NapsterはいまAI企業になった。本気で。

When you give all of your ideas and all of your information that lead to your innovation to an AI machine, you, in essence, have sold your business because you've lost your competitive edge. You've given it away before it's ever been able to be an act on by yourself. Wow, we're sitting here trying to replace humans. I don't know if that's really the way to go. I think we need to be using AI to capture customers. I think we need to be understanding and learning our customers. You know, when you think about an agent, what does an agent mean? You know, these are very real looking people that you as a business can train to be an expert at your business and to have a one on one conversation with your customer. But where is that interactive layer that's neutral, that can sit on top of every model that you can build your entire business on, that's completely neutral, that lets you own 100 percent of it? Welcome, humans, to the Neuron AI Explained.

So if you remember the early days of downloading MP3s, there's a very good chance you remember Napster. And if I told you that Napster was now building AI agents with Microsoft, you might reasonably think that I'd lost my mind. But that's exactly what's happening. Napster has reinvented itself around AI, specifically this idea that interacting with an intelligent system should feel less like typing into a chatbot and more like just having a conversation with another person. Today, I'm going to be joined by Napster CEO John Acunto to talk about why he's betting on video-first AI, how you rebuild a brand practically that everyone remembers for something completely different, and what happens when AI starts looking and behaving a lot more like us. Before we get started, please take a quick second to like and subscribe. And on that note, John, welcome to Neurons. Great to have you, man. How are you? I'm so happy to be here, honestly. Let's get going. I'm ready to go. Heck yeah. And we're happy to have you here. So I guess this first question I'm going to ask is the obvious one, and that is, what the heck is Napster doing in AI?

Yeah. You know, Napster really transforms music. And when you think about the fact that they really invented streaming music, and what it really meant was you didn't have to buy a full album to just get the one song you liked. I mean, what at the time? I mean, I'm old, so I admit it. I was a big Napster fan myself. Maybe I shouldn't have made that. And at the end of the day, we were forced to go in and buy this whole album for $20, $18, for that one song that we really liked on the radio. Now, there's a good part to that, right? The creative community would say that you get to learn the depth of the artist. There's a lot of great music on those albums. And nobody's here to challenge the great music and the creativity that goes into making that music. But they really made it so that, you know, a whole industry was able to change as a result of that and really gave consumers this great opportunity to, you know, get what they want and pay for what they want

in an industry that's kind of preventing that. And so amazing. And so when we purchased Napster from Algorand, we saw an opportunity to take something we'd spent seven years building as a company, which we're very excited to talk about with you today, and really use Napster as a poster child for streaming intelligence. Okay. And so you go from sort of streaming music to streaming intelligence, And we think that at the core of our business is we want our customers to own the data. Yeah. That's important. Exactly. And so that's the disruption. The same disruption that you see with Navs like, oh, what's the link? Well, the link is they were disrupting music by saying you should just buy one song or stream one song. And there was some other stuff, whether you should pay for it or not when you share it. But really, the tour was play the music you want and, you know, just give me the single, not the album.

And here we think streaming intelligence is about owning the data, specifically business intelligence and for companies. Wow. I love that. And it makes a lot more sense when you say it that way. So, like, what does the company sell today? Who's buying it? Like, does music play a role in that or have we shifted entirely? Yeah, we sort of paused the music in January of this year leading up to CES where we announced, hey, here's our agents. Here's how that works. And, you know, when you think about an agent, what does an agent mean? You know, these are very real-looking people that you as a business can train to be an expert at your business and to have a one-on-one conversation with your customer. Look, the interaction between human and machine right now at these early phases is unbelievably important.

If you're in business today and you aren't working on having your customers interact with the machine, then you're a little behind on what AI is going to be able to do for your business. So when we talk to our customers, we give them sort of a roadmap of four things. And the first part of the roadmap is capture. And using an AI agent to capture information on your customer that you currently don't have. So if you think about a website, if I put an agent that you can talk to on a website, and we have a one-on-one conversation where the agent welcomes you to the site, hey, how are you? Let's use Gap as an example. They're not a customer, but let's act like for an example, it's a good example. So you go to Gap.com, and now an agent appears, and you have a conversation. Now, for the Gap, that's very important because they can qualify the customer. Why are you here? What are you looking for?

You know, what is your favorite color? You know, when do you need this shirt by? What do you typically wear? By the way, today, how many websites even have that conversation with a customer? None. Almost none. Yeah. So the capture phase is brand new data in a lot of cases, right? So now that capture is really important because now the business can now go to the next phase, which is learn. So from that capture phase, you go to learn and you're like, okay, what do I need to know? Okay, what questions should my agent be asking? What's important to know about our customers? And look, let's be honest, right? Every single company has some mission statement or some sort of thing that says we want to provide the best service to our customers and we want to do some value for our customer and the customer and the customer. Right? Yeah. So in essence, if you focus AI and train AI to learn about your customer, right, then you're going to win no matter what it is because you're really learning more about who you are selling your products to, which should be what your intelligence should matter.

Right? Like, I'm not going Gap.com to talk about my psychological problems with AI or to get some Band-Aid. I'm going there to buy clothes. Right? And so Gap should know what I like, what I don't like, and how to help direct me to the product that I'm looking for at a great price and value. Right? And really understand me as a customer. And that's what Gap does care about. So that capture phase leads to learn, which allows a business to figure out, how do I use this new data? favorite colors, things that we like, and so on, that then goes to that second thing, that third thing, which goes to the agent, which is the act part. So I've captured all this information. I've learned a lot on the customer, and now I can act. And so the act phase, we've extended our products, which I think you actually lifted up and showed me before we went live of our product, which can allow you to have in the physical world in interaction with that same agent that you meet in the digital world.

So now the act lets me not just act online, but when the customer walks into the Gap store, there's actually a hardware product, the station, the view. We have other new products that we're going to be announcing in the coming future that are hardware products where they'll remember who you are from that online experience. So now the business can now have a conversation with the same customer online in the store and can take what they've learned about that customer and provide them that mutual experience across the whole side of that. The last part is own. This is the most important. If you don't own that trifecta, where if you have meta doing some of your capture, if you have the learn being done by Anthropic or OpenAI or others, if you have the act being done by yourself, then you don't own that sequence of events. And so, therefore, we believe with all of our heart and soul that you are, as a business, selling your identity because your identity is human.

When you think about a business, it's built of humans who have ideas that lead to innovation. And that innovation leads to, you know, customers enjoying what you've created, whether it's, you know, a marketing approach or whether it's an actual product innovation. And those ideas lead to that innovation. When you give all of your ideas and all of your information that lead to your innovation to an AI machine, you, in essence, have sold your business because you lost your competitive edge. You've given it away before it's ever been able to have been acted on by yourself. So owning that is very important. And so our agents sit on your website. Customers own that. On the learn phase, we're now native with Microsoft Azure. We're native with Foundry. And we believe on the learn side that – and by the way, I don't think it's just us that believe. I believe both Microsoft and NVIDIA have put a stake in the ground that they believe that you can run your own models at the edge.

You know that they've announced a laptop. They've announced different machines. NVIDIA continuously talks about having compute for the everyday business. So at the end of the day, we see that learn really being something back to the 90s. Interesting, back to the days of Napster and where they started, where businesses are really going to have to start thinking about controlling their compute. And not just for cost, but I think for that learn phase of keeping some of that important data about their customers and what they're learning from that secret. Keeping that to themselves, creating models they own. And that's being native in Foundry really is helpful for those that are looking to do that and to then go to the act phase, which we think is important. So ultimately, we think our whole ownership plan is in line with how Microsoft is set up to win, and we think it's also in line with NVIDIA. I mean, NVIDIA wants everybody to buy machines, like period.

So – Yeah. And that's their business plan. And they've done a good job of getting them to do it. Amazing. Amazing. So that's how we sort of see the world. And you know we announced it Bill which I think you know you were able to stop by at that time and check us out that we were beginning the process of being native which is a big deal Look it is a big deal There are 17 companies that are native in Microsoft Azure total, and we're the only agentic agent. That's wild. That's insane. It shows the power of our technology. It shows the importance of capture, learn, act, own. So, you know, when we communicate with our customers, we're communicating with that same zeal Napster did, which is don't sell your ideas, don't sell your innovations, don't give them away, preserve them. And using agents is the absolute best way to do it with the approach that we're talking about,

which we think aligns with the rest of how those companies see it. Do you think that at some point that involves some kind of a unified protocol of sorts for the data consumers share with businesses around these things that allows you to, like in an ideal world, any of that you enter anywhere would go back into a bucket you own, I guess? Yeah. robots and it's so crazy right but to think that a business is ultimately training customers to

interact with this machine that's trained on their business ultimately leads to how they train the robot to work but these robots are dead avoid a personality right yeah way to understanding your business you have to know how to train for machine to human interaction you just have to and look you You know what's crazy is that, you know, you and I both come from a generation where, like, we dreamed of this stuff, right? Oh, yeah. Yeah. And in the next 50 years, it gets to be a reality, and we get to play this big role in helping our children and grandchildren have this world we only dreamed of. It's just amazing. It is. It makes us excited at Napster to help make it so that that interactions with machines is still human and where the ideas and those innovations are preserved within those companies so that the machine-to-human approach is human. Wow. Well, you know, and to kind of build from that, you all have made a real substantial bet that the future of AI interaction is going to be this thing that is more visual, is more conversational.

And I'm wondering why is looking at AI in the face better than typing into a box? I mean, we should do this interview that way. Let's just type this interview right now. That's a fair answer. Let's just do it. Let's just shut the monitors off, folks. Let's go. Let's go. This will be entertaining. It'll be fun. It's going to be unfollowing. Follow the chat. It'll be like an Instagram thread piece where we just chat the conversation. And, you know, we're human. We're unbelievably audiovisual people. I mean, we just are. What was Facebook without video? Yeah. I mean, it was terrible. Let's be honest. It was boring. It's just a bunch of pictures. Of dinners. But back to that robotics and training and learning and customers today, again, how many websites are you going to where you're having a conversation with an agent at the website?

How much better would it be? And it's so much better. And that's what we've put into the market with F1 and the customers that have been testing. It's unbelievable. The results are unbelievable. It is a factor of 10 or 100 what it is in that sort of bland, non-visual, non-interactive space. Talking to a real lifelike person, having a conversation where they truly are interested in giving you the best product experience or the best customer experience or the best fan experience. Even in your case, are you able to talk to your fans? We can make a digital 20 of you in seconds and allow you to train that to be you to answer questions and to give people guidance. Talk people through their AI problems. Yeah. Well, exactly. So, you know, as customers start to understand the power of AI is its ability to have real-time communication. and humans are unfortunately visual more than just sort of text.

And so, yes, we're betting big. We're betting big on visual. We're betting big on robotics and the ability for businesses to train towards that. And, you know, we fundamentally believe in something that is called microscope binocular telescope. and everything that we're building is for binocular and telescope. And we understand that we're going to be misunderstood at the microscope level. The microscope is zero to three years where when we're building things and releasing things, that people look at and go, what are you talking about? You know, and then the binocular, that three to five years, three to three to eight years, you know, even that seems like, wow, you're really thinking about like we think a lot about what are businesses doing? Where is infrastructure going? What happens when all this compute goes online? What happens when there's when there's compute in space? What happens when, you know, cell phone carriers start to become less and less because more satellite is driving? And what

happened? You know, what happens? And then that that that telescope part really matters. And so to act like robotics isn't here to act like that, that that human to human interaction, Human-to-machine interaction isn't important. So, you know, for us to have potentially billions of people interact with our product and learn the best possible way for human-to-machine interaction, we think it's unbelievably vital to have that data. So your agents can talk to each other, but they still can't think together. That's the gap OutShift by Cisco is closing. OutShift is Cisco's incubation engine, building the Internet of Cognition, shared intent, shared memory, and guardrails that allow agents to work together from anywhere. Open source protocols and code, a system for intelligence that scales. Read the paper and experience the demo at outshift.com. That's O-U-T-S-H-I-F-T dot com.

Check out the links in the description. Now back to our video. Correct me if I'm wrong here. Not just demand right now. You're building toward what you see on the runway ahead as far as, like, there's going to be more compute than there is now. There's going to be increased demand for things like this. And you're very much on your long-term roadmap sort of targeting some inflection point where you think that hits almost. That's right. We're targeting a point where you are able to stand in front of a piece of a machine and talk to it, and it will do whatever you want, literally. Nice. It can execute everything. But by the way, that's happening in a prompt. Yeah. On your phone, you're prompting to do this or do that, or from your computer, you're prompting control this, do that. There's all kinds of assistants that are doing all kinds of things. But where is that interactive layer that's neutral, that can sit on top of every model,

that you can build your entire business on, that's completely neutral, that lets you own 100% of it? Yeah. Our customers own 100% of all of it. I mean, think about our product prices at a penny a minute. The nearest competitor is $0.25 a minute. That's crazy cheap, by the way. Yeah. Crazy cheap. Crazy cheap. And what enables that? And the developers own all of it. So it's just, yeah, it's a big deal. What enables you to offer it at that price? I heard a penny a minute, and I was like, wait, wait, wait. That is dirt cheap for, I'm assuming, real-time video and some level of orchestration probably between, I don't know, voice. and other needs in the background that are probably a part of such a system. It feels like it would be really expensive. It is very expensive for companies because they're running this all out of the cloud.

They're not using it yet. You know, when we acquired TouchCast, and what an amazing company, what an amazing team. They had spent, you know, the founder, you know, had spent literally a decade focused on AI and had filed patents around literally edge and having conversations between these AI docents and, you know, or what we call agents now and AI. And he filed patents on that in 2018, 2020, 2023. I mean, what an amazing forethought they put into it. But his entire bet then was at the edge compute, not the cloud compute. And so ultimately, you know, a lot of what we have built and a lot of the infrastructures, you know, been built over a decade. And ultimately, it is focused around edge compute. So in essence, your machine's doing a lot of that processing, which is cheaper, right?

Your phone or your device. And that's the big hurdle. So a lot just have not figured out either, A, how to do it, or, B, if they do do it, they run into a patent portfolio problem. Yeah, big patent portfolio problem. I'm assuming as far as the tech goes, we're looking at frames are largely generated in real time. I'm just kind of curious, like, to what extent you can share how it works. Because it's a really cool concept to think of this is in sync. It's real time that you're interacting with this at a pace that feels natural. Yeah, it's insane to think about the amount of time we've spent to not just make it where the experience is high resolution. These characters that you're speaking with are very real. And I would love, you know, to have our team provide videos for you to share with your audience because I think they would be unbelievably blown away with just how lifelike they are.

I'm going to stay away from sort of the magic sauce, and I'm sure there's a lot of tech guys that might post comments. Well, this is what they're doing, and they may be right. They may not be. But ultimately you know our API and how we set this up is ultimately allowing for a real experience but more importantly real memory and personality and the ability for there to be trained experts. You know, we have – and anybody can sort of play around with this for fun. You can download the Napster app, which is still there, and you can play around with all these tools. It's really more of a fun app, sort of a great, fun destination place to play around with what these agents are possible to do. But we've trained tens and tens of thousands of agents to be experts in disciplines. So, for example, a physician, a lawyer, you know, music experts, and so on.

And you can choose. You can actually search in the app for any expert, and they spend 100%, 24 hours a day training on their expertise and learning about their expertise. and that's all they're good at is just that that's all they're good at is just that discipline, financial planning, whatever it might be. And you can go into the app and actually have this interaction in real time with a real personality. And then when you return to the app, they remember the conversation. They remember who you are. They talk to you as if the conversation just happened and it could have been a week ago. And so you see a world where you have a crew of people helping you get through your day. So you could have somebody editing your videos. You could have somebody, you know, helping schedule your appointments. And then you can interact with them. Hey, Billy, come up and tell me my schedule. And up comes Billy, and he talks to you like a person. And I'm working on it. This is what I'm doing. And he's got a personality. That interaction is much better than the keystrokes that we're talking about before.

How many keystrokes do you, you know, I've got to pull up my agent bill and let me find him. You know, I liken this period of time back to the difference between DOS and Windows. It was a big one. It was a big move, right? I mean, to go from, you know, loading a CD into the computer and hitting C, colon, forward slash, forward slash, all this stuff, and then up came the GUI of QuickBooks, let's say, and now I interact with it. Well, Windows allowed there to be an interface before that interface, and that's where we see we are. We see ourselves the interface before the business's interface, meaning that you can have an agent that is sort of greet you, that's this interface, and then you can have other agents that work within those businesses. I'll give you an example. Anthropic, OpenAI can each have an agent. Yeah. Your business could have an agent, right?

Our agent that sits there that's an interface to all agents. So you can say, hey, I'm looking to research this. Great. Do you want to use Anthropik or OpenAI? I want to use Anthropik. Great. That agent joins the conversation. So you're still controlling, but you're using the other agent. It's not a world that's so dissimilar to what we do with interfacing with Windows. But if you don't have a Windows interface, then really you're inside QuickBooks Vortex. Yeah. Right? Without the operating system, you can't load the software. Right. You're just an AOL. And then you're like, I'm on the web. You know, and you're an AOL. Look, here I am. But you really were never on the web unless you hit that WWW, right? So we're really more of that WWW of AI than we are the AOL of Anthropic or OpenAI, which, by the way, we work with all of them. We love them. We know how important they are to our customers.

But so are the other millions of models out there, and so is Foundry where you can create your own model as well. We think those are also equally as important. And GitHub also, you know, and that environment, that whole community. is is photorealism like the end goal game or do you think there's you know maybe people would prefer an artificial character or maybe a brand mascot or something that you interact with do you think there's value in there being options around those things I mean I think we can answer that question around do you think Nike wants to have you know an animal greeting you at their site to have a conversation? Probably not Nike. Okay, well, you got what I'm saying, but name one, right? I mean, so, yes, I mean, I think it's all. Yeah, I mean, you know, I think that I know it's a serious question, but it's not, right?

Because the Uncanny Valley conversation is really because it's always sucked. It's never been good, and it's never been powered by something that can work in real time. And AI is real time right now. And the conversation is in real time. And the personalities are real. These models that we have created, these are real people that sound and look like real people. You can't tell the difference that it's a machine other than it's a machine. Yeah. If that makes sense. So, yes, they're a machine, and they're going to be trained to just have conversations around what that business wants them to have a conversation around. So that makes them awkward. So if you say, hey, do you want to go on a date Friday night? The machine's going to go, what are you talking about? Right? Yeah. So obviously, that's different than the girl at Macy's, but the human at Macy's where you could ask her on a date. But this machine isn't going to respond to that question.

So there's – that is uncanny, right? Because that's not human. Like we're human. Wow, you're very beautiful. The machine may be trained not to respond to that, right? So I know I'm just – I'm saying that is uncanny. That will always be uncanny when a human interacts with the machine because that's human behavior, not machine behavior. Yeah. Yeah, that's fair. Yeah. Something I'd like to talk about a little bit, too, is you mentioned F1 a minute ago and what you've done with them. And when I saw your booth at Build, this was all set up on a screen. And I was wondering if you could kind of explain to us what that project is, what you're working on with them. Yeah, it was great. We really tested out an event, the spa event, and it was really trying to see, okay, how would a customer interact with this machine, this agent that greeted them on the website and navigated the entire website automatically for them?

So, hey, I want to go to the race. I would like a VIP ticket. Great. And now you can have a conversation. What are you looking for? And bring it and literally add the ticket to the cart. Do the whole process for them. And we wanted to learn what was the difference between those customers and customers that preceded them just 30 days earlier or that same year prior to. And look, hats off to that entire, you know, to our client for sharing this data. You know, you have a lot of clients who are testing. many, many, many, an unbelievable amount of clients that are actively testing that don't want to share their data. There's some competitive edge. But it was just amazing that they did share it. And the results are unbelievable. I mean, absolutely unbelievable. They added more to the cart than ever before. They had an absolute, the attrition was significantly less. Now, what I'm going to say is probably the most powerful thing

that we've talked about in this call. And that is that we believe, and I believe a lot of our customers are starting to see, that market share today in their business is really back to zero. Even though they might dominate. Any business could have 100% market share. They could have 40%, 10%. But if they aren't actually following the steps that we're talking about of capture, learn, act, and own. Yeah. They've got a problem. And here's the problem specifically. If you look at the web, if a business is converting 3% of their customers that come to their website, which is a big conversion. Matter of fact, for the purposes of not having any debate here because somebody might say they have five. Let's say it's 5%. That's massive. a 5% conversion from the customer coming to a site, total customers to those actually buying. Let's just say it's five.

Our agents are discovering why, through Capture Learn Act, why 95% aren't buying on the web, specifically your website. So think about a CEO or C-suite officer hearing that they can potentially capture back 95% of the loss that they paid for already. So think about that. Of 100% of the people who come to the site, they pay for that. Yeah, they sure do. Every one of them. Fent marketing. You've paid a lot of money to bring them into your house. Here they are. They're on your website. And 95% walked away without giving you any money. That's right. And you now have a you now today, right now, go to Microsoft Azure, go build this right now. You can capture right now that 95 and that conversation is just now starting to happen where I think businesses are realizing, wow, we're sitting here trying to replace humans.

I don't know if that's really the way to go. I think we need to be using AI to capture customers. I think we need to be understanding and learning our customers, and I think we can capture customers. And that changes. That market erosion will begin to happen because if your business isn't doing what I'm saying, then you're being eroded by the guy who is. And it's happening. You can't stop this. And when we look across the number of businesses that we are actively engaged with, and it's over 2,000 companies around the world. We have offices all over the world. Every vertical is hyper-accelerating with our product because they are really understanding that 95 percent, that AI is actually blue sky, that they can actually capture revenue, capture customers, learn about a customer, get closer to their product. And so we don't see AI truly. We see it more as super business intelligence to provide a super customer experience.

And so we don't see AI in the same way maybe the models do where they're constantly looking to develop a greater super intelligence. We're looking to create a greater business super intelligence for a better customer experience. I like it. I like it. You mentioned – there's something from earlier I want to circle back on a little bit, and that's this idea of data sovereignty. if suppose I go deploy an agent what gets collected does anyone else have access to it and if I leave Napsure am I able to take it with me sure we sit on top of your whole infrastructure we just an API so everything underneath you own and that the point So can you say that about Meta No No way. Whoa, whoa, whoa. I mean, talk about a wolf's and sheep's clothing.

Can you say that about Anthropic? No. No. Right. So that's what I was saying about that Windows interface. We're simply an interface, right? That's it. Everything underneath is what you own. And then when you get to the cost side of it, it's not like you're investing a huge amount of money to actually have this conversation. When you think about, let's say the conversation with your customer lasts 10 minutes. Yeah. 10 cents. If you pay a dollar to bring a customer to your website, what's investing 10? And then what do you learn from that 10 cents? Well, you're going to ask them in 10 minutes. You can find out their favorite color, how many kids they have, what products do they like, What do they like about your product? Why didn't they buy? I mean, ask anybody in tech if they want to know those answers to those questions. And you're the one asking the questions based upon your approach to the market, your business strategy. I mean, that's a heck of a 10 cents. I got to tell you, I pay it.

It is. And you're asking the questions that you want the answers to. And that you own. Yeah. So it's investment. So literally, you're investing and getting answers. I mean, you could append to that same data. Go out and try to append the same questions. I need to know my customer's real name. I need to know how many kids they have. I need to know their favorite color. I need to know what other brands they buy. I need to know why they left my particular website or went to other websites. Like, try to buy that. You couldn't buy that for $0.10. Yeah. I don't even know if it's available. I don't even know my actual favorite color. I say it's blue just because that's neutral. But sometimes I'm like black or red. You know, I don't know. So, you know, when I think about Napster's original path of disruption around access to digital music and who controls that, as you shift into this same question around AI, this, you know, who controls it, who can access it, are there lessons you think to be learned from the original music industry upheaval that can apply in AI as well?

Well, I'm very vocal about that. And, you know, let me say it this way. I want as much as possible to not have these intermediaries between the creators and the customer. and so the music the music business went sideways when we put all these people in between whether it's a spotify whether it's this and they certainly they provide a service um and and and i'm a big fan of all of the industry as a whole but ultimately the creators get sort of screwed very you know and and and and so there's an opportunity when you think about i come up with an idea, I want to innovate and create on that idea, and then I want to bring it to market, how much of that process would cost me for AI to be in the middle of that whole process?

And so I think when you think like that, and then you think capture, learn, act, own, and you think about how you can build and own all of that, I think you're putting the creator, which in this case is businesses or people that want to create music or whatever it is, you're putting the creator into the ownership role of having a one-on-one conversation with a customer. Wow. How powerful is it that you can get rid of these intermediaries and start to shed that and to learn more about how you can talk? So, for example, in that one-on-one conversation that you have with that customer, you can ask them, what's the best way for us to talk to you one-on-one? How can we reach you the best? How would you like us to present the products that we have to you? How would you like us to let you know about discounts? And we can do that however works for you. Visit us. These are questions where you can find out how to best direct your marketing dollars towards reaching. When you think about asking that to a million of your customers, you're going to find out, wow, we've spent all this money here when really this is what they would prefer.

how this is their preferred way of us communicating sales or product. Like, but yeah, how many times, how many businesses really have asked every customer to their website, what's the best way for us to talk to you? How many? How many? Not many. Zero. Zero. They don't know that information. They guess. That's the whole marketing department makes up a bunch of crap, tells the CEO, I think this is the best way. We're going to sponsor this, do that. We're going to throw crap at the wall and just hope that I'll show up at the website and 5% buy. Yeah, and that's what you have to do when you don't have AI. When you're not capture, learn, act, own, you have to just guess and throw stuff at the wall. So that's a change I see. Imagine I interact with the same AI companion for some series of years, 10 years, for example, knows how I work. But then suppose that company shuts down. What should I have the right to? Am I dealing with a persistent agent software that I rent or a digital relationship that I should be able to move?

A hundred percent you should be able to move and a hundred percent you should be able to own it. And that's why the infrastructure layer is so important that you control and own that underneath. And that's what businesses are most worried about. You know, when we talk about us being native in Microsoft, both in Foundry and Azure, you know, this is that fundamental layer of businesses saying, OK, how many other companies do what Napster does? But how many have been approved basically by Microsoft? Like, that's a big question. Right. So I think there needs to be portability of intelligence. But I think underneath all of that is you have to invest in you have to invest in the infrastructure, particularly you. Right. As someone who has an audience who, you know, use you as an example, you have an audience, you've invested so much into that audience. You know, you could make a digital twin of yourself. You could then build and invest in that infrastructure. You could figure out how to continue, which you're doing a great job of delivering great content to your audience.

But now your audience have a greater response. in over 10 years, you've just built an unbelievably targeted persona of your digital self, your digital person. Hey there, humans. Quick update on our Neuron IRL event in San Francisco on November 18th. We are getting the Neuron community together for drinks, bites, special guests, and a live recording of the Neuron podcast. Basically, it's a chance to hang out with a bunch of people who are just as obsessed with AI as you are. Huge thanks to Slack, Alumni Ventures, and Sonar for helping us make the night happen. Now, if you're already registered, we're excited to see you there. Keep an eye out for more party details from our events team. And if you haven't signed up yet, join the wait list. If more space opens up, the team will let you know. See you in San Francisco. That's cool. That's a cool idea. What does success look like for what you're building five years from now, to the best of your ability to predict with a crystal ball? And what's something we're really missing with still thinking of AI as a chatbot?

I feel like those kind of line up with what you're doing. You know, we think so big. We really don't want to end up being more socialist. And we believe that, unfortunately, when you listen to how the large language models sort of see the world, it's very socialist. Like we just heard today someone make, and I'm not going to name drop it, somebody said, hey, this is going to bring us back to a three-day work week and we'll be able to be home and we'll be – and government can supplement you. The point is money won't matter in 10 years or whatever. It's just – I think that the reason why they're saying that is that these companies see that they're going to own our intelligence. They're going to own our human, our mankind, our human part of us. And and and and the purpose of us going to work gives us purpose.

Right. I want to get up and go to work. I want to get up and do something. I want to get up and motivate or or or help make a difference. And and and but that's what that's the goal that they have for the machine is for the machine to be our purpose. And that's that's scary. So so so by by. By allowing our customers to own, which is that core disruption, we're removing the ability for there to be silos. You know, and I think that's where Jensen sees the world. You know, hats off to him because he's saying something that's so unpopular and he's still saying it. And what he's saying is I want – and he's continuing to create products around individual companies being able to do their own compute, to create their own models. I mean he just bought a $10 billion, $13 billion model sharing company, like sending us a message that he sees that same world where he's got to unfortunately break up these monopolies.

If we end up with a monopoly, we go socialist very fast. Very, very fast. Unfortunate. It's just because ultimately our dependency will be – the only way to be competitive would be to depend on the model that owns information we don't have. And that information is on our customer, right? And so ultimately every business has a customer. It doesn't matter. Everything is consumer-based. The whole world drives off consumers. Even audiences need you and you need advertisers who have customers that are going to buy a product. So at the end of the day, it's all driven by a customer and a business product. So owning that intelligence is at the core, and we really believe we're doing that part. And you may look back 25 years from now and look at what Napster becomes, and I think you will see that we're beating the drum on something that we think is at the core of humanity.

We really believe that. I love it. John, thanks so much, man. Thank you. This has been a great chat. I really appreciate it. Where can people go to learn more about Napster today and what you're doing? Napster.com, of course. You know, same destination. Download the app, the Napster, just to play around and have a good time. And if you're a developer in the community, check us out in Azure. If you want to build a model in Foundry, you know, click the button and get an active agent out there. I think you'll find the results will be spectacular. Excellent. Well, if you're watching and enjoyed today's conversation, like the video, subscribe to the channel. It means a lot to us. Don't forget to pop by and check out the Neuron AI newsletter at theneuron.ai. Join 700,000 other people who read it every morning. Thank you for joining us. And that's all we have for today. So, farewell for now, humans.