20VC: Nikesh Arora on the Frontier Model Problem: Breadth vs Depth | The Future of Token Costs | Memory Becoming the Moat | Where Value Accrues: Infra, Models, or Apps? | Why Enterprise AI is Not Ready & Systems of Record vs Systems of Intelligence

I think the long-term token pricing should be one tenth of what it is today. Mitzus ended up, I think, ends up being an accelerant to cyber security. In technology, you missed one trick, you can survive, you missed two tricks. You're partly impaled, you missed three tricks, you could be obsolete. I came to the United States with two suitcases, $200, and I was willing to do anything, anything at all to make sure that I made a life for myself because there was no way to go back. When I came to the United States, I was a security guard, I took notes with the disabled, I flipped burgers at Burger King, I had $200, I had to find a way of being my tuition. This is 20 VC with me, Harry Stabbings. In the hot seat state, we have Nikash Aurora. Nikash is the CEO of Palo Alto Networks. They have a market cap of $225 billion. Nikash is one of the most respected operators in technology. Before, he was CMR Google of all things, and I questioned him on marketing in this show, and then he's like, "Well, I was CMR Google. God, there are some moments when I think Harry should stop talking. He's an old friend." But this was a very authentic and honest discussion, in a way that I don't think you could have had without that friendship. It was in person in London. Nikash here is one of the best that I've ever seen him. But before we dive into the show today, you have the idea, but often with AI tools, you hit a wall. Well, Base 44 is where that friction disappears, turning how you talk into how you build. Full stack, web, and mobile apps, sites, or autonomous super agents all built in minutes, not we can spend on damn configuration. Base 44 ships it all out of the box. The back end, the database, the authentication, and the hosting. It handles the heavy lifting, so you can just stay in the flow. It doesn't just replace the busy work. It multiplies you. It makes you so much more capable and effective version of yourself. In this market, being fast is the baseline. But to win, you gotta be first. And Base 44 is that edge. It's the move that lets you skip the troubleshooting and get straight to the breakthrough. Launch your next big thing at base44.com, that's base44.com. After Base 44 helps you launch, Korgi helps you cover what comes next. My world rodent arresting first line gets you are asked covered with Korgi insurance, and I'll tell you why, if you're running a business right now, you already know this pain all too well. Getting insurance, it's really slow, it's confusing, and my word, it's full of paperwork. Well, that's exactly why Korgi is here to change the game. Korgi is the first and only insurance carrier designed specifically for tech companies, allowing you to get covered in minutes instead of days. Korgi provides essential coverages for all growth stages, such as D&O, Eando liability, cyber, commercial, general liability, and more. Get your ass covered. I love the way we say ass with Korgi insurance alongside thousands of other startups at Korgi.com Ford/20VC today, that's Korgi.com, Ford/20VC, you won't regret it. While Korgi handles the coverage, cheering handles the talent. Frontier labs keep facing the same limitation. Models perform well on benchmarks, but they fall short once they enter real coding tasks, real tools and real workflows. That disconnect between synthetic evaluation and actual system behaviour is now a core block of foragantic models. That's why Nvidia and Thropic Salesforce Gemini and other leading lab partners partner with Churring. Churring is the research accelerator focused on post-training reliability. They build realistic RL environments, nest generation data quality systems built from real-world operational traces, and coding data sets that stress models under conditions where failures matter, state changes, workflow branching, brittle tool cools, and the coding errors that break RL agents but never appear in benchmark reports. In reality, a model may demonstrate correct reasoning in your evaluation setup, yet still select the wrong parameter or mishandle a code update in a realistic interface. Churring makes that failure visible and gives teams the signal they need to fix it. For labs advancing agentex systems, Churring provides the structure required to understand why these failures occur. To find out how, visit churring.com/20VC, that's t-u-r-i-n-g.com/20VC. Nickesh, last time we did a show, I was 22 hours into a 24 hour fast, and I listen back now, and I just think my word, you had the audacity to be with one of the OGs of this business, and be hungry, like I was, I was hungry with you, and short-term. Well, good news, as I told you earlier in the chatting, I had total memory loss, I had no recollection of what I said and what we talked about, so I had to go back to listen to it. I was happy to be here. It was a great show, we've put on this for a while, it was wonderful, I didn't know what that was about. I think we should stop doing podcasts remotely. God, I agree. I think you should force people to come in here. Well, it is. So I actually do. Good. And then I see about sizes, they do. And the only challenge is they actually sometimes come quite jet-lined because they come like all the day, and that's a little bit tough. Maybe you should do it 11pm. Maybe. I've got people from California, it's perfectly nice time for them at 10pm, maybe you need to. And maybe I need to adjust. Yeah, I'm the selfish one. I've never seen more countries that have something to do with me. More countries that I've really actually embraced it, haven't even got the first question if I could. You said, you need to embrace. How is it all your fault? If you embrace everything in life with how is it my fault? Actually, a lot of the world changes. Actually, let's spin that a bit. How do I make it better? Change the outlook. What can I do to make this better? That's how I run my day, and I run my life with my company. How can I make it incrementally better today? And how can I make it radically better in three years? Have you ever had something that you couldn't make better? Not for the lack of trying. So all we can do is try. If it works out, it's great. If you try a lot, you succeed more often than you think. We were talking downstairs about your personal brand, and it's because I'm making it better. I don't think about it like you do, because this is what you do for a living. I think about running my business. Okay, so I think this is fundamentally wrong. Look, it's funny. I clearly treat Harry. You know what I mean? I'm here to learn, teach me. The audacity of podcast. That's all right. No, but I actually think your personal brand is your business. I think if you build a great product, a great company, and people like your product, eventually your brand survives all of it. I think you can have a great brand, shitty product, shitty execution, and your brand goes to hell in a hand basket. So I flip around. Do you think that is still the case today, though? I think brand can be such an excellent today, especially in a wild bar. It's just so noisy. See, I spend many years at Google as you know. I spend actually was cheap marketing offers for five years. And if you look historically in technology, there are companies which have died who had great brands, remember Sun Microsystems? They're drawing, you know, 30 years ago. It doesn't exist. Why? Because their brand went dead. Well, the product went to hell. What about Yahoo? Remember that company? It's a great brand. Amazing. Why it's way before Google. I think it's probably a fraction, probably even a two decimal point number versus what Google is. So I think product helps make brands. But I think those were founded and broke into the public diaspora when there was much less noise. And so I think if you were to put that today, you need to have a brand first to get into. Look, let me speak against my own thesis. There's a spectrum. On one end of the spectrum is only have real differentiator product. In which case, the product helps build the brand. Google search is now Google something, right? On the other hand, it's a commodity. It's water. Not on a white, this is called Evian. But this is a commodity. Here, only brand matters. So it depends where you're on the spectrum. If all you are a commoditized product, yes, brand matters a lot. If you are a differentiator product, then you build the brand of the back of the differentiator product and you can decide where you want to be in that spectrum. Speaking of that brand, I saw your tweet and I was like, this is masterful. But you tweeted, the frontier model problem is a brat versus depth problem. Yes. Can you explain that to me? Yeah, look, I've been listening to some of your podcasts and I've been paying attention to what happens to market because I want to understand where all this settles down, not just because, you know, I want to understand it, but it also impacts how I build my own business. We got these phenomenal frontier models and they keep sort of, they frogging each other every few days. There's a net new model that's delivered by OpenAI or by Google or by our friends at Anthropic. And the question becomes like, if I and these models are moving in this space, what do I need to build? What do I need to do with these models? And as we came with that mid-haust moment, when everybody was busy chasing mid-haust and that was kind of important, you realize even the best model has a high false positive rate. But for some reason, in the consumer space, we don't seem to care. I was talking to my sister this morning, saying, I just went to chat GPT and asked all these questions. It was very helpful. So I guess like, what happens is the consumers are way more tolerant of false positives because just kind of always the person in the middle, right? There's always somebody who's understanding what the model says and making their own judgment whether they believe the model or not and some happy people get rid of some false positives, some happy people don't care about the false positives. Sometimes, you believe the false positives. So the consumer is highly tolerant on this notion of false positives and doesn't seem to distinguish and it just seems to get better and better. I literally had Gemini produce an investment memorandum for something I was looking at. I looked pretty accurate and I give or take a tweak a few things, but it seems possible. So on the consumer side, it's a breadth issue, right? It wrote an investment memorandum for me, which is cool. I had to hire a banker and a bunch of investment analysts had taken me days and I didn't for minutes. So the breadth is there, which means it's my go-to place. And as you know in consumer, if you become the go-to brand, talking about brands, it's hugely beneficial, right? Whether it's YouTube, I've touched the only place to go look for a streaming video or Google has the only place to go to a search, it becomes hugely multiplicative from a distribution perspective. So our frontier model friends are chasing the consumer brand and the false positive doesn't matter. On the enterprise side, false positive doesn't matter a lot. They matter because if you imagine a future where an agent's going to make independent decision. and act on it, you have zero tolerance for false positives. Now, take Waymo. My view Waymo is the biggest agentic product that is out there because guess what? If you replace a human being called a driver, right? All decisions are made by AI machine learning. It decides when to turn around to stop what to do. But think about the amount of edge case training it took to replace that human agent with effectively an AI driven agent. I don't know, tens of billions of dollars. So that's what it takes to take one use case and train the hell out of it. And if you think about what happened there, they could have used the equivalent of an AI model. But didn't they build so much context and intelligence and edge case training and proprietary data to make that happen? That data is not available in the internet. You can't stick the next model of an anthropic into your Mercedes and say, "Okay, drive me home." It's not going to be able to do it. And that's the depth issue, right? Because you need the depth of the context and understanding and the intelligence in around the model to make it useful for the truly agentic use case. So you think there's this constant tension. The frontier models want the consumer attention because that drives both training for models. It drives the consumer brand of the model. On the other hand, the real enterprise revenue is going to come from use cases that are required a lot more context. The one sort of stand out use case, we all know is coding. Coding is a universal activity. Every does it. So every data is helpful in training the model. And that becomes a large enterprise application. Hopefully there's a few more out there. But I think that's the tension that I wrote about. When you think about the workflows and the way the enterprises are engaging with AI stand specifically models, what extent do you think we will be locked in a frontier model dominant world versus a the majority of enterprise workflows can be done with open source. And we will be more cost efficient moving towards that most of the time. I think more than half the enterprises are still not getting it right on the use of AI perspective. I think we're still busy trying to incorporate AI into our current business practices. So how do I take what I do today? Use a little bit of AI, get marginally more efficient because I don't want it to be the old way. I think the opportunity is to rethink your workflow fundamentally with the AI. That's where the true benefits going to come. I think the winners in the long term will be people who actually rethink their companies with the AI, not people who adapt their current workflows marginally with the AI. How can you do that if you're an enterprise? So we have thousands of CEOs of big enterprises that listen. That sounds great. What do I do? Do I do brainstorming session? No, I think there's going to be perhaps true to different categories. One category is let's take the workflows of today. Then we have workflows around ERP. We have workflows around sales team management. We have workflows around human resource management. There are existing workflows which have been satisfied. As in some software company decided, we all have common processes. Let me build a container where these common processes can be marginally customized by enterprise. They can code their workflow into my SaaS application and we're off to the races. Now, that workflow required a lot of human judgment and human interaction. The software is not intelligent. It's been coded, right? You define input, you define output. I do the input. I know what the output I'm going to get. The idea is, imagine workflows where, in the hiring process, most of your workflows are containers. Imagine where AI actually is helping you make judgments. If I put every CV into AI and say, these are 20 people in the inner view, this looks at the CV. You should ask this person these following questions. Send the note to Harry saying, "Interviewing this person asked the following 10 questions because it's three other colleagues only asked the following five. We needed no false positives to human beings." So AI could be hugely helpful in informing and making the process more intelligent. But that requires us to give up human control and let AI do 80% of the thinking for us. I was not how we're doing it right now. All we're doing is, let's take this invoice, let's scan it, abstract the data, put it into AI and say, "Look at that, it's happening 20% faster." Do you think we are willing to give up that human control? You see cases like Matters Day, where I think 1700 people have signed a petition that they are unwilling to let it track their mouse and key stroke behavior. We are seeing resistance to giving the data. Our total of points. I think two different points. I think mass collection of data to inform AI is a little dangerous because people see the outcome of what that's going to happen. I've heard of companies where people are using cameras to track people, folding laundry and ironing clothes because they want to be able to raise train physically in the future to do those things. So that part of the side. I think on the enterprise side, we can actually run a business much more effectively and efficiently if we decide where we are willing to relinquish control to AI. You're a perfect example, marketing. Anything that is required to train a marketing model is already out in public domain. By definition, marketing is public domain. If you didn't market it, it's not in public domain. So I have the best training data in marketing. I don't need to train any AI model with more marketing content. I may need to train it for tone of voice and what my brand is. And I'm pretty sure an AI model, if I throw my marketing collateral into it, will tell me this is not consistent with your brand. If I look at all the things you've been talking about the last 10 years, I think some people have done that. They've actually analyzed earning scripts of companies and said, "The last 20 years, what is company X done?" And when does the CEO start getting away from a topic because perhaps it's not going so well? So this is really smart. You can do that stuff. So it's the best marketing training database in the world. The frontier models. Why do I need 400-600 people in marketing? Because my biggest problem in marketing is I 600 people, but I'm not sure. They'll fully understand how to consistently deliver my tone of voice, my value, proposition, and how not to break my brand by having different collateral than public domain. You have 600 people in marketing. You ever take 21,000 people? Well, it's not going to be 600. No, what is it going to be? My rule of thumb is that in the next three years, we'll probably have half the people in G&A type activities in companies. Things like marketing, things like finance, things like HR. Because there's a lot of process management there. And a lot of process management can be made more intelligent using some version of an adapted future AI application for life of a better world. So SaaS applications will give way to AI applications. The difference being, SaaS applications have no opinion. AI applications will have opinions. And that's a fundamental rethink we need from a workflow perspective. Can you just help me out? What does that mean in terms of how you use them in terms of the app that they have? What would it mean? Everything, right? You're a scholar, what do you want to call it? A near assistant, a marketing assistant, a HR assistant, say, I look to your copy, it sucks. It does not good enough. It's not consistent with the tone of voice. Here's what I will recommend. That's an opinion. That will make my average employee much smarter than they were today. Then I don't need so many of them because they're doing most of the work for you. What would you say to the people that say, you're wrong? We won't see that halving of those functions. And actually marketing teams will create more copy, more content, being more places. I think the places where people could be wrong is how many technical resources we need in the future. I think we need more or not less. I think there's this fallacy people believe we're going to have less people working because AI is going to take over our jobs. I don't believe that. I think what's going to happen is you can't imagine the number of people on my team who want more technical resources, more AI savvy resources because they wanted exactly these things. They say, well, I've got an amazing project to transform marketing. I've got an amazing project to transform HR. What do you need? Well, I need more people to understand how to prompt frontier models, build harnesses, bring proprietary data into play, bring modes. I need more compute, more storage because I want to learn everything. So I think we're going to need more technical resources. I think we're going to need more sales resources. Because if your product's really good, you need more people to go out there and cover the universe because not enough people know about it. I'm in Europe. I met 20 customers last week. I still see half of them don't know all the stuff you have. I'm like, dude, we've been around for 20 years. Why is my team not out there pounding the pavement, telling them everything we do? And the problem is, not a time and a day, because I'm too busy dealing with some arcane piece of software at work, which I have to go feed. Well, if I had that software be really intelligent, it'd tell me what to do. In terms of wanting more technical resources, tokens I would put in the more technical resources camp. How do you think about effective token allocation stay? You're seeing very different camps from your matters and newbies and my resources you put in budgets to your free-for-all be creative. How do you approach it? I know the soil of the work of docking maxing is going to gone topsy-turvy with the whole composition that our having to hook people using. The challenge right now is 90% of the enterprise employees are not AI savvy. They're not. They have to learn. I can't send them to the university, but there's no course you can take in any school anywhere. They have to be able to learn of their own. I think we're back to a Darwinian moment where everybody has to figure out who's really good. Now, you've seen people like Brian Armstrong and Jack Dorsey go out and say, "I'm going to decimate my organization." I'm going to start building from scratch and they've gone to some version of 30, 40% less people because they figured out there is no redemption. I can't train these people. I'm just fine. The people who are going to come in and help me do this stuff. That's one model. The other model is sort of gradual. We've been hiring people only through hackathons now. Right? And we see natural attrition of 2% give or take a month. And we just replace them with people who actually are AI savvy people who are from hackathons. Give me 12 months out of the transform 20, 25% of my team. Give me three years. I have hopefully enough AI savvy people working at Ballot. So there are two different ways to get there. I think Ballot, what you're seeing in the token-maxim world is people are learning, people are experimenting. The risk is your smartest employee who knows how to use AI really well. Could be using 20 times of tokens that an average employee uses. And if you get into this whack-a-mole moment saying, "Oh my god, I'm going to stop people spending too many tokens." You actually will hurt the best AI savvy people more than you will hurt this average employee. And by the way, I think the best talent will want to go where they will be best equipped with the most advanced frontier models, the biggest budgets. I think that will almost be like an employee benefit. Yes, possibly, but I think part of the challenge is, right now everybody's experimenting and everything. And you have to figure out what is it that I need to build as an enterprise and what can I get off the shelf. All right, if I can get an AI-based tanking application that does marketing for me, I don't need to build it. It's a generic problem. Everyone needs to solve. I can tweak it. I can customize it just the way I did with SaaS applications. But I don't need to build my own. from scratch. So I've made sure that everything my team is building is proprietary to us. So where do we have unique distinguished knowledge that we bring to bear, which nobody else can do on the outside? Let's put that, let's package it, let's use it. Where it's going to be a generic application, 12 months, 24 months from now, let's just wait. So you have a free-for-ruled on token, say, to allow your team to do the best. We have a used judiciously model for tokens. It's not a free-for-all. Free-for-all sounds like you can go token max to halal, but we haven't used it judiciously. And we keep track of it to see what people are doing. And if we find somebody who's using it well, we want to constrain them. If you find somebody's gone a little over the top, we'll find a way to. Benny Off has said the other day that he's been 300 million on anthropic a year for his devs. And that works out to be about 3.8% of developer sound respawned average. If it stays there, the valuations of anthropic and open AI are grossy overvalued. And if it moves to 20%, they're actually very undervalued. And if it becomes what Brandon McCour said, which is, we'll spend as much on tokens as we do on salaries, they're grossly undervalued, grossly undervalued. I want to talk about where you think percent of developer salary spend on tokens will be in three years. There's still a narrow lens for me. If I abstract, if I step back today, there's not enough compute for what the world is demanding. Unequivocally not enough compute. Can't buy compute. Compute is costing two to three, X of four X more than it used to cost two years ago. There's not enough compute. That scarcity of compute and that excess cost required to build and deliver compute, which allows us to go make AI useful, is causing the constraint and forcing pricing. And interestingly, more than half of the compute is going to feed the consumer, which is a fundamentally lost making entity right now. I don't think any of the frontier models make any money in trying to get you and me to use chat GPD or you know, cloud or Gemini every day. It's free. That's a lot of compute. Imagine there's billions of people around the world using for all kind of queries every day. That's sucking away half the compute, which is making the return. Guess where the pressure goes. The pressure goes on the other half of compute, which is being used for coding in enterprise applications. So now you're saying enterprise applications coding have to pay until we build transaction models or advertising models on the consumer side because they're not ready. Now you can say, well, that happened in search too, right? Google search was around that happened YouTube. People used a lot of YouTube, a lot of compute, but it wasn't paying for itself. The problem is the computer requirements and the cost is now 10X of that when we were in that era versus today. That's forcing token prices to go up. I think the long term token pricing should be one tenth of what it is today. But that happens. You will see that people will consume more. You can decide if 3.8 or 15.8 is not not sure we can tell the answer right now. This pricing will move very drastically in the next three to five years. Sorry, so I'm saying you say, even we'll see dramatic reductions in token price. I think so. I think in the next three to five years, we will see reduction in token pricing. I think at some point in time, the consumer user via will get constrained by these frontier AI companies because they have enough post training data more than they need and each user is inherently unprofitable in their activities. They do on frontier models. Do you not think they just build advertising engines light open eyes doing now to pay for that business? You know, that's an interesting question. It has to come from somewhere. When I started Google in 2004, we were 2% of the global advertising revenue and global advertising revenue estimated between $500 or $600 billion. I think online is about 70% by last count of total advertising revenue. And I don't think the overall number has changed by more than 3% a year or 5% a year. I don't think the total advertising pie is going to increase. You've already taken away 60, 70% of the advertising pie in the online world. Unless you tell me it's going to explosion at the top where more people are going to spend more money in marketing, that money that you're hoping to fund consumer AI from advertising will have to come from current advertising revenues. So I don't think that changes the equation drastically to make the consumer profitable. I do think there's an opportunity that AI ends up taking more transaction revenue, which has not been into the purview of AI. Well, think about the marketing chain, right? We do advertising advertising is inherently efficient. What's the best conversion rate you get in online advertising? You think one one and a half percent? That's fine. The thing the best of breed is 7 to 10%. The average probably 1 and a half to 2%, which means 85 to 90% of marketing is wasted. So now if you get really smart, you have memory, you have context and you get smarter and targeting Harry when he's trying to buy something out there in the world, then your conversion rate goes up. If you look at the entire value from the time you decide to buy something till the end, you get the product. You take the case of consumer goods. Today, I want to say the cost of consumer goods is probably in the five to eight percent of total list price. The 92% is distribution and marketing. It's highly inefficient. So you could imagine a world where AI makes marketing really efficient and you get more dollars coming from traditional marketing into the online world because coming the form of transaction. If tokens get cheaper, why are we not seeing front-end models get cheaper? We will think that this would be cheaper. This would be cheaper. Right now they're figuring out that all your front-end model companies are value-maxing, not talking maxing. They're raising money at a trillion dollars. At some point in time, they realize, oh my God, where's the next hundred billion dollars that compute going to come from? Financial markets are not going to bear the cost of another hundred billion dollars at a trillion dollars or a trillion and a half because I'm going to need again the following year. So you say, well, I get a build a sustainable business model. It starts showing some degree of gross margin profitability. The only lever they have is to take the fastest-growing thing that they have on their portfolio from an economic perspective and charge us more for it. That's where you get the price of tokens from. I think the price of tokens I now you can imagine is the price of tokens high. Every technology is trying to figure out how do I make my compute more efficient? Right? In the future. So I'm sure we'll see a whole bunch of advances in the world where memory and computer can start getting used more efficiently from modeling perspective. I still believe I don't need fable five or mid-tose five to do 90% of what people do in the eye today. Why is the last model not good enough? Certain tasks. Having an unwavering needs. I think the model overhang in terms of capabilities and especially consumer and most enterprise model. But the models from two years ago were good enough for most of the queries that we all understand. That's right. The problem is they were inefficient from a compute perspective totally. So you're seeing the efficiency come in. The problem is the cost of R&D is now being spent in terms of what the tokens have to pay for. So I think the token prices come down. I think the amount of compute that we need is going to be huge to the next ten years. I think the frontier models are in a position to capture a significant amount of the future economic value of the user AI. Everyone is fundamentally looking at the stack and this is me being very open as a venture master. That's why the show has been successful. And I'm just saying what every venture master feels. We're all looking at it again. Oh Jesus, I have no idea where values are occurring. And you've got, you know, you've essentially got info, which is going to top. And then you've got models and apps to be totally. Yes. Look, making money, infrastructure is more expensive than it's ever been. That's you're seeing trillion dollar market caps in the infrastructure because of the scarcity of compute and this need for speed. Do you think we're in an infrastructure bubble like people think? What's the just? I have a question which I don't know the answer to and you can tell me, so you spend more time with people here. I have to go to my day job is at what point in time does physics kick in and we just can't produce the compute as fast as we want to? Like infrastructure people are gearing up for large amounts of capacity, large amounts of demand on the infrastructure side. And then one of us says, you know what? It's only so many data centers we can build. It's only so much energy. Well, I mean, now I mean, I can show you as a copper. But do the panthalassa which is building data centers that see or eat on building them in space. Yeah. So I think there may be a digestion period at some point in time when we think the demand for compute is there, but the capacity to execute is now limited by physics. And the infrastructure players have built up too much capacity for this demand. I don't know when that rationalizes. Maybe rationalizes and causes us to go think about a different sort of time frame for putting all this compute out. That doesn't take away for the need of compute. Well, still want as much compute as we can deliver as faster as we can deliver. I think some of the model companies have outstripped anybody else's ability to build frontier models of that capacity and that speed in the training. So I think you are seeing perhaps a settling down or who's going to be the frontier model player in the future. The question becomes in the economics, what value of crucible model, what value of crucible the application layer is you said it. And I think the application layer is probably is a simplistic term because for the first time you have memory and applications, applications understand context, understand context as specific as to what you want what I want. You're seeing that in the consumer space that has not yet come to the enterprise space from the enough. That shows up in the enterprise space, which means the demand for compute and memory goes up on the enterprise side. Like how many I haven't I don't haven't used all the coding models myself, but over time, coding models have to get really smart about understanding individual context of enterprises and humans. That's how they'll be more effective and more efficient for that. We're going to still need more compute and more memory. So I think that'll start defining where the value of cruise. I think the value gets shared between frontier models and the context that gets created in enterprise play. I mean, the frontier models are fully understanding that this is where the gap is. I suspect the frontier models as a crystal ball, they will spend a lot more time the next year or two building memory around consumption. What do you mean like expanding context windows? More than that. If you look at the consumer interaction that you have with your favorite favorite frontier model, right? It's starting to remember, oh, you asked about this yesterday. You asked about that. Should I take the question? Just ask me in the context of everything I know about you or should I just limit the answer to as if I don't know anything about you? Now that having context of what I said to you the last 30 days or last 60 or 90 days requires you to store a lot of information. It requires a lot of personalized interaction that needs to have. Now if you want to maintain your mode with Harry and Nickesh in the future, the more context you have about me, the easier it becomes for you to give me the answers in the future. And as you start building context on a user base, it creates stickiness and that becomes your mode. To what extent does mythos cannot be? applies a business for you towards the standards and make a business for you. Mithos ended up being an accelerant to cyber security. I think what happened when he saw Mithos came out. It demonstrated that all the training we've been giving these models on how great code is written. The models are able to turn around and say, well, I also know how to find bad code. So what happens is you point the gun the different way and the model says, oh my god, look at all this code that you have. There are so many flaws in it. As you talked about the challenges like area model, it also suffers from false positives. If you're an offensive actor and I point the model against, must say, 20 BC enterprises and I go look at everything you've done, somebody left a web socket open, somebody's done some mess up with IP addressing, etc. So it finds the flaws outside in that allows the bad actor to go figure out how can they daisy chain vulnerabilities and get into infrastructure. It's not good enough from a defensive perspective because I can't use the model and say go take every vulnerability you found and build a patch and go patch my system and protect me. Look, that's what it's going to patch 30% things that are not wrong. Who knows what that's going to do to blow up your infrastructure. So when with those game, we looked at it, we treated with respect, we ran it against our code, we discovered it finds bad stuff much faster than human's can. We found in six weeks where it was five to six years. We got it. We ran around, we patch it, but a cloud code helps build the patch, but you have to run it through human evils, through testing, through production testing, to signboxing, to see does this patch break anything in infrastructure? Only then after six weeks we were able to go patch everything. So what does this mean? This means that every enterprise better fix their stuff faster because if I point the next generation of models against an infrastructure, it's going to find security flaws, security vulnerabilities, misconfigurations, things that you've not been paying attention to. So it creates a bit of an urgency on the parts of the customers to improve their cybersecurity posture, which I think generally is a good thing for cybersecurity companies. This is not fundamentally bad. When we just say, some are what you just said, it allows you to weaponize bad actors to find holes, but isn't good enough to provide the solutions. Well, look, the solutions are there. The challenge is getting, sometimes getting the attention and focus of the customer saying, let's say I got to go fix my stuff because it's important. What this has done is let a fire under the security practitioners are under the world saying, this thing is not good. This is going to weaponize the bad actors. I better make sure my defenses are in place. Now remember, the way cyber defense is done is it's fundamentally just some security is too fundamental things, right? One thing is, if it's bad and I'm at the gate, I'll stop it, which means you have to have somebody at the gate. Now, we have 150 million sensors in the world where we stand at the gate protecting our customers. If I can find a way of infusing, yeah, at the gate and taking all these vulnerabilities and finding a way to protect you, I'm good. I don't have to change the gatekeeper because there's no clawed endpoint agent that exists out there. There's no opening an endpoint agent that exists out there that I can replace by a lot towards the other people in the space with. The problem is not at the gate. What happens is, despite all the perimeter defense you put in, things come in, things leak in, people make mistakes, people passwords get breached, their vulnerabilities people get in through. Then the question becomes, all right, oh shit, the bad actress in my infrastructure, how quickly can I find them and get rid of them? That becomes an AI task. That becomes the same conversation I'm having so far is I need context, I need intelligence, I need to know what this means. So creating the context that intelligence within the enterprise or what this intrusion means and how to protect against it becomes a challenge. This is the AI-Scibe Security Challenge. You know, again, this is not trying to hear to pitch my book, but we spent five years trying to build that capability inside enterprises. So in the net net, it ends up being excellent. Does that mean I have everything I need? Not everything. Does that mean I need to get AI models to start helping me? Yes. So we're going to infuse more AI into our defense infrastructure. Do you think it is good or bad to have government intervention when you have models as powerful as we have them? I think we're going to a discovery process. I think this notion of Godrails has not been built robustly enough because these models seem to be easy to get past. Remember the early days when you used to read about somebody had this conversation with a model and found a way to, you know, albeit the Godrails because it asked questions differently. The model was able to get sort of jail broken and became a hobby amongst people and they had all kinds of conversation with models. It is the same challenge. How do you make sure that you can put enough Godrails around AI model that you built to make sure it's only used for the purpose that you had? That's the challenge that there is. I think the Godrail needs to get better. To the extent the government feels the Godrails aren't robust enough, it's trying to tell us that it's a national security issue. You need to go fix the Godrails. But the thing is, it's a tip of matter of trying to fix, to treat the Godrails as a real problem is all that. Is it possible? I'm hoping it is. I hope it is too. If he was taught and this was, I swear to one of the world's best cyber ambassadors who were my name list because he asked some spicy questions to him. Oh, okay. He asked questions from me. Yeah, he asked questions to me. He asked some interesting ones. But he asked if you were to start power out in that vice versa, a company again today, starting today in the age of AI, what would you do differently that you're not doing now? The paranoia I have, if I look at self-driving, there are broadly two or three approaches out there, right? One was, my car is not going to have a human in it. It's called WIMO. I'm going to keep bounding it. I'm going to keep bounding it, training it until it learns by itself to drive and no human is should be holding this steering wheel ever when my customers are in it. When you see it's out there, as many cities have WIMO, I saw one done the street from here. So that's one way of doing product development. What you do is every edge case gets discovered. You build training around it. Every experience is a learning experience. You build training around. You keep training it. Get to a point of total autonomy. The other version is, I'm going to start taking segments of the driving and start automating that segment where I get really comfortable. So my car drives 50% of the time. The other 50% of the human gets it on the edge cases. That's my Tesla. I used to drive just a highway for me. Now it's getting better at other streets. It's slowly getting better. But still I'm holding the steering wheel very often. It'll tell me you're not paying attention. Look at the camera. Otherwise, you can't drive. And I know FSD gets better. But that's another way to get there. I'm going to keep training and fix this. I'm going to start working the edge cases as my business continues to evolve. The third one is, I'll infuse some degree of self driving into my car because I've come from the traditional model of having amazing cars, eight V8 engines, beautiful sleek cars, and I'm not into the technology aspect. Do you see them out there on the street? We have all three versions out there. My fear is, do we all need to stop and start thinking the way more way as enterprises or is the room for the Tesla approach to self driving in our businesses? Because we have an existing set of customers to satisfy. We're not going to take kindly from me saying, you know, guess what? I change my product. It's right in 80% of the time. And I'm going to take those three years to train the product to be right 100% of the time. And am I pivoting fast enough in my product strategy? That over time, my products become more self driving than they are today. Or do I need to go faster? And can I get there by automating or AI enabling certain parts of my product where I can apply the models that are capable to do certain aspects of flabby security today and keep doing the others from machine learning and managing S cases edge cases to machine learning or is it time for me to pivot? My view right now is you have to have the Tesla approach if you're in enterprise that is building AI in fuse capability. But you can't have the approach of traditional car manufacturers which are trying to stick a little bit of AI and sort of, yeah, I washing their cars and saying, I'll get there eventually. Would you like to do the Brian Armstrong Jack Dorsi? I often think a good question is what would you like to do? But you'll help back from doing different courses for different courses, right? I don't think in our business we can go implode the organization because I don't think the underlying application software is there. I don't think all the things that we've talked about that we need to get done are ready from an AI software perspective. I don't want to build a lot of software that is proprietary to me for things that should be available for everyone. I don't want to build an AI marketing stack. I don't want to build an AI HR stack. I don't want to build an AI ERP stack. I'm hoping that somebody goes and does that much more effectively and efficiently for the world at large. Perhaps it's the next iteration of Salesforce, the next iteration of SAP or the next iteration of Workday that is going to help me do that because I do believe it needs to become more intelligent. We talked about needs to be more AI enabled or AI controlled. I do want to build things that are particular to me. There I have all the information, all the intelligence, the context and the memory from an organizational perspective. I think the right way to get there is to hold people accountable and I run a meeting twice a week now called AIO. It's kind of funny. It's like all my doll had a farm. Yeah, yeah, oh, because everybody in my company wants to do AI. So if you use it as a converging function, as a converging to do brain storming across my team, how do we think about this? How are we building this? What happens in that meeting? Everybody comes and shares how they're adapting to the new world of AI. What are they doing from a product development perspective? How are they thinking about it? How are they including agents in their products? How are they going to go build the back-and-infrastructure? How do we think about how are they using tokens? How do we think about the capability of resources? Remember, for me to transform 21,000 people organization, I have to get the hearts and minds of the leaders to make sure we're all swimming in the same direction or pulling in the same direction. So this is my way of ensuring the top 15 or 20 technical leaders of my company are pulling in the same direction. What would the direction be? And as you know, if there's no expert, then a group of smart people do better than an expert. I don't think there's an expert for future enterprise design yet in terms of here as the blueprint. Like you said, as if you see you're saying, "Well, these shit words of value are going to accrue, is it going to be in models, is it going to be in the application there?" Trust me, we have to have a point of view on that stuff and what's going to emerge in the future before I can start re-transforming my company. How should I build my products? Are your leaders AI-pilled? I interview CROs as well, some of the best CROs, some of the best CPO's. And in all honesty, the bigger the company, the less AI-pilled AI-maxed they are. To the extent that we won the other day, a public company doing 5 billion plus and around New CROs, you know, we don't have that AI-telling internally, but we'll bring it in. Yeah, look, I discovered this in 2004 when I was in Europe and I googled Europe, I used to go around with COs. And there was this FAD where COs would hire this 24-year-old Sherpa. They were called the AI-pilled AI-maxed. web share paths or internet share paths. It's like chief internet officer. Remember, those companies had chief internet officers at one point in time. I was eight. So now you don't remember what they'd use. We've seen this movie before. Sometimes it's fine to have seen this movie before. And the task of the chief internet officer was to make sure the organization was ready for the internet because I'm too busy in my traditional business and I don't know who Amazon is. I don't know who Google is. So this wonderful 24-hour who understands this stuff is going to help be my savior. And then CEOs would wash their hands about the internet because they have this wonderful team of people who would be frustrated because they can't get anything done because nobody's giving their attention. The risk of that happening is true with AI as well. I'm so busy doing what I did yesterday. I have no time to think about tomorrow. So meet my chief AI officer. It was probably a researcher at some amazing university before and has low execution skills. So until I can get my leadership to understand and agree the extent of the AI challenge and the AI opportunity we're not going to make progress. What specifically are you not focusing on today because of the burdens of today's problem? Look everything when you go to our product manager any large company I'm pretty sure they have a product roadmap that exists in their heart in their hands. It's six months or 12 months long. I can go fix all these things. I'm like that's interesting. It's in a six to 12 month thing. There's nothing called agents in there. I come like the world is talking about identifying everything and your product roadmap doesn't have that. I'll get to it once I get this done because of what customers want right now. I'm like no that doesn't work. How do I get you to do more agentic work? How many people are you going to free up doing development the way you're doing it today so I can use that access resource to go make new things happen. So those are all important conversations. I can have them one at a time across 14 or 20 people or I can have them twice a week with them and people demonstrate how they are and those fascinating remember you have to make sure your leaders are ambitious. You have to make sure they're competitive. You have to make sure they want to win. You have to make sure that they have a learning mindset. When they watch their peers around them do cool shit. They want to show up at cool shit the next time. So for me it's getting 14 people together saying hey Harry tell me today what have you done for AI in the last three days since I last talked to you in your organization and whatever motivates you whether the fear of Nikesh asking you three days again what you did or you're sort of getting a learning ambition or it's your team pushing you you will show up or something. Then you'll see what the other guys are doing. I'll say oh my god I'm doing a lot or I'm not doing enough. So it creates a little bit of Darwinian competition amongst them. It creates this urge to go embrace this new technology and I think hopefully I get 14 people fully motivated and then they go to that with the next set of people because I need to transform from the top down not from the bottom up in this topic. There's a bottom up experimentation of course right. People using tokens to see who's really good at that allows me to find the best talent. So you've got to find a way of transforming 20,000 people over the next two years in that direction. Bottoms are top down. You got to get into these organizations. That sounds normal. I wish I was a VC. I just don't know what that was. Clearly I need to get out more when that's. I just sit in this dark room all day in a catch. I'm sorry. The question is how do you get in effectively and how do you get implementation and adoption done while I've had guests on the show say before you cannot do enterprise adoption without after ease today and then I've had my tank come on the show off from factory and say if you need FDs you have a shit product bold and I love my tank and I think it was a great clip. So grateful for the validity that came with that. Okay. Is that that's one of our? That would do very well. Yes, I am from Palantir then obviously Chime Derniff. Of course. My job is to create a discussion. What is true and what is right? Do you have to have FDs to sell into enterprise? What is true is we've only been chasing the enterprise's dream for AI for the last 12 months at best. If you think about everything that happens on a weekly basis we see new things come which we don't quite fully understand and grasp right. We're all busy trying to get our arms around LLM's and how is there going to be great for chat bots to talk to our customers in enterprise and suddenly agents showed up. Oh my god. I got to figure out agents. I'm going to start working internally agents going to do a lot of stuff. I'm pretty sure you could still have a agent fest and have everybody tell you what an agent does is still to walk out and say I'm not quite sure that his agent or her agent is the same as with the last guy said. So because AI is moving so fast I don't think the products are fully there yet like the enterprise products that the application where it don't exist in their entirety because we haven't invested against the enterprise's ask. So FTE is a short form for saying my product is not fully there because it's evolving as the technology evolves. I'm going to send some people across who are going to sit in your office and build my product while I adapt to your needs. That's what it is right. That's what we saw from Palantir that's what we've seen from all these companies. So what you're seeing is I'm going to send my product engineers or developers to your enterprise. They're going to build my product. If you get it if you do it right again FTE is a different version. Some people are just trying to get you to consume AI which is actually not an FTE. It's just a technical sales consultant who's trying to help adoption. On this add an FTE truly is somebody who actually brings the code back from a customer site and goes back to your product and I say listen I built this at the customer site because they had this need. We should incorporate this into our product because everybody's going to need it. That's an FTE in my mind. I think FTEs are needed for the short term because remember all the enterprise AI startups are hungry for revenue. For some reason we've created this notion that don't worry just keep selling it. There's a huge sort of bent up demand around AI applications sold it before the product is fully ready. That's what we're seeing. Do you think that's right? I think that's the case. I think as we think things evolve in X12 or 24 months people will switch from one set of products to another because something will emerge as a better product. Look at the coding conversations. How many coding companies have you heard of the last 24 months? I think we had Windsor, we had Devon, we had which is now cognition. Windsor got sold. Those are the early guys in coding. They don't exist in their den form. Now you've got codecs and plot and anti-gravity, factory doing SDLC. You've got cognition doing SDLC. You can see as the market evolves people who have concentrated on different parts of value chain run coding are getting formed. The product is getting more and more formed over time. Who knows in two, three years is going to be leader in that space because the product's not fully ready when it's start. You want to make a bet on here? Well, no, you do that. I think you're a good one. My teams can use. I need to make a bet. How do you choose who you decide to help? I spoke to a daughter, Aisha, before. She said one thing that I should lot of things that many people don't know about you. She said, "What is you help a lot of people, a lot of founders, and you ping them?" Yes. How do you choose who you ping and who you help? My town obviously being one of them. Well, my current paradise I told you is this market is moving so fast that based on what you read, that an open clock comes out. Suddenly, there's this thing that people are going to have agents. There was a moment if you heard, if you saw, there was agentech browsers. Remember, we don't hear about them much, but there was a moment when everybody was going to have an agentech browser. Your browser was going to be your computer and that's going to be sort of dual the agentech task. When I hear about these things, I'm trying to assess which one's going to work. If it works, how does it impact my product portfolio? What do I need to build in anticipation of this technology becoming mainstream? That window from the idea to execution is shortening in the air world, as you can see. The way these companies are coming out and getting formed in 12 months and 24 months and getting 100 men air are probably the fastest ever, which means if that's what my enterprise customers are using, I have to figure out if I'm going to secure that stuff. Well, my team doesn't fully understand all this stuff. I'm really listening to podcasts, listening to people watching people tweet, watching people in LinkedIn saying, "This is an interesting technology helping the founder." My first step is to ping somebody who's doing something interesting, which I don't fully comprehend. They seem to be getting to a degree of success, which gives me a feeling that this could be something relevant. Perhaps not this company, but the construct that they're working on, the concept they're working on. I'm an investor in a world of investing in a world of uncertainty. You go later where there's more certainty. Yes. We spoke about this downstairs. You know, I have that luxury in terms of a flexible mandate to do that. You have that luxury too in terms of the benefits of scale and acquisition budget. Can you not just sit on the sidelines and wait for the right things to populate and then buy the middle billion? Yes, I know. Yes, we can wait. That doesn't mean I don't need to learn. If I'm not paying attention to eight different players in the space, which I'm sure you do too, if I'm not paying attention to eight of them, trying to see who succeeded why, what did they do wrong, what are those guys who got it right, do. It's very hard for me to assess what made it work. Was it a fundamental, there was a bad idea? The technology is bad. Agents are not good. Agents are not going to work. It could be that. Or this company didn't implement, right? The agents are still a phenomena. Somebody else is going to execute, right? I need to understand the underlying technology for sure for to make sure that my team's thinking about it and do we adopt it, adapt it and secure it in the future, right? We bought a agent AI company gateway six months ago. It didn't cost a lot of money. I figured out, saying, listen, if everybody's going to agentify the enterprise, how are we going to know how many agents you have running around the enterprise or keep track of them? How are we going to govern them? How are we going to secure against them? I said, the only way to do that is logical is to find a way to aggregate agent traffic somewhere. If it goes through a certain gateway of firewall or some router, I can watch all the traffic and I can stop an agent from acting. That's the only way it works. So I said, the first thing you need to be able to do agent security staff some sort of a gateway. So we bought a gateway product. Now I got it at the right price. If I wait, look at what's happening now. Suddenly, people are waking up to the idea we need some sort of a router or gateway that all traffic needs to go through because of optimization reasons, because of routing reasons, because of token maxing reasons, which you have paid double. Maybe it wasn't a big price, but I could have paid double. That's not the point. The point is things I buy, either they're going to help me 10 x or 100 x, I was going to fail spectacularly. It doesn't matter if I paid one or two x at that point in time. Of course, I shouldn't be paying two x and having it fail spectacular all the time, but the one two x doesn't make a difference. The one is to 10, one is to 100 is what you do. That's what we'd like to do as well. Not from an economic return perspective, from a business value perspective in our business. Are you more involved in Corp. app today than you've ever been? I've been always more involved in corporate. This is not about normal. I think I say I'm more involved in trying to learn what's happening out there from a technology perspective than I've ever been because the stuff is moving so fast and if I don't have a point of view and if I don't encourage my teams to pay attention to it and we talk about it, I think there's a risk we miss a trick. And if you miss a trick, you remember in life in technology you miss one trick, you can survive, you miss two tricks, you're partly impaled, you miss three tricks, you could be obsolete. A lot of SaaS providers are feeling obsolete today, their share price is telling them they're obsolete. Do you think that the majority of SaaS vendors have been oversold? Would you think that is an accurate reflection of where markets are moving? I think what the market is telling us is that the system of work or the system of record will see a reimagination of workflows as Unitarkin. So going from software that doesn't have an opinion to software that has an opinion and expresses an opinion and also does a lot of work for the human so the human doesn't have to repetitive tasks. I don't think those AI applications have been created. I think the SaaS versions exist, we all use them. I think at some point in time we'll see AI applications that do a lot of the task and workflow gets reimagined. I do think a lot of SaaS has built a lot of analytical capabilities to sit on top of the systems of work and system of record. I think it's a lot easier to abstract that data into some large data lake and have analytics analyze that data for you and give you the answers. I think the analytic world is getting reshaped already where we can see people like snowflake or glean or data bricks. All these people boast enterprise data lakes where you can bring the data and run all of them against it and get you much more synthesized analytics and outcomes than you ever had before. I think that's a third question which we started off. I think people are not sure how many people are going to work in these enterprise in the future. So if you take the confusion on how many seats are going to survive in the SaaS world, take the confusion around analytics are going to get done differently and system of work gets reimagined. So once you mean analytics done differently and so I'm again I'm disclaimer, I'm a podcast of all you get. Yeah, you're a podcast investor managing lots of people's money. That's true, but disclaimer podcast. I understand the seats question. I understand the work of like, can you tell me I'm standing on the analytics race? Well, you look at most SaaS software, right? In the past many years once you're fully deployed to the company, they say, listen, I got all the school data about all your employees in my HR system and I can help you get more insight into your system. If you take a sales force, they have a sales force marketplace with 300 apps you can use which are analytical apps that feed off your own system record, go to market data and it helps you analyze the data. You know, Neil Meta. Yes, of course. Yeah, I love Neil. I do too. I think it's one of the three phenomenal people. He always says the one question is, I other companies best days at a head or behind it. And that's a very helpful one. That's very question. Good question. Yes. Our sales force is best days at a head or behind it. I don't know. That depends on how they execute from here on. If I were to paint a backcase fee, what would that be? In the very cases, we don't get this transition right of the world going to a AI first future. Look, these false positives will keep reducing over time. Agents will become a real thing. Agents will do a lot of work for humans which humans have been doing manually in the past. All that needs to get embodied in your product. If I can't make that transition happen with my team in the next three years, yes, there's a bear case because somebody else will build a better master up. And the bull cases, you understand it better than any other security provider and you become the default. Bull cases that we get that transition right, there's already a trend in our favor underlying that where people are realizing they can't have 40 to 60 cybersecurity companies that they have to manage themselves. So we've been driving this trend of platformization already for the last 24 months or 36 months. We already see the fruits of that where people are saying, I don't want 40 people solving my problem. Let me put it all to it solves the problem that 20 different companies do together on our platform. The good news is because we are all coming to our senses and saying, yes, we need a lot more enterprise context and enterprise data. It has to be stitched, it has to be seamless. That's what we deliver with a current proposition. I just need to bolt on the right, ear nut bolt on her and brace the right ear capabilities and that stuff. Does the platformization remove the ability for venture scale returns? Remember, I need like 10 billion dollar companies. This is the big thing that I think most founders still don't kind of fully comprehend. I sound awful. A billion dollars doesn't do it anymore. It needs to be 10. It needs to be 20, 30. With the platformization, I can get the billion dollar out of it to you. Hopefully, please buy any of my companies for a bit in the cash. I'll give you the account. You can take them. But you know what, in Paytan? I'm not going to fight against innovation. I think there will be venture scale returns in cybersecurity because remember, where's the most innovative industry in the world? The bad guys are always looking for a new way in. They're not saying, oh, I exploited that two years ago. Let's try it again. Maybe somebody hasn't deployed a patch against that. We fixed that one. You got to go find a new way to attack people. It's all too tired to come up with an innovative way to hack into it. That's just trial. It's all too hard. Yeah. Exactly. Exactly. So it's highly innovative. There are new attack vectors. People are going out there trying to chase them. I'm not going to build everything myself. People will build great stuff. And sometimes people will great stuff and build a platform around it. And that's fine. Remember, we come at a different vantage point. When I started up all out, oh, we were less than 2% market share in the entire revenue cybersecurity. We're closing it on 8 or 9% right now. There's still a lot of room between 89% to 20 or 30 or 40. That means there's still 60% of market cap out there to go enjoy in different companies. And that's not all going to be existing players, including us. There is room to build companies which have tens of billions of dollars in market cap the next 10 to 25 years. Fucking enormous. Beautiful. You're well think about it. The S&P, what percent of the S&P is tech now compared that from 20 years ago? Year to day again. So 86%. Yeah. Well, it's a total S&P market cap. What percent is tech? And what is that 20 years ago? What will that 20 years from now? It was less. It'll be more. You know, I am great. So the entire tech space, like, you know, what do you call marketing tech in the future? Or marketing spend or tech spend? What do you call HR tech in the future? HR spend or what do you call the spend on all the tokens which are placed rapidly to human tasks? All becomes tech spend. You said bad guys is China in many people's eyes. And we see a huge amount of incredible open source models which are being used extensively to stay at a much cheaper cost. Do you think the proliferation of Chinese open source models is something to be concerned by or is an inevitable feature of a burgeoning ecosystem? So for a second, let's play the thought experiment. Take the word China out for a second. Answer the question. Do I think open source models? No. I think open source models are a dangerous. So does it matter where they come from? Yes. Okay. So you're not worried about open source models. You're worried about Chinese open source models. I'm not saying I remember there's a large tech company which also had open source models for a while. Sure. So it's interesting to watch open source models. The question becomes in the future. Do we end up with in a horses for courses? Do we end up with models that are very tossed specific and very helpful in certain tasks? And do they always need to use this mega frontier model for everything? And you already see that with the level of labs and you know, the voice models that are out there which are specific to a task and they probably do that task better than what the frontier model does. So over time, if you believe the world bifurcates into many task-specific models which are going to be useful for that task, that task-specific model could be better training across the depth in a vertical space. That's going to happen physically. I, for example, I don't think physically it will be as easy as having a generic frontier model because there's no consumer use case for physical AI. Right? It's a depth use case only. The question is, is your physical AI model that helps you fly planes going to be the same physical AI model that helps you drive cars most likely not? It will be the same physical AI model that does robotic manufacturing, probably not. So you're going to see depth in these models. You're kind of trying to see a world of bifurcated models. It's some sort of orchestration layer, as we've talked about in your busy finding orchestration layer companies that can allow you to pick the best model for the right task. Those orchestration layers have to get smarter and smarter. That to understand the context, that to understand the memory that has to the question is do I store my memory and context in the orchestration layer or do I store that in the frontier model? Which one do you think it will be? I know I'm the investor I should know, but I don't know. I think that the challenge is right now the frontier models know this problem and they're aggressively moving to incorporate memory and context into their models because they understand that's the mode. And the challenge is you have to pay for twice. If you say no, I don't want to use your memory and context, the model may not be usable if you use an orchestration layer. The orchestration layer today is not as well funded as these models. Prisk is you end up in architecture where the model has a lot of context and you cannot be model agnostic. You actually be model captive to get maximum efficacy and value for what you get done. It's like you have a choice. You have to go all in on the model or you can't go all in model. You can't do with one what you can do with the other. And if you want to do it with the other, you have to redesign your entire application that is deeply embedded with the capabilities of the second one. So in the world of bifurcation and horses or courses, I think opens this is a good thing because it allows you to play the cost curve. You don't need the smartest model, the smartest thing. It opens up so good where it comes from a certain country. The question becomes what back doors are you worried about that these open sources model have models have? What are you worried about? That's true for any nation state. If there's a nation state sponsored open source model, what are the back doors? Can I get in? Does the model wake up one morning and it's got a sleeper agent and it starts sending all the data somewhere else? Those are questions. Those can be secured. That's why you come to Paul also to help you secure the models. Always go to secure your back doors. I don't know. What kind of night do you have? It's good he's not a hangry night. It's the different kind of night that I'm dealing with. This Monday morning. It's terrible. What is the best time for me to show up here? I have your time zone. I am well rested. I'm not gentle. The only state means that you should secure your back doors. The only state that you seem to have gone on to. What things are your mind is going in the wrong direction. No, I've got two questions and then we'll do a quick fight. One is with the incredible success you've had. You've made a lot of money. It's just a question I have. It's like, what does no one know about having money that they should know. Like one thing for me is I've become much more impatient. We have very different, your phone more, so it's just me, I've become way more impatient. No one told me I'll become impatient. I'm used to a good quality of everything. And now when it's not that, I'm very pissed. I don't like that in myself, but no one told me it would have-- - What are you gonna do to fix it? - Therapy. (laughing) This is a common West in solution. - Yes. That's somebody else telling you. He gone back and said, "I must feel better now "because I told somebody that I was doing to take you." - I told me I should put myself more. - I did. - I did came back to being impatient because that's how you put yourself first, more important. Believe in yourself, right? Is that what you supposed to do? - Yeah, yeah. - Off to my spiritual self. - Yeah, yeah. - Put myself, be confident in yourself. - Be confident in yourself with your dad's faults. - Oh my god. - Is that what your therapist told me? - Yeah, yeah, yeah. - It's must be British. - Yeah. - But no one told me that. I kind of wish they had done. What does no one tell you about having money that they should do? - I think it's not about money as much as it is about success. I remember we all follow Maslow's hierarchy. I came to the United States with two suitcases, $200. And I was willing to do anything at all within reason, as the rights out of law, to make sure that I made a life for myself because there was no way to go back. I was gonna use the different word, but I'm not gonna use it because you go crazy again. So, there's no going back, right? It was a one way ticket, which I did not have. I had no recourse. So I was willing to do whatever it took. I took notes, became a security guard. I tried to pump gas for a weekend. - He became a security guard. - When I came to the United States, I was a security guard. I took notes with the disabled. I flipped burgers at Burger King. I had $200. I had to find a way of paying my tuition. - Was one quite transformative to your mindset? Did you hate them? Did you love them? - It's not like-- - It had to be done. It's karma. You come from Eastern philosophy, it's karma, right? It's destiny. This is what you need to do to break your destiny. So you do that. So you don't worry about what you have to do. Now, at that point, there was no-- - You always know you were gonna be successful. - I don't know. Who knows? Nobody knows. It's gonna be successful. You just come in and throw your best and hope for the best. See what happens. So that's very Eastern philosophy, right? If you believe in karma, destiny. How do you manage billions of people in the world? You make sure they believe in destiny. They believe in destiny and say, "Oh, this must be what was my destiny in the end. I tried my best. This is what I ended up." That's better than theotherapist. It's just like keeps you center. Say, "Okay, I tried my best. I gave it everything I had. But perhaps this is what God intended for me." Do you find that hard to believe? - If you were my therapist, I don't think I could afford you. - That's the problem. - Let's say this is free. You're getting too-- - I've been getting crazy. - It's why it's difficult. - Sure, I fully embraced all of that. But that was where I started. So if you start from that perspective, over time, you climb up Maslow's hierarchy. It was about food and shelter. And that became about ambition. That becomes of actualization conceptually. And Maslow's hierarchy. And you get to a certain amount of money that you decide there are some things I don't have to do anymore. I don't have to be a security guard. I don't have to flip burgers. But that very quickly goes up further than I think. I don't have to tolerate certain things that I told her in my life. Because I don't need it in my life. Because I don't have to adapt to the circumstance because I can walk away. You have to get worried that the willingness to walk away makes you softer. - The willingness to walk away makes you softer. - No, actually it's the other way around. The willingness to walk away makes sure you optimize the outcome. When you negotiate, if you're fully vested in the outcome, you fold at some point in time saying, "Well, I can't let Larry Steppings walk away 'cause Harry walks away, I have no deal." But if I say, "No, Harry, it's gonna be these terms or no terms." I'm willing to walk away. Then it depends about the widths. So he says, "Harry, you want it more, do I want it more?" - So to make yourself-- - I don't want to make it political, but it's not Donald Trump's, like the art of the deal. - I don't know, I've not read the book, so it's quite a good book. - Is it good? Like at the end of the day, I don't think being willing to walk away makes you softer. I say, being willing to walk away makes sure that you understand the pros and cons of what you're dealing with. It makes you understand whether you should spend your time over there or not, which makes you understand whether you can get an outcome that is useful for you as well as the other person. - No, we have a lot of choices in life. Once you have a lot of wealth, you have. - By no one, and then we'll do a quick fight. I care a lot about kids, actually. I love kids, and I want to be a really good father when I am one. - You see, you have a public company. It is incredible. You've had an insane career, and I've had the pleasure of meeting in one of your children. She's amazing. - Cheers. - What advice do you have for me on lessons on how to be a great dad, but also not losing in short work? I'm not willing to sacrifice much on the work side. - You know, this is the hardest problem in the world. I think there's about 20, 30 billion people who have been born since the beginning of civilization. Yet there is no AI that can train us on what we need to specifically do to create the outcome we'd like to create. So we do many variables, right? All kinds of people in the world, and I'm sure their parents, some of the parents are amazing, some of the parents are not as amazing. So I think part of it is you can do your best from your perspective, and I think kids absorb a lot by watching you, your work ethic, they watch your values, they see how you interact with them, because my daughter probably has a better sense of why I'm the person than anything I can tell her, because she spends time around me. She sees me interacting every sort of micro situation, what makes me impatient, what makes me patient, what makes me do certain things. And at the end of the day, your child believes that he had the best intention for them, I think that goes along. - I had a guest on the show, and they said, "Watch National Geographic if you wanna be a good parent." And I said, "What?" And he said, "Look at the elephants, the children follow." And so if you want your child to be nice to waiters, be nice to waiters. If you want them to work hard, work hard. - Yes. - But it's true in organizations too, by the way. Organizations take on the form of the leader. I'm pretty sure if you close your eyes, and you've rattled off, you know, five or six attributes of a company and said, "It's a company as a founder," and say, "How do you compare the company's cultural values vis-a-vis the founder?" And you'd find a remarkable resonance between the two things. Companies act, because remember, the organization is trying to please the founder, 'cause they figured out that's the way to achieve success. If my CEO is impatient, and my CEO is exacting, and my CEO is ambitious, my CEO suffers no fools gets stuff done, then that must be what they want at a war. So you suddenly find, if you, this is getting depressed, you know, the right values are not. And you told me a story about a guy who had different values and they had to shut the company down, but you had the right values. People will watch your behavior, and want to emulate your behavior. - I wanna do a quick fight, 'cause although I thought, "Take up all of your time," what is a belief that is held by most top investors and founders in Silicon Valley today, do you think is wrong? - My concern would be at this point in time, given the base at which technology's evolving, given the uncertainty in terms of what's gonna work, what's not gonna work, I'm worried that it might be too much euphoria and a bit of formal going around in terms of, "Oh my God, if I don't invest in something that's interesting in the right founder, I'll be left out." Because people have seen this happen, look at what's happening in the topic, right? You've missed the first round, the second round, the third round, the fourth round, the fifth round, and you look like a guy who's not moneyed. Now, you had 20 years to invest in SpaceX, you had three to invest in an topic. That pace is fundamentally different. I'm sure as many people are happy that SpaceX finally went public, as many people are probably saying, "Mope, I'm saying, damn, I should have done the "entropic round two years ago when they showed up on my doorstep." So I think there's a lot of formal, coupled with euphoria on the other side, and I think the risk is that we think every company that's gonna show up now is gonna be the next end-tropic, so we better get into it. My next one, she was, "any moment, any board meeting." What was the biggest, oh shit, in a board meeting? - I got a very interesting insight for one of my board members. You know, we're prolific buyers of companies because I'm constantly paranoid that we have built it. Somebody else is gonna build it, so we better go acquire it and find the team to go get it done. And there's one particular acquisition that took a lot of effort to get the founders to the table, get them to agree, grind through diligence, figure out whether it's gonna work or not. - It's a substantive amount of money, relatively speaking, a hundred million dollars, close to almost a billion dollars, and I call one of my board members and I said, "Hey, what do you think about this?" She said, "You're calling me. "You don't call me all the time, "but all the acquisitions do. "So this must, this one must be different." And I said, "No, it's not different. "I'm just thinking hard about it. "It's taken a lot of effort." She says, "Go for a long walk, ignore all the effort you put in." She said, "Because sometimes what happens is "you confuse effort with one team to get the outcome. "Because I spend a lot of time at effort trying to get it, "then you feel like when you get it, "you better take it because you put all the effort in." And she says, "You haven't spent a dollar yet. "You just put in three months of effort, "but remember once you put the dollar, "then it becomes zero to make it successful." So you still have one more chance to decide if you want it or not. I go for a very long walk and say, "If this walked in the door right now, "and there was zero effort involved, "all I had to do was write to check what I'd take it or not." - Forget the sunk cost. - Yes. - Have the same in the investing business. You spend so long. - Yes. - Yes. - You spend a lot of time, you're just like, "Oh my God, I'm the one getting the term sheet, "and nobody else has it. "I nailed it. "I've beaten out eight VCs to the questions. "That's not how many VCs you put. "Beep to get the deal." The question is, if this deal, can this deal stand its own merits? And would you invest in it if there was no competition? - What's the best advice you've ever been given? - The best advice that the really old man gave me on a flight once was, you know, life is simple. If you wake up in the morning, you're really excited about going to do what you do for a living, your blessed. And if you're done after a long day and you're really excited to go home to your family, you're blessed. Did I like that? - I do. I actually tweeted last night, I hated school when I was a kid. Sunday night was the worst. And like, my Sunday night last night was like thinking about our show and the comedy. What a great Sunday night. - What a great Monday night. - What was it you were going to go with that? - No, no, no, no. I'm lucky, am I? Seriously. It's amazing. Final one, what are you most excited for when you look forward next five to 10 years? What are you most excited for? Is it becoming a grandparent? Maybe. I might have just seen my mother become a grand mother. It's amazing. She's amazing. Is it the health benefits? See, I'm so excited that AI might be able to solve multiple cirrhosis, which my mother has. That being incredible. - You never know what tomorrow's going to bring you. The only way I've been able to do everything I do is not to get too hung up on what's going to happen to me in year from now or five years from now. Because it's too far. I think you wake up in the morning, You have an amazing day. day and everything's working around you. If you're kids are happy, your family's happy, you enjoy what you do. You have good friends. And I was at a different space, placed the other day, and this weekend they asked me, like, you're not a 996 CEO, what do you do? And I said, like, I try to make sure that I can find something to enjoy every day. Because I have enough things to worry about. I could really get myself in the wrong headspace by worrying about a lot of things. I run cyber security for crying out loud. I live off the fact that somebody's going to hack somebody at some point in time. My phone rings and can you help us? It's like, why didn't you spend the money before? But can I help you? It's like, can I help you? I think it's a state of mind thing. Can you get your state of mind to be optimistic, positive, and one of gratitude and happiness every day? If you can, it's going to be great. Well, guess what? If it helps, then if it's come, I can start being Benjamin Button, be amazing. If my kids are continuing to be happy and successful, be amazing. My mother lives for 150 years and she's happy to be amazed. There's so many amazing things that's going to happen. And perhaps may not happen. So let's just focus on tomorrow. And, Cass, I so appreciate you being willing to come back for a second. I mean, after the first, when I suggested I was like, there's no chance you'd be. Good thing I've got back a little bit. But thank you so much. You've been incredible. Thanks for having me. But before we leave you today, you have the idea, but often with AI tools, you hit a wall. All stack, web and mobile apps, sites or autonomous super agents all built in minutes, not we can spend on damn configuration. Base 44 ships it all out of the box, the back end, the database, the authentication and the hosting. It handles the heavy lifting. So you can just stay in the flow. It multiplies you. But to win, you got to be first and base 44 is that edge is the move that lets you skip the troubleshooting and get straight to the breakthrough. Launch your next big thing at base 44 dot com. That's base 44 dot com. My world, rodent arresting first line gets your ass covered with Korgi insurance and I'll tell you why if you're running a business right now, you already know this pain all too well. And it's really slow. It's confusing and my word, it's full of paperwork. Well that's exactly why Korgi is here to change the game. Korgi is the first and only insurance carrier designed specifically for tech companies allowing you to get covered in minutes instead of days. Korgi provides essential coverages for all growth stages such as D&O, Eando liability, cyber, commercial, general liability and more. I love the way we say ass with Korgi insurance alongside thousands of other startups at Korgi.com/20VC today. That's Korgi.com/20VC. You won't regret it. Models perform well on benchmarks but they fall short once they enter real coding tasks, real tools and real workflows. That disconnect between synthetic evaluation and actual system behavior is now a core block of foragantic models. This way Nvidia and Thropic Salesforce Gemini and other leading lab partners partner with Churring. They build realistic RL environments. Their generation data quality systems built from real world operational traces and coding data sets that stress models under conditions where failures matter. State changes, workflow branching, brittle tool cools and the coding errors that break RL agents but never appear in benchmark reports. In reality, a model may demonstrate correct reasoning in your evaluation setup yet still select the wrong parameter or mishandle a code update in a realistic interface. To find out how, visit churring.com/20VC, that's TURR-I-N-G.com/20VC.