Verborgh: The way we're building it today is completely irresponsible for many ways. But, um, what is very striking is that you see that the companies themselves also have lost control.
[00:04:27] Ruben Verborgh: Like make no mistake, like yes, big AI is harming certain people. That's also the companies themselves. Like one of my favorite examples is, um, Google a couple of months ago was talking about how Gemini is being attacked. Like people are trying to reverse engineer internal prompts by... But like, dude, I use Gemini every week, and at least once a week it will leak its own prompt to me, and I didn't even ask for it.
[00:04:50] Ruben Verborgh: So what you clearly see is that we've got a technology that is so strong that companies themselves are not in control of it. And it's not a solution to say like, oh, let's just, you know, do it a bit better. No, no, like fundamentally the way we package AI, the way we package personal data, that's what's wrong.
[00:05:08] Ruben Verborgh: And I mean, just to say it very simply, like electricity is also not packaged as like, um, uranium that we just hand out to everybody, right? No, no. We package it in formulas, stay for mass distribution, and now we all benefit from it. AI, the way we see it packaged today is completely irresponsible for everyone.
[00:05:24] Ruben Verborgh: Uh, because there's the fact that they don't have it under control, but also make no mistake, they don't know how to make money with it. Uh, like nobody knows how to make money, and that's a real problem because that's where you see the shortcuts. So what we see today is shortcuts of companies playing a shorter term game because we haven't cracked the economic puzzle.
[00:05:44] Ruben Verborgh: So by all means, I want it to be just a very simple economy and not whatever it is we're trying to do today because it's not sustainable. We need to do something with this immense power of AI, but the way we package it is it's not working for anyone really.
[00:05:58] Per Krogslund: Hmm. So when you say about packaging, we are, we are, right now we are packaging it heavily subsidized for easy consumption.
[00:06:05] Per Krogslund: It's available everywhere. Um, and then, yeah, the companies behind it doesn't know how to make money out of it, which I guess is a problem when there's like billions and billions of this invested in it. Um, do you think we are heading towards like a, a collapse here where we are kind of getting people used to like AI is practically free to use and then at some point it's either gonna be like the prices are gonna change heavily to make this like sustainable or do you think this is just gonna collapse because everyone loses faith in the, the financial like structure of AI companies?
[00:06:37] Ruben Verborgh: I think what's gonna collapse is the current economic model that we try to impose on data and AI. It is clearly not sustainable, and companies know this. Uh, so I think they very openly are saying, "You know what? We don't know, so we're gonna play a much more shorter term game," which fair enough. Uh, and it's good if you wanna make a lot of money.
[00:06:56] Ruben Verborgh: Just I don't find it interesting because, like, they haven't invented a business model. They have just found a way to make a lot of money in the short term. But what they haven't cracked is the puzzle, how do we keep on doing this? Uh, so, and when I say electricity and other utilities, it's, it's not a coincidence.
[00:07:13] Ruben Verborgh: Like, I bet it was the exact same when electricity was discovered. Like, clearly it's something important. It's a, it's a big force we need to do something with it. The fact that it was not just like companies selling batteries or uranium or boxes of coal, so to speak, but no, no, we need a distribution network for it.
[00:07:28] Ruben Verborgh: I, I think that, that one was a game changer. Yeah.
[00:07:31] Per Krogslund: Mm.
[00:07:31] Ruben Verborgh: Uh, so, so it gonna implode in a sense that, yeah, the, the models won't work out. And if you look, another thing, the OpenAI, um, predictions, for instance, like, yeah, we're gonna lose billions this year and the next, and so on. And then suddenly, I believe in, in 2029, after years of billions of loss, 100 billion of profit just a single year.
[00:07:50] Ruben Verborgh: Just magic. Um, well, if they believe that I'm happy, if it was hallucinated, I'm even happier because at least it wasn't a human dreaming this up. But it ain't gonna happen. That doesn't mean that any of those companies are doomed. Like, clearly this is a very powerful thing, and knowing, having a lot of knowledge about this powerful thing will be beneficial.
[00:08:09] Ruben Verborgh: But the packaging is completely, completely gonna change because the way it's packaged right now, they're not gonna make money. It's like, you know, it's... The way we deal with AI is like if we're just, like, shipping uranium instead of, like, having sockets for electricity.
[00:08:22] Per Krogslund: Mm.
[00:08:22] Ruben Verborgh: Not a wise idea, but you can make some money in the short term, sure, because look, it glows and, uh-
[00:08:26] Per Krogslund: So we're kind of rewarding, like, irresponsible behavior in these markets.
[00:08:29] Per Krogslund: But what, what, what is the alternative then? What's your idea of this?
[00:08:32] Ruben Verborgh: Ultimately, we're still stuck in advertising really. Yeah. So, so what we see... So basically after big data came, came big AI, so to speak. Uh, and, and big data was never really resolved. Like, we haven't found a sustainable way because the problem with data is that you can keep on copying it, right?
[00:08:51] Ruben Verborgh: So it's, if you have it good and I have it, like, there's no rivalry between us. Like, we, we can both have it, and a third party can have it. So basically with data, it's about making sure no- nobody else has it. And in a way, AI is simply liquid data because that's what it does. Like, an AI model by itself is only as powerful as the new energy you inject into it, right?
[00:09:12] Ruben Verborgh: So any algorithm, be it generative AI or whatever, um, you need to feed it data to get something out of it. Uh, so the whole shift towards, um, AI now is, on one hand, facing the same problem because of the heavy data dependency, even bigger than with big data, I would say. And interestingly, we're also going towards the same solutions like, "Oh, you know what?
[00:09:35] Ruben Verborgh: Advertising. We just gonna start advertising." As if the answers of LLMs are not unreliable enough already. So that's the issue, like this high dependency on data, and it's not a fallacy that, you know, more data will solve this. But it's only gonna stop when companies and every single company realize that they'll never have enough data.
[00:09:55] Ruben Verborgh: There's no such thing as having enough data. So the core of the issue is that personal data is a dangerous stream, a toxic stream because we send it into the world and it's dangerous because it's true AI is a very dangerous tree. Now we just send them to the world, it's dangerous because it's not true or maybe true, who know?
[00:10:16] Ruben Verborgh: And to get results, what are we gonna do? Dangerous personal data send them to the world all the way over to a model, and then they sit there. No, that, that... The solution is algorithm to the data. And I know technology-wise, that's, that's a known thing, but reason is not technology. Get economy ready, huh?
[00:10:35] Ruben Verborgh: Because raw data is like stem cells basically. Like, like I cannot... When I give raw data to you, I cannot imagine all the things that might happen to it. You cannot, nobody can imagine what can happen to it. What, what are all the things that you could do with my date of birth or the fact that I like curly fries to use one of- Mm
[00:10:53] Ruben Verborgh: the classical examples, huh? I don't, by the way. But, um, I cannot imagine that, so I cannot, but we can derive from it. If you bring- Yes ... an algorithm to the data, and if you say things like, "Actually, Ruben is willing to pay a premium for, for this product," or, "Ruben is allowed to buy a bottle of whiskey because, uh, he's of a legal drinking age," and things like, these are raw.
[00:11:16] Ruben Verborgh: And from raw data, we go to insights, okay, which, which are valuable at the moment. Okay? So, and all of this is possible because I have much more data than you will ever have about me. And like, no matter how much you try to get from me, there's legal barriers. It's just expensive to acquire- Mm ... to maintain and so on.
[00:11:34] Ruben Verborgh: So you don't really want raw data, just like you don't want raw uranium and oil and... No, no, no, no. You want the insight you need at the moment. So- Yeah ... I invented an architecture and economy in which, uh, what companies really want is the value from, from the data, but the value changes at every single point because the real value is in data they'll never have, and they'll never have enough.
[00:11:58] Per Krogslund: What you're describing here is really... You're, you're modifying people being consenting adults handing over their data as like in a certain, like negotiation rather than just being passively extracted from by various trackers that's trying to deduct a lot of things about their behavior and so on.
[00:12:15] Ruben Verborgh: Yes.
[00:12:15] Ruben Verborgh: And you know why? Because companies are absolutely terrible at extracting data. In Europe- ... with the GDPR, it's medical data. They don't- Okay ... wanna know, but I'm like, uh, you spend millions trying to guess whether I'm a dog, um, but I'm pregnant on a diet, and I will bloody tell you how to make money. I will tell you that I'm gonna pay more for products with less sugar.
[00:12:37] Ruben Verborgh: It's as simple as that. And the reason this example triggers them is because they know very well it's data they'll never have. Yeah. So I'm trying to promise them the value from data they'll never have. But then the question is, how do we make it happen? Li- like, I, I wanna tell them, and I have incentives to know because it helps me.
[00:12:56] Ruben Verborgh: They have incentive because those products are more expensive. So why are we bothering with all the data you're trying to guess, whereas the most valuable data you're not even willing to accept? What are we doing?
[00:13:05] Per Krogslund: Right?
[00:13:05] Ruben Verborgh: So we need a solution for that.
[00:13:07] Per Krogslund: So I think, uh, a anecdote to my, my life in, in tech is, uh, I worked for, like, I worked for Zalando at some point, and, uh, we did, like, this GDPR, like, policy rollout.
[00:13:17] Per Krogslund: And, uh, where we really... We, we cut off so much data from... And we also cut off, like, sharing a lot of data with, like, Facebook and all these other, like, platforms and so on. And I think, I can't remember... Well, I can't remember the exact percentage, but I wanted to share it. But there was, like, a fairly high percentage of data that we lost out on, like, sharing with all these, like, social platforms that sucked in the data and ac- worked as a, as a broker, basically.
[00:13:39] Per Krogslund: But interestingly enough, like, we didn't see any negative impact on, on sales or, like, people coming back to look at that thing. You know how, you know how retargeting works? So you go and look at a shoe, and then the, the ads for that damn shoe is gonna follow you around on the internet forever. Um, like, because that data was not shared, or at least, like, targeting, uh, algorithms, that, that stopped happening.
[00:14:02] Per Krogslund: But, but, but people didn't stop going back for the shoe. Like, like, there was, there was zero change in actual customer behavior. So all this, like, data that's shared with these, like, enormous platforms, like, it didn't, on the surface at least, seem to make a difference. So I think, I think the industry as a whole is, like, grossly overestimating the value of these, like, very simple signals of, like, person looks at shoes, therefore they must want this shoe, or they, they have to be reminded constantly to buy it
[00:14:30] Ruben Verborgh: Yeah, exactly.
[00:14:31] Ruben Verborgh: And there's almost a certain technological arrogance there as well. Like, and, and it's also a paradox because on the one hand we're saying like, "Oh, you need more and more and more data. Like, you can never have enough data. Keep on collecting." Hmm. And at the same time, with the small data set they have about me, which is nothing compared to data that I have about myself, they think they know me better than I know myself, but that's impossible.
[00:14:52] Ruben Verborgh: Of course, with all the data that I have that you'll never have because it costs too much or you're not even allowed to have it, of course, if the insights start being derived from my side, you'll have much more. And, and, and that's kind of the, like, if you go back to basic economy, like, like we have almost forgotten that we're not enemies.
[00:15:09] Ruben Verborgh: Like back in the day, like, like the first supermarkets, no, like, like they're not enemies. The supermarket's in competition with the others and the butcher and...
[00:15:17] Per Krogslund: Hmm.
[00:15:18] Ruben Verborgh: But we're trying to do each other a favor. Like, I give you money, you give me vegetables. That, that simple as that, and you make it really easy to do this transaction.
[00:15:26] Ruben Verborgh: But suddenly when, you know, the customer loyalty cards came along and then when it was about data, we start believing that we're in a fight. Now you're trying to steal my data, I'm trying to... No. I will tell you what I want. I will tell you how you can make money by giving me the stuff that I want. That's a transaction.
[00:15:42] Ruben Verborgh: So help me help you. But at what point did they think that they, they can predict better what, what I want? And in fact, it's, it's counterproductive because the economic force of supply and demand have to play. Hmm. And if you take some of the thinking or guesswork out of the equation, if you move to the other side, then the forces do not play.
[00:16:01] Ruben Verborgh: And maybe in the short term you can think you can do something cool, but I cannot believe for one moment that any returns you get are higher than all of the investments in all the infrastructure, sucking up the data, running the algorithms. I will tell you how to make money. And the fact, I, I know this for a fact because my supermarket still thinks I have a dog, only because my kid once wanted a squeaky toy, which was an animal toy, not a kid toy.
[00:16:28] Ruben Verborgh: And like, I will tell you, come on. It doesn't work.
[00:16:32] Per Krogslund: Yeah, I guess that, that is the, like, fundamental trust problem right now that would you trust an OpenAI or like a Google Gemini agent to, to strictly work in your interest?
[00:16:41] Ruben Verborgh: I mean, OpenAI doesn't trust OpenAI for its agent to work in their own interest.
[00:16:45] Ruben Verborgh: Like, we don't have a technological... Even, even if they wanted to, like the agents do whatever. Like, like they- Yeah ... they kill people. They literally kill people and they say the stupidest things about OpenAI itself and, and, and, and, and so, so what an arrogance. Like, so we, we cannot even do that. So even if you wanted to, it is not possible.
[00:17:01] Ruben Verborgh: So no, nobody, nobody trusts any company to make AI that we can trust because it's impossible. And it's also, I mean- The reason that I cannot trust an agent is because the agent cannot trust me back, as simple as that. Yeah. Like, trust is a human thing. What has changed, of course, is like back in the '60s, the number of trust relationships was relatively low.
[00:17:26] Ruben Verborgh: Like there's your family, friends, a baker, a butcher, a brewer. That's about it. With the worldwide web, the number of relationships we have to maintain, uh, as, as we see, we see it, it increases time and time and time again. That's not gonna change. But still, relationships are a human thing, but technology can help us in supporting those relationships.
[00:17:50] Ruben Verborgh: Uh, so make sure that my interests are always guarded in every relationship. But developing a trust relationship with technology is not possible because they cannot trust me back.
[00:18:01] Per Krogslund: Yeah.
[00:18:01] Ruben Verborgh: So the question by itself, um, leads us to problems that cannot be answered meaningfully by technology.
[00:18:07] Per Krogslund: So you're saying when, when, let's say, my assistant goes and negotiate with your AI assistant or with Walmart's AI assistant, there i- there is not a, there is not a concept of trust between them?
[00:18:17] Ruben Verborgh: Let, let's take it from the start. So suppose that we don't have AI agents. Uh, we, we got people, so we got like I got a person negotiating for me and, and your Walmart got a person negotiating for you. Um- Yeah ... I think that they, they do a pretty good job, and they don't have to trust each other necessarily.
[00:18:31] Ruben Verborgh: Like, by just acting out of our own interest, they can find something that, that we need. And so it's more like rather than asking is there trust, it's like- That's not a precondition, uh, um, it's more about what kind of guarantees that you can offer. So in this case, I'm asking for the guarantee that my assistant only asks, I only, um, acts rather in, um, my best interest, huh?
[00:18:52] Ruben Verborgh: So that needs to be hard guaranteed, testable, verifiable, and so on and so forth, huh? Um, and that's very hard with big and complex models because nobody really knows what they do apparently, huh? So we need, uh, specific models, small models there to, to certain tasks like so. I don't expect to have the one agent that knows everything about me.
[00:19:12] Ruben Verborgh: It's not gonna happen. That's, that's not how the world works, eh? I think dedicated agents for specific things with... But the most important thing is that you can control it, uh, so that you have the tools to say more of this, less of that. You know what, I'm gonna do it myself.
[00:19:25] Per Krogslund: I actually wanna step a little bit back because now we talked about data and we talked about...
[00:19:28] Per Krogslund: But also I'm actually curious, you had, um, you worked with Tim Berners-Lee at some point about a project called Solid, about decentralized data. Whatever happened to that?
[00:19:37] Ruben Verborgh: Solid is one of the next evolution of the web, like one of the, the next steps that we need to do. Because if you look at the way the web was designed, it started out as an information system.
[00:19:47] Ruben Verborgh: Mm-hmm. But, but then you see that people are using it in, in very different ways than, than you anticipate.
[00:19:52] Per Krogslund: Yeah.
[00:19:53] Ruben Verborgh: Like what started as a hypertext system for scientists to share information with the world suddenly was also used for private information, um, to order pizzas online, to, um, start entirely new projects nobody had anticipated, um, like eBay and, and Facebook and, and like completely novel concept.
[00:20:13] Ruben Verborgh: At the same time, you see that some of the basic principles then, then are ins- insufficient. Uh, for instance, um, one of the major decisions, uh, is like, well, when the web started out, we're not gonna have identity. And as a consequence of which, every single website has their own login. As a login that is not part of your browser, it's part of the website.
[00:20:35] Ruben Verborgh: Whereas, um, other things like change font size, for instance, is part of the, of the browser. And, and so, so, so what Solid does is re-examining which standards that we need. And it's a couple of, of- Things that are, are, are lacking, uh, given what web's actually used for. One of them is identity, so the fact that my browser can, uh, manage who I am and so on.
[00:20:56] Ruben Verborgh: And if a browser does that, not only would it, um, save us a lot of effort always signing up and remembering passwords, but also in terms of security and data breaches and so on. Uh, so completely changes, um, what you can do with the interface. Um, and another aspect of course is that the web for years and years and years hasn't consisted of just hypertext, of course, like there's images and video and also data and, and, and lots of data basically.
[00:21:24] Ruben Verborgh: And every website does identity and data in, in different ways. Uh, so Solid is a set of technologies built on the existing web to structure data and identities in a way that makes it easier for people to do stuff. Very simply said, um, today, uh, the way we use the web, it, it's application storage are, are, are tied together.
[00:21:45] Ruben Verborgh: Uh, so if I, um, store my document on Google Drive, I have to edit it with Google Docs. I cannot, um, share it with you and then you use Office 365 or whatever it is that you... No, no, we, we have to tie those together, uh, because of the way the web works. So if we separate this, we, we, we gain choices. If I can decide where I store my data independently of the application that use it, um, there's many more cool things we can do.
[00:22:10] Ruben Verborgh: That was one of the starting point of it.
[00:22:12] Per Krogslund: Okay. So really separating the, the person doing something, the application they're using, and where they decide to store the data. Why, why, why wasn't just, that just the default choice from the beginning, that we had a stronger sense of like you can say data and identity sovereignty?
[00:22:25] Ruben Verborgh: Well, the beautiful thing is all of this is documented. Like, if you go back to one of the oldest web pages, which is like Tim Berners-Lee, like, um, the proposal of, of, of how the web... Well, like you see the thinking and, and he didn't say like, "Well, identities are an important pri-" No, no. He said, "Look- We need to make a choice here.
[00:22:43] Ruben Verborgh: Like, we're designing a system that is intended to help spread knowledge to different scientists, because the web started out of the frustration that we had all those scientists at, at CERN, at the European, um, Center for Nuclear Energy, and lots of knowledge remains stuck on machines. So the primary driver was like, how do we make sure that the knowledge is accessible?
[00:23:02] Ruben Verborgh: And other concerns like videos and images and identity were... They were all seen, but like, "This is not important right now. We can add it later." Yeah. But what happened then, this is the beautiful thing, of course, is that like the web became so successful that people started using it for many other things that it wasn't designed for.
[00:23:22] Ruben Verborgh: So to Tim's credit, it's not like, like, um, you can tell like, "Oh, he forgot to think about it." No, no, he, he actually did think about it. It's like-
[00:23:29] Per Krogslund: Okay ...
[00:23:30] Ruben Verborgh: this is not a problem right now. Because there are many competitors or competing technologies to the World Wide Web, most of which are much more complex.
[00:23:37] Ruben Verborgh: They're trying to, are trying to solve, uh, more complex problems. But the web apparently found the right mix between utility and simplicity that it conquered the entire world. Yeah. So of course, you never know what history would have looked like, but it's not hard to imagine given some of the examples that we have that like a system with, that required much more implementation effort, like, well, you need to have identity, you need, uh...
[00:24:00] Ruben Verborgh: Would have spread much more slowly and simply would not have made it.
[00:24:04] Per Krogslund: Yeah. At
[00:24:05] Ruben Verborgh: the very worst case, we wouldn't even have had an alternative. Uh, so, so bit of a victim of its own success, uh, I suppose, but at the same time, it doesn't stop there, uh, because it is possible to keep on adding new, new layers. Uh, but, but- Yeah
[00:24:18] Ruben Verborgh: back then, who knew that, that we would use the web to order pizzas and, and, and, and look into space?
[00:24:25] Per Krogslund: Fair. Very fair. Um, so if we go to the original purpose of the internet of, like, sharing and distributing information, do you feel like with, with how the internet is evolving now, do you feel like that original purpose is kind of slowly going away?
[00:24:39] Per Krogslund: For, I'm not saying for, for- Well, I think- ... scientists or, but, but more like for the, the common person just using the internet through Google or whatever, like, you know.
[00:24:46] Ruben Verborgh: I still think it's one of the best things to happen to humanity ever, because, like- Yeah ... I'm a very curious person myself. And, and I think it's so important that knowledge spread because...
[00:24:55] Ruben Verborgh: And even with data, make no mistake. Uh, like, everything that we see today, the problem is not that, that, that, that, that our data's flowing too easily and all this. No, no. The problem is the opposite. It's so incredibly hard to move data in the right ways that all that we see are the shortcuts and all the ways you're not supposed to do it.
[00:25:12] Ruben Verborgh: So intrinsically, the fundamental problem is always that it's super hard to move information. That is the hardest thing. Um, and, and, and all the rest is, is, is, yeah, they're important problems, but getting data moving, getting information flowing is, is, is the important bit. Um, so the world has become a lot smarter for it, but the wo- the web is an amplifier, of course.
[00:25:36] Ruben Verborgh: Like, so it amplifies the good and the bad of humanity. And as long as we believe, and I do, that there's more good, then the amplifier does its work, right? But we have to make sure, and then, uh, to a real point, that, um, the amplifier is, is linear in the sense that, that we don't see, um, a high concentration amplification of, of, of bad content.
[00:25:57] Ruben Verborgh: And now if you switch- Mm ... to something completely different, but, but social media. Yeah, there's a big thing about the irresponsible youth today and, and so... But, but have you ever tried starting a new account, like a blank account on any social media platform? Like, apparently it's so easy for bad things to, to go in.
[00:26:14] Ruben Verborgh: And if you go back to the utility thinking work, like, look, I've got, uh, water. Yeah, I've got a tap, and if I open it, I'm sure I can drink it. That's safe. Electricity, I can, well, plug my finger into a socket. I don't recommend it, but I can, and I will survive because it is safe. And then there's the internet cable, and somehow it's so much easier for, for garbage to, to come in and, and to go out apparently, and than, than for the good things.
[00:26:40] Ruben Verborgh: And where's our control panel? Like, how, how can we deal with it, huh? So there is more good in the world than bad, to get very philosophical. But apparently the filters and the amplifiers are set up such that some of the, the, the bad stuff, some of the garbage has reached us too, too far, and we need to take control of that.
[00:26:57] Ruben Verborgh: We need to change that.
[00:26:58] Per Krogslund: Yeah, that's what I'm thinking about when I'm asking you this question about like, are, are the internet losing its purpose as, as information. Because it seems like how people interact with, with the internet today and, and the way they're perceiving information is through, like, a highly curated algorithm that gives you them, like, very specific things, which is maybe not so much optimized in their interest, but more optimized in this, like, interest of creating, like- Uh, engagement based on rage.
[00:27:23] Per Krogslund: Or even when you just go and, like, go do a Google search from the company that wanted to make everything available to everyone, like you now see like an AI summary instead of the actual, like, choices you have for, for content. So all of a sudden you have this, like, facade in front of all this, this information that's available to you.
[00:27:40] Per Krogslund: The information's still there, but can everyone actually find it anymore?
[00:27:44] Ruben Verborgh: Yeah. And, you know, people sometimes, uh, come at me like other room, and you're a computer scientist, like, "Tell us. Tell the world. What do we need to do to change the algorithm such that instead of promoting garbage, it promote good things?"
[00:27:56] Ruben Verborgh: Yeah. Like, what is the magic sauce? I'm just like, "But we know the answer." Like, like, we can optimize the algorithm for anything you want. What we don't know the answer to is how to still make money if we show things that make people happy. Like, the problem is that we don't know how to make money with it.
[00:28:10] Ruben Verborgh: Like, apparently algorithms that promote content that upsets you, uh, make more money, and it would be irresponsible if those companies did anything else, frankly. So the problem is that those incentives are, are, are misaligned. So we know technology buys perfectly how to give people content that they love, uh, we just don't know how to make money with it.
[00:28:29] Ruben Verborgh: Um, or at least we don't have, we haven't deployed anything at a large scale, um, that does it. But it's not an unsolvable problem. What we see right now is, is, is a transition. Short-term business models, but, but in the end, we will be able to align what people need with what makes the world go round.
[00:28:46] Per Krogslund: Yeah.
[00:28:47] Per Krogslund: So if we, so if we look at, like, ex- responsible use of data and responsible AI in some extent- And do you think that if, if you kind of enforce the notion of that being the, what companies have to do, that could eventually make products like less addictive, maybe less extractive, and I guess to some extent, like less profitable.
[00:29:05] Per Krogslund: But I guess for you, for your po- from your po- your point of view, it's like there needs to be a balance here. You need to have a responsible way around data and AI, but it does need to make a profit, otherwise it's not gonna happen.
[00:29:15] Ruben Verborgh: Yeah. And so in that sense, like there is no responsible AI, just like there's no responsible electricity or responsible water.
[00:29:21] Ruben Verborgh: Water can kill you, but you also need it to survive. Electricity can kill you, you need it to, to survive. And you know what? AI, you need it to survive. It needs to be, it is the basic human rights that we have decent AI and decent data, okay? So let's stop blaming the data. Data, a- and let's stop blaming AI.
[00:29:38] Ruben Verborgh: Like AI is not responsible or irresponsible. It is the way we, we package it, the way we deal with it, the way we deal with it, the way we set up our infrastructure, um, to do these things, the way we set up our economy and our world, that it's responsible or irresponsible. Uh, and the good news is that we have a long-term incentive to all the responsible things because this is how we survive.
[00:30:03] Ruben Verborgh: Uh, and the short term, sometimes people think that we can do, do differently, and they become very successful financially in, in doing so. And you know what? That's a quirk of the system. That's a quirk of the system. But in the end, um, you know, technologies are not responsible or irresponsible, people are.
[00:30:22] Per Krogslund: Mm.
[00:30:22] Ruben Verborgh: And we as a society need to organize our technologies in such a way that it helps us all move forward.
[00:30:28] Per Krogslund: So if you look at today and, and, and say, what are the kind of tech- technical, and I guess also maybe policy decisions we might need to make today to like, uh, make this a better thing in the future.
[00:30:40] Per Krogslund: Like, what is it that we are kind of like not doing right now to, to make this better in the future? I guess like you can, you can like, depending on if you're like overly optimistic about AI, you're overly like pessimistic about it, you can kind of see two paths forward. Like either people think that AI is gonna save the world, like da, da, da.
[00:30:56] Per Krogslund: Or there's this notion that AI is just gonna make everyone dumber and, uh, take away all the jobs. I don't know either of those is gonna be true.
[00:31:04] Ruben Verborgh: Well, the one thing we should really do is get out of the denial phase. Like whatever the five stages of grief are, nobody knows them because we only remember denial, but we're still stuck like AI, like, like look, the world has changed.
[00:31:16] Ruben Verborgh: It will never be the same again. That's one thing we, we, we know for sure. And so trying to fight it is wasted energy, and that's irresponsible honestly, because our children deserve better, really. Um, so, so that is not gonna work. Um, what I also think we should do, like after the denial phase, I think is the finger-wagging phase, especially the EU is very, very good at, at this.
[00:31:35] Ruben Verborgh: Uh, like, no, you're not supposed to. Um, we also see it in the US like the whole appeal to, to, to ethics a lot. Like, but businesses and, and CEOs have, you know, morals. Um, they might, they might not. It shouldn't depend on that. That we should not have dependency on any individual company, these morals. No, no, no.
[00:31:54] Ruben Verborgh: Like, we should set up a system so that that doesn't really matter. So that this whole... I mean, look at GDPR for instance, like the US massive data regulation. It's 2026, we're celebrating 10 years of GDPR. Yeah. So what? We wag our fingers really, really, really hard. Has anything, no, become better? No. Become worse.
[00:32:13] Ruben Verborgh: And AI needs even more data than this.
[00:32:16] Per Krogslund: Yeah.
[00:32:17] Ruben Verborgh: And this is because I think do not try to legislate with ethics and finger-wagging what is fundamentally an economic question. Like, you know, what is fundamentally the reason that I'm not stealing a pizza from a supermarket? They make it really easy to pay, right?
[00:32:34] Ruben Verborgh: You've got three sorts of people. You've got people who always do the right thing. Well, don't bother, right? People who always do the wrong thing, they will always steal. But there's people like me who will try to do the right thing if it's easy enough. So you make it easy to pay for my pizza, I'll do it.
[00:32:47] Ruben Verborgh: What we see now with like AI and, and data is that we make it so expensive to do the wrong thing and the right thing, but the wrong thing, whatever that may be, is still cheaper. So therefore we all do, because we're in denial that this is just an economic system. Like, we need to just accept that AI will be here.
[00:33:06] Ruben Verborgh: We need to assign a commercial function for it. It won't be what we see today. We should start regulating this. But, but the whole, "You're not supposed to," and, uh, "behave ethically," is... No, no, no. Create the incentives to do it. Align the ethical incentives with the technology, with the economic incentive, and stop wagging those fingers.
[00:33:25] Ruben Verborgh: Like, do something. And above all, believe in the future generation, because that's maybe, uh, what upsets me most is that in all of this, like AI is making our kids dumber kind of, uh, rhetoric, basically. You know what? The irony is that we say, "You know what, young people, social media, we ban it. AI, we ban it.
[00:33:44] Ruben Verborgh: Don't use it because you're not thinking critically." The irony, eh? Like, like look at us, like just banning everything we don't understand.
[00:33:52] Per Krogslund: Yeah. Right?
[00:33:54] Ruben Verborgh: But the irony is that apparently big tech believes more in our children than we do because- They, they want their children's data for the wrong reasons.
[00:34:03] Ruben Verborgh: They wanna advertise. I, I get, I get it. But they believe in the potential. And we're like, "Oh, no, no, you can't think critically anymore. We're just gonna ban you from using it." No, no, no. We're gonna have to work with young people today together, believe in their potential, um, and believe that win-wins are possible.
[00:34:20] Ruben Verborgh: Just believing that they are changes completely how you look at the world. It's gonna be tough, it's gonna be challenging, but we have to believe that there's a solution and that people will do it, and I do, so I'm, I'm not worried really.
[00:34:33] Per Krogslund: I think that's a, that, I think that's a perfect note to end this conversation on, Ruben.
[00:34:38] Per Krogslund: We deep dived on a lot of things. I really appreciate all your insights here. Uh, and you come with, like, a vastly different perspective than talking to, like, a, a random tech VP that I also occasionally end up talking to on the show. So I think, um, thank you so much for your time and your, like, um, very Ruben-specific, uh, perspective on this.
[00:34:59] Per Krogslund: I very much appreciate it, and I look forward to, I look forward to, like, your book on the insights economy. Do you know when this is gonna be done? I'm not putting any pressure on you, but just like-
[00:35:07] Ruben Verborgh: My deadline for writing is the end of next month, the end of June, so it's, I'm aiming for, like, Christmas. At the end of '26, early '27 is when it should start appearing.
[00:35:17] Per Krogslund: Okay. Really, and thank you so much for being on the show. And then I have to, to go with my, my little outro here saying that Ship Happens is a podcast sponsored by, by Docker where we talk to really smart people about building cool things or just the economic incentives of AI as it was the case today.
[00:35:34] Per Krogslund: Thank you so much all for listening in.