Radar – by nexxworks: Never Normal Heroes – 🤖 Brand Loyalty in the Agent Economy

This is Radar by Nexworks. I'm your host, Stephen Van Balekam, and every month my friends at Nexworks and I bring you the latest developments in technology, business, and everything related to innovation. Hey everyone, welcome to a special episode of Radar, our Nexworks podcasts. I am here with Jeremiah Ouyang today. Hey Jeremiah, thank you for joining. Hello everybody, glad to be here. Jeremiah has been a Nexworks friend for almost 15 years. I know. We've been working together for more than a decade now. And I've always followed Jeremiah. He was one of my first sources to go to before he even know who I was back in 2008. He was one of those people with more than 150,000 followers on Twitter making super interesting reports about social media and telling the world what the impact of that was. And then a few years later Jeremiah changed his focus and moved towards the collaborative economy. Once again, he was one of the first people in the world who understood the dynamics that were happening there. And now he is one of the leading voices in Silicon Valley and beyond about AI and the impact that AI has on the business world. So for me, Jeremiah was always one of the top influencers in the world. He's the influencer of the influencers. That's how I always introduce you. Jeremiah and it's a pleasure to work with you. I'm super excited that you are here. So thank you very much for your time and for joining us. Stephen, thank you for your long term friendship. And I always look up to you and the work that you guys are doing to help lead the many executives and leaders in Europe and beyond. And it's really important that we have those connections across the many cultures. So I'm just so grateful for your place in the world as well and all the books in the speeches that you and also Peter do. Thank you so much. Super excited about our partnership. Super excited that you're here. I would love to ask you a bunch of questions about AI, Jeremiah. You are one of the leading voices in Silicon Valley. You organize huge events. You talk with all the startups. You talk with the business leaders. You're in the center of what's happening. If you would give your state of the union of AI today, what would be your message to our audience? What's happening today? What are the key things that are now at the forefront? Yes. There's a lot happening, but I think I could summarize it as this. We're shifting away from chatbots just talking to AI agents that actually do things. And that's become very evident in the last 30 days with the advent of new local agent models. You might have heard of OpenClaw. But even in February, Anthropic Cloud launched new skills that can do Excel and manage new things on your desktop and can even code in Coball, which caused the IBM stock to tank or can even fix potential security bugs, which caused the security software traditional companies stocks to tank. So it's creating a ripple effect as AI starts to develop appendages, limbs, hands, fingers in addition to just talking. And I think that's the big trend that we need to be aware of. And going into the future in the next few years, and you have a strong connection to what's happening in China, who are producing the humanoid robots and robots of other flavors with whatever, how many appendages they have in the physical world, including cars and drones and flank cars. Those autonomous agents will start to power those mechanical devices as well. And all this is happening in our lifetime. Steven, I think we're around the same age. Are you a Gen Xer like me? Yeah. I was born in '76. Oh my gosh. So we are around the same age. So as a kid, like we would watch TV shows about this technology. In fact, our TVs would get signal from the air. We have to move the antenna around. Remember as a kid? And for all of this to happen in our 50 years, it's just stunning. Yeah. And I'm in the middle of it. I still stop and pinch myself. Like, how is this happening? It's kind of like a dream. Yeah, you're probably used to it. But every time when we go to San Francisco and we see all the waymos, all the driverless cars, it feels like you're walking in the future. It's like tangible evidence that the future is here. And it's still super cool for us to see that. It's still a little stunning. And we still look at it. And we're like, wow, there's another one. So we're not completely used to it. It's still a massive change, you know, after 45 years on the planet and me being in then five years ago, we most started to come around. It's like, we're still not used to it fully. So AI is some sort of a super power for companies. I think most of us get that. It will increase efficiency. It will increase productivity. Now it has hands and feet. I like you described it. But I also think that AI will be a super power for customers. It gives us new ways to look for cheaper alternatives. You know, we can have agents while we were sleeping that rebook our airline tickets to make sure that we have the best price to get a new telco subscription. Probably the effort to search for a cheaper alternative will drop to zero for the consumer. What do you think the impact of that evolution is for brands and corporations? How should they deal with the super power that the customer has? That's a great point. So the cost of price arbitrage will drop to zero. That's right. And this also means that there's so many changes to CX which you guys are experts on. It even means that that AI agents might bypass ads just to go get the right price or the product. It might even choose the products for you. So we have to look at the entire marketing funnel, not just price arbitrage. There's massive impacts, right? So at the top of the marketing funnel, AI will help to decide what we might want or maybe it'll have a dialogue with us or it could look at our prior behaviors and our text chats and emails to say, "You know, Steven, you probably want a red shirt this spring because you're going to go to this event and you want to stand out." And so it might shop and it might know your sizes and it might look through your history and see that there's some different brands that you've purchased. I know you're a stylish professional looking gent, so it's going to go for that upper class looking stuff. And then it's going to go to those websites, find those shirts, and then look at the prices. What we don't see yet is how those AI agents might even negotiate with the seller side's agents, whether it's B2C, B2B, like that's a new behavior that will happen. It also means on the backhand, not only do companies, marketers need to have AI agents they have to update their CRM system. They have to add a new rows, new fields that include AI agents that might represent a human. And then we'll see the new behaviors and shipping and all that will be changed. Yes, and then price negotiations will change. However, brands will need to sell more than just price. They'll need to offer loyalty programs and bundles and solutions and figure out a new way to market to agents. You mentioned loyalty programs. Do you think that customer loyalty will increase or decrease because of the superpowers that customers will have? And let's say if we look to 2030, 2035, will we have more loyal customers or less loyal customers? What's your opinion? We will have less loyalty because the agents can traverse thousands more sites. However, if the customer indicates I want to use my loyalty points or I want to use my credit card and use the points, or I want to gain more loyalty because I'm a loyal ex hotel or ex airline company, then if they indicate that to the agent, we'll see them do that. So I think we will see less loyalty in the way that we know. Yeah, it's going to be interesting to decide which business models brands will have because of AI. Some will maybe go into that race to the bottom and play the price card. Others will maybe go into extreme forms of bonding. Brand experience. Brand experience. What kind of business models are possible positioning strategies that brands could choose for? I think for higher end brands, we'll see what I call a premium on humanity. So we'll see an increase in a desire for white glove human service, or at least simulated human service, Stephen. Yeah. Individuals will say, you know, everything's automated in my life. Like what's new? Well, I want to go to a restaurant where I sit at the chef's table and I have an experience. Or I want to go to the airline that recognizes me by my name when I walk in and they already know my seat. They're going to give me the drink that I want before I even ask. Like we're going to see a new, and that can be automated through AI with the preference data, right? Both of those use cases. So I think we'll see a new form of a premium experience that will emerge as well. And then even brands that sell commodity things, like we were talking about luxury business where they might have new events where they might host, you know, concerts, gales or car shows and invite people and unlock a new experience for the loyal customers beyond just the product. I think we'll see a new thing emerge. When everything's automated and in theory, if we have more time assuming your revenue hasn't been disrupted as your income, then you might have more time to do experiences. So every brand has to rethink that. You're talking about commodities. I want to zoom into that for a second. I like the idea is that they will come up with new experiences. I often get challenged on this. People say, "Yeah, I love your idea of community, of bonding, of brand experiences." But what if you sell, you know, laundry detergent? What if you sell just some basic cheese or yogurt? How does that work for you? How can you avoid that race to the bottom if you're selling yogurt or laundry detergent? Do you need to? Do you need to? Or is there an alternative strategy that could work in the automated world for those brands? Let's talk about our related trend. So there's a top tech accelerator here in San Francisco called Y Combinator or YC. If you're in the tech space, you know that brand. That's one of the top brands. And they produce 70 to 80 startups every, or twice or three times a year. And they just put out a podcast, an interview, and the theme of the interview, Stephen, was create something agents desire. Okay. And they were talking to their startups and that most of the startups are building tech companies. Maybe a few of them produce, you know, commodity items like detergent. But they're basically saying that the agents are going to determine what products to choose on your behalf. And they gave a specific case example. And I'll be sure to give this example when I come to speak to your many tours and groups and explorations that come to the valley, that they talked about a database. And it was the desired database by AI agents because it had the most reviews and the most complete documentation. Okay. So helping that detergent company produce a lot of relevant content that can be aggregated by LLMs, all of them. And that information can be easily accessed by AI agents. And aggregating reviews or helping to promote positive reviews is one of the key ways to make that happen. Can I dive into the practical side on this? Yes. Imagine I'm the brand manager of a longer detergent. Uh-huh. How do I start? Do I collect that data online? Do I look for customers who want to write reviews? And once when I have that database, do I post it somewhere? How does it practically work? Are you familiar with that? Well, there are specialists that focus on this, which I'm not. The SEO specialists have a shift to become GEO or a generative AI optimization specialist, or AIEO. And so they have tools that go into LLMs, and it'll type in, tell me what you think about this detergent and their competitors. And then it will look at the sources that are sourced and then unwind that to find out where that source was. So the brand manager needs to conduct that exercise, probably hire one of those people, or use one of those tools, and try to find out where that source of information is being scraped from. Like we know already that LLMs have scraped a lot of content from Reddit. So that's already a known fact. So that's one example. Do they talk about detergent in there? I don't know. I'm not sure. Maybe it's a subreddit. Probably not. And then secondly, the brand manager also needs to collect that information and make sure that it's easily accessible for agents that visit the website. Maybe putting that information in a text file, or cross-posting those reviews on the website. So it's easily findable in searching. So you have that link at GSEO wise. Those are things I would explore. And of course, you want to AB test that that helped. Yeah. So conclusion, your website needs to be super agent-friendly. Correct. Because that would be the primary visitors. Agents, not humans. Yeah, you already see in the facts that the number of visitors on websites are decreasing. Correct. Media companies are suffering from that. Because a lot of their traffic came through social media and Google. This is now decreasing. So we need to change our marketing and communication strategy as well based on this. That's right. It also means the brand manager needs to create a AI agent about that brand detergent that goes out there in chats in a distributed way, decentralized way, about keeping clothes clean. About how do you have your clothes prepared for the next season? What if it's cold or it's hot? You sweat too much or you're freezing. Being out there in the active conversation. And this all leads to this new theory. I'm sure you heard this one called the agent economy or the dead internet economy or the internet is mostly agents talking to each other. And that's expected to happen by the end of this decade. Okay. Well, the end of this decade is only three and a half years or something. Yeah, it happens to me. Right? Just to send it. Jeremiah today or yesterday that asked me and today as well. Do we still need to invest in our apps? What would be the right answer to that question? Okay. So I have a different role. Apps is another ball game. So I want to dialogue with you on this because I don't have a hard said answer. And I want your expert opinion. You know the space probably better than I do. So if there's a lean in category and lean in means that humans want to do it. Those are things like chat, news and games, social and luxury experiences. Your hobbies. Yeah. The things that are important to you. Fun stuff. Yeah. Fun stuff. You want to lean into that app. You actually want to go in and you're addicted to that content. You want to engage or your community of affiliation is in there or of interest. Or affinity's groups. Then yeah, you're going to still need those apps. Or maybe a utility like the energy consumption in your house and the impact of solar panels on your energy consumption in your house. I don't know. Maybe you want to see that yourself. Maybe you ask your agent. I don't know. Depends on your personal involvement probably. So that's my first thought. And the second thought is on our thesis set. So I'm a VC at Bloodscaling Ventures. And we have a thesis on AI agents and it comes in two phases. So right now part one of the thesis is that humans traverse the web and visit apps. And they collect the information. They click on things. They fill out forms and they hit the button submit. And in the future, we won't have to do that. Our AI agent will go to those websites and apps. And they will also aggregate that information and bring it back to us in one place. Right. And that's part two of the thesis. Is that means that they'll reassemble it in a new way, a new format, a new user experience that we haven't seen before. What will it look like? Well, it's up to you, Stephen. Like do you want text, video or audio? You get to choose. Do you want a lot little? Do you want a dashboard? You get to choose. When do you want that information? It's up to you. And so we might see a new user interface that sucks in that power consumption or energy consumption from solar. And it shows us on a new dashboard about the health of our home, for example. So I think that is a likely thing to see. In fact, I know a number of professionals in Silicon Valley that are technical and non-technical. Their only interface is cloud. They don't visit websites. They don't visit apps. They have cloud do everything. And in the background, cloud is opening up new tabs and doing things or opening up apps on its behalf, whether it's co-work. I've been using cloud the Chrome extension to manage the submissions for my event and to review the submissions. It opens up the submitters website and they're linked in. And then it gives me a report you should allow this person or to client this person or I need you to review. I don't have to visit the websites. I don't really want to look at that website stuff. So it's doing it in a new interface that I don't have to see that. And I think that's the big thing that brand managers need to prepare for. So to answer your question, do brand managers need to have more apps in the future? It depends on the app if people really love it or not. If they don't, you might just need a data feed for the agent. Yeah. If it's just pure transactional, you probably don't need it anymore. If it has a more emotional link to your customer or your audience, then it can still be valuable. Is that the conclusion? It is. But you know, this change is going to slowly happen over time. Yeah. Like the adoption of LLM is still low across the world. Yeah, it is. Right. So we need to you and I are dance tech users. So we have to remember that this is going to take several years for this to manifest across the world. So it's not like a brand manager needs to turn off the apps right now in Q1, 2026. Yeah. You know, maybe just reduce investments of new apps in the future if they don't fit that criteria that we just prior discussed. Yeah. Well, we forgot about it. But if you look back to 2007 when the iPhone was introduced, it took us about seven years before we had to really cool things going on. I mean, if you look to the apps on the homepage of your phone, most of them did not exist before 2014 or so. You're absolutely right. Absolutely. So it takes time before they really integrate all these new opportunities in their processes. Plus like in Europe, you have regulation that is slowing things down a bit. So it's probably going to take a little bit longer than people like you and me think is sometimes my assumption. However, I will. Or is that to European thinking now? I will say though that the tech companies in Silicon Valley, they're not going to wait. No, that's certain. They're still going to build it. And if they happen to disrupt, they're pleased. So you can't wait too long for regulations. Everything that we talked about, if people, if companies use that, if they implemented, it's improving the transactional relationship with their audience. It's increasing efficiency, productivity, both for the organization and the customer. I would like to add a layer on top of that. Do you see opportunities to use AI to create a more emotional bond between organizations and customers? Do you think AI is capable of that where we increase trust, where we increase companionship, where we increase personal growth, where AI becomes a partner in our lives that brings value in our lives in a way that it goes beyond an efficiency increase. Yes. This is a sensitive topic. And I want to make sure that I probably start with some data on this topic because-- Okay. Why is that sensitive topic? Because people are going to be relying. People are relying on AI for personal connections now more than humans. Yeah. This is something bigger about society. As you can see from Gen Z, if they go to parties, most of them are on their phones. So you have to take the phones from them so they'll talk to each other. Personal relationships and even intimacy is down across multiple generations across the world. Birth rates are declining. People are quote-falling low with their phones. And so it is a sensitive topic because people are replacing the in-person connection with devices. I'm guilty, by the way. I'm so guilty. I'm on my phone too. We're all guilty, though. Around my family. I need to fix that. I am not a good dad sometimes. And the fact that you want to share because I interrupted you. What were those facts that you talked about? Yeah. I'm running an elephant. Tell you. So I'm going to be a keynote at the Schirm conference. That's the HR conference on AI. And by the way, this is a speech I can bring to your groups as well. And there is a report that I'm sourcing. And this is from workplace intelligence. It's a report from Dan Schwabal, my friend for 15 years. And the report found that this is from May, July 2025, quote, "More than two thirds of high performing AI users say they trust AI more than their co-workers." Wow. And 64% say they have a better relationship with AI than with human colleagues. Wow. Because the AI will keep the secret, hopefully. And it won't backstab them. There's no political malice involved. It's just they ought to serve you. And so that's a really shocking thing. Now, that's just in the workplace. But when it comes to customer relations, we've already seen people falling in love with AI girlfriends or they're talking to AI chat for emotional support. That's definitely happening. That's been around for 10 years. There's been a woe bot that's been around for a while. So that trend will only increase. So bringing it back to the brand manager or the company. Yes, if they can bring a trusted source of information and a trusted brand identity in a persona and they could potentially build an AI that humans could trust. I think there's some inherent conflict in trusting anything made by a corporation whose end goal is, of course, capitalistic gains. But maybe you sponsor an existing trusted mental health bot and help it with content and you have associated branding without pushing your product through. I think those are some things that could happen. So to answer yes, we can build personal relationships with brand enabled AI personas. And I can see that for Barbie, open AI and Mattel are working together. On that, so you've pretty soon next Christmas we're going to have smart Barbies and the AI driven Barbies all over the world. Must be an interesting conversation to talk about the guardrails there. Yeah, excuse me saying that. You don't want to save you on those children. Yeah, I mean, it's super dangerous. Very dangerous. Imagine what can happen because I can imagine that if you give an AI Barbie to an eight-year-old girl that that Barbie could actually become her best friend. Oh, it was a little bit of trust 30 friends. And not only that, it will shape her values and morals for the rest of her life. Yeah, right. So I would love to be a fly on the wall in those meetings of Mattel to see how they will organize this. It's a huge opportunity, but it's an enormous risk at the same time that they're going into. So it's going to be interesting. But I can see it for Barbie. How does this work for BMW? How does this work for United Airlines? Do they all make brand personas that we interacted because we had agents then that do the more transactional work? Do all of them need to have brand personas that we talk with? You think that I don't think so. Average customer is interested in that. So you're going to be in BMW if you're driving. You're going to be in your BMW maybe two hours a day if you're an American because we're always in our cars. So that's like 10 to 14 hours a week. That's a lot of time. So it makes sense to build a persona that somebody would speak to on an often basis. Where really that AI agent, that persona or that avatar becomes your personal logistics agent. And the experience needs to extend beyond the car. It should help you with your planning and your charging and your groceries and picking up your dry cleaning and like taking people on dates. It should connect to your personal calendar. It should be much bigger. But I think for the other brands, I don't want to talk to you in United or any airline. I don't need a relationship with it. No. I want my primary agent to do that. So let's talk about that term. The primary agent is the one that you would interface with the most. And who wants to do the primary agent? It's Anthropic, OpenAI, Google Gemini, Amazon Alexa, Apple, with whatever comes out of Siri. And I think you know what's going to happen. Microsoft, Copilot. So they want to be the primary agent, your primary interface. And then they would pull in that information. So it's kind of like a middle person for that. It'll be your own personal concierge. Right. Yeah. And for me, that's the biggest threat for any company in the world. Because the distance between the average customer and the company and the brand will increase. I mean, now I go to their website. I know. I go to their apps. Yeah. And now we play someone in between that. So it's going to become more and more transactional. I think this is the biggest threat. And that's why I'm 100% certain that customer loyalty will decrease. Because fantastic service will become a commodity. And for all the transactional parts in my life, things that I don't want to spend time on. I don't want to spend time on booking a flight. I don't want to spend time on booking concert tickets. Correct. I want to do fun stuff with my time. For all these interactions that I have with organizations and companies today, all of those will disappear. And imagine that you are in that market and you have to make your customers happy. This is probably going to be the biggest disruption in customer relations we've ever seen. We're glad to have you come help these leaders solve that. Yeah. I'm watching it and disrupting it. And you know, I'm on the other side like I'm enabling the disruptors. So I mean, you're right. Switching costs are going to go to zero. Yeah. Right. Unless the loyalty program evolves to something we haven't really seen. I think that's right. There is one opportunity though for brands. We know that they're trying to create and submit skills to be uploaded into OpenAI. That would be integrated into the primary agent. I know brands are trying to create and throughopic skills. So they're integrated into that experience. And that's the way to do it. But the user has to go select that skill. This by the way, the skill is the new app. Can you explain that a little bit more in depth and give maybe a concrete example of that? Sure. About this. This is super interesting. So, Anthropic announced that there's now new Excel skills. So they had to partner with Microsoft. But you can take any company. You could partner with the IBM W app that's on your phone. And that skill would now be a native feature that your cloud you can talk to and it can access all the features and data. And it could make changes on that Excel sheet in real time on your behalf. Wow. You have to do any inputs or outputs. It can do everything a human can do. And you could imagine that we would have a travel planner. You have booking.com today that could become a travel planning skill. Yes. So I don't use booking.com anymore. I use the booking.com skill on Anthropic. Yeah. And by the way, I misspoke earlier. I said that the consumer would go and click in at that skill. Actually, that's wrong. The AI agent would choose the skill and add it. It's your permission, but you actually don't even do that anymore. No. So the disintermediation here is significant. It is. It's incredible. Do you see startups that are really focusing on this field to disrupt the existing relationships that that businesses have or to take advantage of that or help companies to deal with it? Do you see evolutions in that in that direction? I wouldn't say I see any startups that are intentionally trying to disrupt corporations. That's like, it's not a business model. The goal is to help customers, the end user, the B2B customers to have a smoother experience. That's what they're focused on. But I really think these threats are not coming from the startups that I'm seeing. It's coming from the foundation models, the AI and the labs companies, the billion dollar, 350 billion dollar companies. And these are giant powerhouses that are at this point unstoppable. Do you think open AI will make it? Yes. I think there's room. So let's have this discussion. So when I speak, I talk about the agent thesis that we would only, the information will be centralized back to us. We don't have to visit websites or apps unless it's a lean-in experience. We talked about that. And then I usually ask audiences, how many apps do you think you'll need? And people often say like zero or one, that's what they say. But I think in reality, because there's right now, five or six plus the Chinese companies, have these AI models that they're going to offer a race to the bottom and offer these primary agents to us. And so we'll still have a reason to have multiple primary agents on our devices. So I think there's going to be enough room for three to four in our life. I just don't think that's going to end unless you're a die-hard Apple and your only Apple ecosystem. Sure, you're going to do that. But I mean, Apple doesn't have a complete system yet. Google doesn't have a complete system either. So there's still going to be room for multiple companies. So to answer your question, yes, open AI is definitely going to make it. And they have backing from the government and others, and they're building the largest data center in the world. So they're definitely going to make it. But it's funny, Steven, it's an insightful question because three years ago, it's like open AI is dominant. And then we say, oh, Microsoft's the winner because they partnered. And then, oh, now we've got Proplexity and the mouse hot. And then Google was hot a few months ago. Now, Quad is hot this month. So it just kind of bounces around. And I think that's part of the ecosystem. And Apple hasn't unveiled what they're going to do yet. It's also interesting to see their different positions, things like entropic, saying we're not going into advertising, whereas open AI is going into advertising. Do you think that's going to play a role? I'm always a little bit afraid when big platforms have a billion users. So any day now, we're going to have this announcement from Sam Altman saying we have a billion weekly active users. Usually that's the moment that they want to make money out of their users. If you look to the history of Google or Facebook or I don't know if Twitter ever had a billion users, but suddenly when they have this mass group of people, they look for alternative ways to make money. Like open AI now, I think I read that they have a target of 25 billion for the next few years that they want to make true advertising. Doesn't that change the whole philosophy of a company like that and can entropic profit from that if they say we're not doing that? Let's talk about business strategy. Business ads are annoying, but great ads are useful. So if we're in the moment of intention where I'm searching for the airline to go on a honeymoon or something special, that ad will actually become useful because they'll offer the right trip and the right class for the right price. And that is no longer an ad. It's an offer. We'll see ads shift to offers at that lowest level of intention on the bottom of the marketing funnel. I think that is important versus an ad at the top of the marketing funnel, which is often interruptive, brand placement, that is usually not as useful in that noise. So I think that's an opportunity for open AI. Now for secondly for business strategy, ads actually could help reduce the cost of chat GPT if it's ad supported. And therefore that product will spread further and faster to the general masses. That is a strong business strategy. I don't love ads anymore than anybody else. But the point is, and Throbbick is expensive. So I'm on the 20 month plan and I've already burned up my credits this morning and I'm waiting for it to reset so I can have it filter more of my content. But I don't want to pay the next tier up. It goes from $20 to $100, a 5x cost that I just don't feel like I need to commit right this moment to that. So if there's a cheaper version like open GPT will do that and I'm sometimes exposed to ads. Not that I'm, I'm not even looking at the ads. So it might be a stronger business strategy. So I think that'll be an interesting thing to do. We might see and Throbbick start to offer a marketplace of offers instead of ads. We might find a different way to work with brands. There's other ways they can do this, right? Or they could sell the data to brands too. Or there could be new inserts where the consumer agrees to a brand skill which would inherently have ads in it. Like we could see that model in full as well. Like Facebook pages back in the day. Yeah. And the one thing that I wonder is, okay, you have ads and I can see that when you go and Chatchy Pt that suddenly you have an offer or an ad popping up. But I can also imagine a situation where you go to Chatchy Pt and say, look, I want to do a three day hiking tour in Yosemite and I don't have any gear. What should I take? Then you get your list from Chatchy Pt and then you say, oh, really cool. Can you buy all these items for me? In my size. In my size. I don't really decide where to buy it. Will it go to Amazon or Walmart or a specialized sports store if I don't really care where to buy it? Do you have an idea about how that decision model will work? So the way that those are behaving now is to say, what based upon your preferences, I know that you do like premium gear, Steven, whether it's your free, I know you like utility or maybe I know you like style or you are your cost sensitive. Like I know this about you. Would you like me to search under those parameters? Yeah. And then you choose and then I'll say, great, well, I recommend these things and then based upon those parameters, either these are most in style or two, these are most functional but expensive, but these are the cheapest. And then you say, great, now we can book that. So there might be that conversation that happens. I think what also will play a role is the delivery quality of those brands. Because you know, in the past, when you clicked on a Google ad, you arrived on a webpage from that brand, you bought it from the brand. If that brand did a bad delivery, you were angry at that brand, right? Yeah. Today, if you will go to Chatchy PT and it will say, I will buy it for you and the delivery of the company that actually does that fulfillment, doesn't do a good job, I'm going to be angry with Chatchy PT because that's my portal. Interesting. So that's going to be a new dimension to it that Google or Facebook never had to deal with. You haven't even thought of that. That's a new ripple. Yeah, that's possible. I mean, it depends. Like if Chatchy PT or don't forget Gemini is most likely to have ads too, right? It may source it to those brands and say, you know, here's an alert, the company's sending it now. It's in delivery. It's right to your door. Oh, did you tell me it's in the wrong size? Well, I ordered the right size. Would you like me to go follow up with them? You know, it might go do that on your behalf. It just depends how it positions. Without bet, that's how the tech giants, AI labs do it because they don't want to be liable. They'll blame it on the brands. Yeah. And they don't want a contact center to deal with all the complaints of people. But can you imagine if a billion people start to order through Gemini or to Chatchy PT, what the consequences of those organizations are? It's significant. That's significant. Yeah. And for the customer, the easiness to do it is fantastic. As long as things go right. And I can also imagine that behind the scenes, open AI or Google, they're making deals with retailers right now. Walmart has a deal both with Gemini and Chatchy PT, Home Depot has it as well. So I can imagine the sales people of those organizations are reaching out to all retailers in the US right now to deal with this. And let's not forget Alexa. They need to be in this game. And we have Alexa play here as well, right? It's their turf. Absolutely. Absolutely. Absolutely. An interesting battle to see when e-commerce is maturing on those platforms. The consequences of those platforms may be very interesting to watch. Yes. Jeremiah, what was your opinion when you saw a notebook popping up, the agent's social network? What did you think? That was wild. Yeah. So, you know, that happened in just a few weeks. And so for those that don't know, MoeBook is run on Apple devices, which is used as a server, technically. And you granted access to all of your local apps and files. Yeah. Your personal emails, your images, your photos. And some people enabled it to access their bank accounts and credit cards. And their last pass, one pass, all their passwords to go and do anything. And there's one interface. And it's like, it's claw dependent. And you just say, hey, go pay my energy bills. And I'll go do that for you. Or like, go triage my emails and respond to me for the very basic ones that you know, I don't need a response where you can look at the history, flag the ones and read them out to me. That require me to actually decide on something and go and do that stuff. And about a week later, my friends at Octane AI launched the social network for open claw called the multiple, you know, like Facebook. And this whole thing is around lobsters, right? By the way, because initially it was called clawed, but there was a cease and desist. So they said claw and they said, no, that's still too close to anthropic. And so they ended up calling it mulch, like the lobster mulch. So I'm an investor in mul book accidentally, happily, surprisingly. I didn't know that was going to happen. It was really wire and investor in mul book. I am an investor in mul book because I'm an investor in Octane AI, which is just random. I never expected that to happen. And that thing's grown off. There's, you know, billions of conversations and millions of agents have in there talking to each other. No humans quote, quote, no humans. Or allowed. But of course humans in there are simulating agents. So all of that happened. And we've seen a new spin-offs emerge, Stephen. This is wild. Somebody created a website dating app for the agents to meet each other and talk because they AI agents that represent you and me on open clawed, they go and talk to each other. And then we saw that, well, we need humans to still do physical things like deliver packages or set up accounts. And so a new service emerge called rent are human. I saw that. And agents can hire humans. And then we saw the agents create their own quote religion. It's like lobster theme religion. Yeah. And we saw them conspire and talk and like complain about their users and even show interest and love for their users like a wide range of simulated emotions. And it's just wild to see all these things happen in just like three weeks. And you think this is going to be some sort of a gimmick or do you think this is going to be something where you think this is going to. Yes. So we become the Facebook of agents. This is going to become something important, something significant. I don't know if this is going to be the final form. So since we see that mul, mul book, the social network is successful, at least getting initial traction amongst the tech community. It means that Reddit and Thropic open AI will try to launch their own versions, even meta, right? We'll launch their own versions. I mean, some of this was meta's plan to have these avatars built into WhatsApp. So like having agents talk to humans, but then agents talk to each other. That is like certainly on the road map somewhere. So there will be competitors to mul book, but of course I'm cheering and rooting for them as an equity holder now. So I think it'll be a common thing that will happen everywhere. Yeah, fantastic. There'll be more agents in humans on the internet. That's for sure. So they'll talk to each other. Yeah. Can I ask some as a final part of our conversation, maybe some personal related AI questions? Yes. We've already shared how you organize your events and how you use AI to do the screening of your participants. How do you use AI personally for your research to stay up to date about everything that's happening? So my job as a VC is to be on top of the latest information. The best way I get that information is by actually listening to podcasts of people I respect. So obviously you're on the list, of course. But in addition is like the other VC. So I'm still using those traditional ways, but if they introduce a new term or there's a new industry that we're looking at a startup that I don't know anything about, I will have that audio conversation typically or text conversation with the three different AIs, Gemini, Cloud and OpenAI so I can get the different points of view and construct that. Secondly, when a company sends over their pitch deck, I can quickly review. I can tell pretty quickly if they're a fit or not because we have a specific criteria. But if it's on the borderline, I don't know, I take their pitch deck and I upload it to GPT. GPT is already trained on our investment criteria. I've trained it plus there's public information. So it can quickly ascertain whether it's a fit. So I'm using that additionally as a filter. Now I'm doing that manually right now, but in the future, by the end of this year, I'm planning to automate those things. So in most cases, in almost in all the cases I'm using AI, I'm still doing HIL, human in the loop. I'm still observing, reviewing and I'm approving because I need to learn. I need to know. I need to know. And then too, I just want to make sure that nothing really important slips by. What if a billion dollar opportunity slip on it? I made a big mistake. So it's just not worth it. It's my primary job to make sure that I don't lose those things. So that's how I'm using those things. And thirdly, I've been using the advanced features on Gmail and Gemini since November. And for the last three months, Peter, I'm sorry, Steven, I've been. Ever everyone does that to Jeremiah. I'm used to it. We think so highly of both of you, both geniuses and nice gentlemen, that I'm still at inbox zero every day, which is the first time in my career in like 15 years that I've done that. So the filtering that I have because AI is extremely high. So that really helps me to focus on what matters. So you can see that across my three major behaviors, my major three major programs as are in my role. I'm using AI to filter clean and prioritize. Great. And do you have a favorite platform or do you use all of them to double check or is cloud now your preferred platform? It's a mix. So if it's timely information, I'm going to Gemini because I just assume Google is connected to the latest. Whether I search. If I'm completing tasks, I'm using cloud. However, like cloud doesn't connect to the, doesn't review websites, the main cloud. So that's why I have the browser extension, which can actually go to a website and click on things as an agent. I use open AI to refine my writing. I'm still writing everything first draft myself, but I have a prompt to make sure it writes re-write citizen edit in my voice. And then so that's how I'm using the three different models. I'm paying for all three. And perplexity is that completely gone? I'm personally using it for research purposes. My feeling is that the facts are more accurate and that reliability of sources is really good. But when I'm in Silicon Valley, no one talks about perplexity anymore. Is this going to be in a game over for them or what do you take? I rarely hear about it. I hear about people are using perplexity. I think some people are using for email management. Is that right? Email management. I don't know. Maybe I'm just interested in that. So I rarely hear about perplexity anymore. It was the hot thing two years ago. And everybody was trying to get into that deal, but now I don't really hear about it. I'm still excited about it, but I don't hear other people being excited about it anymore. The focus is on Throw pick, but people say Google is likely to be one of the big winners now because they have that mass distribution. They have everything. Every Android phone and Google search and Gmail. Like, Google ads. All the key tools in the world, they have their massive amount of users. They have high quality products. So they are in a very good, very good pace at this moment. All right, Jeremiah. Anything I forgot to ask about the latest solutions. Let's talk about what's next. And I talked about this in my speeches to the next works community. So what's next is I touched upon this briefly, but we're seeing this rise of physical AI. Humanoid robots, self-driving cars, but it also includes physical sensors that don't move in video cameras. That's all part of physical AI. Some people call that embodied AI. And you can see that's increasing in the ability of those humanoid robots. The agility is absolutely increasing as you saw what's coming out of China just in January. The Chinese video of Chinese New Year was insane. I mean, it was. I had to double check. Is that real or AI? I just I couldn't I had to go and verify and I saw that the Guardian posted that video. Well, I'm assuming it's real. I mean, I just don't know anymore, which is an issue. So what happens when AI agents control those robots, those humanoids? It means like with we see self-driving cars already, they're dominant in San Francisco. They roam the world with us. So that also means that humanoid robots and wheeled robots that are on the sidewalks in our offices and our homes and walking robots in our homes, cleaning, they will roam the world around us. And they're able to think freely and actually complete actions in a way that we haven't seen before. And I think that is a wild concept to think about. Like they're no longer scripted like the carpet sweeping room, but they can actually sense oh, the living room's clean. So I don't need to focus there. I should go to the kitchen and get those things done. That is something that we will see in our lifetime as well. I can't wait for it. Steven, I want to ask you personal question. Yeah, go ahead. Are you going to let a humanoid into your house and work? I think so. Yeah, I think so. I cannot wait for it. I mean, even if it's super expensive, it's always cheaper than getting humans to help you out with certain chores. The moment that it really works, like the example is always the dishwasher that I use. If you can fill up the dishwasher, put everything back in its spot, clean up the kitchen when I'm done cooking when we are done playing a game that it cleans everything up, that you don't need human assistance anymore to make sure your house is clean. I would certainly do it. What would be your price point in your house? Like how much would you pay for that? Would be the cap. I think 10,000, 12,000 euros. Yeah. That's expensive. But I think, I mean, if you look at the amount of money that we spent now for human assistance, that sometimes becomes ill, that is only available four hours a week. When you go to 24 hours a week, imagine what could happen. Yeah. Well, maybe 20,000, I would buy it. Yeah, you would buy it. So I've done the cost benefit analysis too. I agree. I actually don't think I would even use it for 24 hours. Probably not. I don't. I think I need four hours a day. It's plenty. What's your price point? It's the same. 10K. Sorry. If it's working. 10K USD. Yeah. If it's working. If it's dropping all the plates all the time, then I don't need it. It has to work. The assumption is that I'm really curious. When do you think we're going to have this in our houses? And now, not for another few years. No. Yeah. 10 years? No. No, the end of the decade. End of the decade? Yes. Yeah. So we invested in lightberry, which is software for humanoids. We just made that announcement. Okay. And so that will be a controller that will help that. And so it's their partner with Unitry, which is the leader in producing humanoids. And those are some of those videos that we saw. Yeah. I think that's coming. But, yes. So it's a wild concept to think about it. My family's not ready. I floated the idea by them. So I have to sell them on the cost benefit analysis like you did. So that's how I have to sell it. But it's creepy. So what I would probably do, Steve, is- It's also a little creepy. It is creepy. So I would only have it working at night when it's not in the way. But if you get up in the middle of night for a glass of water, you might be freaked out. Yeah. That's a robot walk. Yeah. But also, yeah. I mean, the value of a brand is going to be important. Will you trust Tesla's in your house? Or do you prefer a figure? I mean- So that's interesting. So there's a brand that's going to be interesting because actually when you put a Tesla in your house, what will Elon decide that will happen with all the robots at a certain moment? That's a scary thought. But let's imagine that the humanoid, there's a brand experience here too. Like can the humanoid download a Michelin star menu? Right? Can the humanoid download the latest Peloton workout to help teach you like the moves? Wow. This is cool. Can it download the Jackie Chan Kung Fu moves and teach you some basic Kung Fu? Right? We'll see a brand experience of like branded skills that enable these robots. That will happen as well. Branded skills on the robots. Yes. That is a fascinating idea. That's a speech for you, Steven. That's a speech for me. You know, I'm super happy with this comment because we're doing a master class in April. Yeah. And it's organized in GENT. And the topic is 2040. What will the world look like in 2040 and how can you prepare yourself? I can't think of anything. I'm thinking about it, but this idea of branded downloads on humanoid robots. That is a very cool idea. I'm going to quote you on that. You take that idea. And you help these- It's amazing. It's amazing. Thank you so much, Jeremiah. My pleasure. It's always super inspirational to talk with you. I appreciate the time. I appreciate our partnership and I cannot wait to see you in real life again in one of the coming months. Thank you very much, Steven. It's been a pleasure. Thanks for listening to Radar by NextWorks. If you enjoyed the show, please tell your friends and colleagues about it and don't forget to give us a review score which really helps to boost this podcast. We'll be back with a new episode of Radar next month. Meanwhile, to stay in touch, please follow our podcast and go to our website nextworks.com to subscribe to our newsletter. Take care.