Vinod Khosla: trust — not data lock-in — will be the moat for personal agents, and startups have an edge over Meta

Speaker 1: develop over his career. He's the perfect person to talk to about this. And so we're very pleased to be joined by Vinod Khosla, founder What's going on? Of Ventures. Vinod, how are you doing?

Speaker 4: I'm doing great.

Speaker 1: Welcome to the show. Thank you so much for joining us. I would love to hear about how you're thinking about this moment in personal agents. It feels like a new category, a new feature, a new product, a new winner take all potentially, how are you viewing the market? How are you investing against it?

Speaker 4: Well, it's a new category. It will be a broad category. I do think it will be very diverse and I don't think there will be a winner take all. Okay. Every one of the big labs or big companies, Meta included and Google included, will want to have their agents. They will have different trade offs, and I think we'll see a lot of diversity in requirements. As to Alex's comment, I think he hasn't used agents enough to understand much of that diversity that he's thinking about. Like, The we want to best way to browse is have the agent do the work for you and give you the options and ask you questions and be interactive with it. It's amazingly good when you can do it that way and I think some companies will optimize for it. I think the fundamental difference will be, do you want to be an advertising product or be something else? Do you want to trust Meta with your data Or you want to and some people will, and they do today, and I think Meta will be successful. I'm not trying to say they won't be. But there's others who will have different trade offs, and I do think agents from startups will move faster than agents from the big companies. Interesting. Okay. Yeah, that

Speaker 2: makes a lot of sense. I wanted to ask you about the knife fight that's unfolding between agents, which are right now deeply aligned with users. Even fundamentally, last year we went through the cycle of labs and different players launching new browsers Oh, yeah. Which I thought was fascinating at the time because it felt like for me, ChatGPT was a browser. Yeah. It would replace my need to go to 20 different websites to understand like a topic or a certain product sort or anything 20 like Chrome that. Tabs. It's already was a question. And I think that consumers like the ability to do a lot of things from one single interface without bump, you know, jumping around and they can ask a question about a product and then navigate to another and all these different things. But it seems like there's this, know, something that's been basically ongoing, this fight between new agents that wanna just do what the user what the user wants in a way that's aligned to them and these platforms like Amazon that have big businesses that are under attack by these sort of new products?

Speaker 4: I think you're right on there. There's a massive difference between a company wanting its agenda. And the browser battle you referred to wasn't for the consumer. It wasn't the consumer wanted to move from Chrome to something else. It was the company saying, Google controls what happens on the Chrome browser. I don't want that, so I'll have my own browser. That's a very different need than what the consumer wants. I think from the consumer's point of view, with all the assault on the consumer, the marketing assault, I should say, from AI, and all the AIs ganging up on changing shades of blue, which Meta does to make you click more or other things TikTok plays. Do you want to be subject to the consumer or fight all the AIs with your own AI, your personal AI that only has your interests at heart. And I think that's a very, very large and important category. I think I talked about this first about four years ago, the largest opportunity being protecting the consumer against all the AI assaults and getting all their work done without it being biased. Today, you can't even look at the reviews without understanding whether it was paid for by whoever's marketing it. Yeah. So, representing the consumer and having an AI that loyal to the consumer, not to a particular company, a particular product, a particular point of view, is going to be very, very important.

Speaker 2: Dude, so so Silicon Valley has long said consumers consumers won't pay for things at scale. There there's

Speaker 1: consumers lot of people. Wanna be entertained. They don't wanna be productive. Yeah. If you're building a tool that is productivity tool, you ultimately need to sell that business as an enterprise. Yeah. And yet it feels like that is gonna be the most aligned model for

Speaker 2: agents because your agents can go fight different things. It can save you money, all these different things. It feels possible that this time could be different. Don't wanna maybe it's too early to say, but an agent that can get $20

Speaker 1: a month and it saves a couple hours a month and it That was Mark Zuckerberg's pitch. He said Muse will save you money. And we're already seeing examples of people using it to cancel subscriptions that they're not using, get refunds. And so there might be some economic calculus where it looks more like paying for TurboTax than, you know, browsing Instagram Reels potentially. So how do you think about monetization?

Speaker 4: ChatGPT Yeah. With its subscription service proved that consumers will pay if there's enough value. Yep. Yeah. Massive business got built. Google always assumed consumers won't pay, and that was the conventional wisdom, and Chad GPT, in my view, destroyed that. Sure. For enough value, consumers will pay. Now, saving you money is really, really valuable for most consumers, and they will pay if something does that. Knowing that it represents them, knowing every time they're about to spend some money, they can find a more efficient way to get the same thing, that's a big deal for consumers because whether you're looking in travel, whether you're looking in shopping, there's always ways to save money. Subscriptions is a great example. I have way more subscriptions than I know about. And so representing the consumer's interest and acting on their behalf as if it was omnipotent personal assistant is what matters, and personal is the key. It's for them, the consumer, not for some company.

Speaker 1: How do you think about moats in this era of personal agents? There's a there's a thesis that's percolating right now that says once you install a personal agent, once you've given one personal maybe people won't trust Meta. Certain people obviously will. But the people that do and they go in that bucket and they start daily driving that particular personal agent, they give it access to their calendar, their email, their text messages. They are not going to move to another agent even if it's a little bit better. At the same time, I feel like there's a risk of if you're on instinct, you can go to Muse and say, Hey, just off board me from instinct and set up Muse or vice versa. And so, it might be a little bit less of a hassle to off board, not quite as strong of a mode as traditional social networking. But how do you see the moats question playing out in the I era of personal

Speaker 4: think off boarding will be possible, a little bit clumsy, but then if you have a good agent, you tell it to do the off boarding. Yeah. And the onboarding onto the new one. Yeah. And we saw that with, with OpenAI and and and Anthropic

Speaker 1: saying, hey. Go go dump this prompt in the other chat app to get all your memories. Come and paste it over here. And they were fighting over that because, you know, it was possible to move from one LLM to the others. What I would say is the complexity

Speaker 4: of tasks, especially as companies make it hard to offboard, will no longer be a barrier. If I had to look at all my information on Google or Meta and figure it out, it's complex. If I want to know, I don't want Google to know where I was on September 3 at 5PM, they probably know, and I can probably disable that, but it's hard. AI is just going to be good at doing all these complex things, and so I think that moat disappears. So what's the moat? I believe the moat will be trust. A company that develops trust with consumers, that's where Meta has a huge disadvantage. The trust will be the moat. That'll be important. I also think completion is going to be really important. Consumers will be loyal to products that get their shit done. Excuse the language. Yeah.

Speaker 1: It's true. It's true. It's a good point. It's that simple. I mean, it's like we miss it a lot, but that is If the look at the Vagio agent

Speaker 4: Yeah. You don't need to know what it's doing, but it knows when it's stuck. Yeah. Most agents then just tell you they're stuck or don't even tell you that they're just stuck. The audio agent, for example, will pull a human in as a tool. Oh, interesting. You lose a human to complete your task. Interesting. So the idea of trust first, your information, your most private information, what you like, what you don't, what you buy, what your preferences are, that's trust. Then there's completion. If you can get all my stuff done, I'll be loyal to you.

Speaker 1: For the very large companies, it's really hard to have humans in the loop. Mhmm. Oh, sure. Sure. And I think Yeah. You need to scale it. I mean, it's it's it's, yeah, very difficult to come over the top with, you know, a product that can service a billion users that has humans in the loop because you have to stand up. Yeah. That won't force advantage

Speaker 4: to being small and agile and Interesting. And okay. Right? And and frankly, for people who need stuff done Mhmm. It's much, much better if you pay a little for the human in the loop and get things done. Yeah. The other thing that's really important, and you talked about the Chrome browser, not everything will be in the browser. The right agent won't be an app. It just where do most people live? They live in their email, they live in text messaging, They live in Slack. They live in other channels. And so an agent has to operate primarily with all those channels. In the app, if there's one, and there may not be one, it's just in the background. So you don't have to open the app, you just text message your agent like you do with Foe from Voyager, and it takes care of it. Or sometimes you just pick up the phone and do that. And sometimes the agent just has to pick up the phone and have a conversation with the merchant. Yeah. So, know, meeting the consumer where they are, the consumer experience is really, really important. Not get stuck in, we do text only, we do browser only, you have to be in our app. Users don't like those constraints. How do you think about it? I don't wanna install one more app on my phone.

Speaker 1: Yeah. It's reflected in the App Store. It's it's it's all just AI apps. How do you how do you think about the way you will interface with personal agents over the next few years. If if you if you want to send a message to me, John, will I receive that message from you? Will I think it's coming from you, or will I think it's coming from your agent? Will your will your agent identify itself as Vinod's agent with a different name, with a different identity, a different persona perhaps? Or will you want to tuck the AI beneath the fold of your digital life so that it's sending emails as you?

Speaker 4: So I was actually doing some ordering from Petco yesterday. Oh, yeah. And it said some things to me, and I said, just pretend to me me. And it said, I can't do that. I won't do that. Mhmm. So this goes back to trust as a key mold. Interesting. Right? So I do think for both parties, you have to be transparent. The agent is your agent, but it's not you. Mhmm. And I think that's important, and that'll engender trust. Mhmm. And we'll see much better

Speaker 1: I want to stay on Petco. Dog, cat, bird, what type of animal are you most fond of?

Speaker 4: Oh, dogs for sure. You go. I mean, dog. I mean, what kind of dogs? I'm

Speaker 1: sorry? What kind of dogs?

Speaker 4: Oh, I'm a big fan of Newfoundlands.

Speaker 1: Have thing that does. I have a Newfoundland. I love Newfoundland. Awesome. Then I had two. One made it to eleven. He passed away. Rest in peace. But we have another one that's going strong. A

Speaker 9: a 130 pound, 2,400.

Speaker 4: Lovely doggy. I I had I had two Newfoundlands that died last year. They were both around two hundred pounds. Wow. Wow. And so we now have a puppy, but I was trying to get a cast done. Yeah. I could have looked at six websites for this specialty type of dog collar after My puppy had surgery last week. It could call lots of stores. Yeah. And look for, do they have same day delivery, which is the key thing I was looking for, and can it pay for me and all that stuff. One of the great things is payments is going to be a real differentiator.

Speaker 1: Sure.

Speaker 4: If you can be absolutely secure and give your agent your credit card, that's a big deal, and there will be lots of abuses or mistakes in that domain. I feel comfortable with the Wagyu agent giving it my Amex card. Sure. Because it has certain types of protections that are going to be patented, so I won't go into it. But I will be absolutely sure that card won't be reused for something I didn't authorize. Yeah. Yeah. You can sort of set up

Speaker 1: payment rules if it's over this threshold, actually in code. That's not something that can hallucinate. It made a mistake and the model thought it was a $100, it was actually a $100,000. It it makes a mistake like that. So, yeah, makes sense that we'll be in this blended regime of both, you know, deterministic software,

Speaker 2: stochastic software. Last thing for me, do you think we'll get to the point where users will wanna give, like, something like power of attorney to their to their agents? Dan in our chat was asking about that, and I and I've thought about this before. Just working with executive normal executive assistants, always very frustrating where you get to the point where a certain service will be like, hey, they actually need you to call now. I can't sort of like take this any further. So it feels like we'll get to that point eventually. All

Speaker 4: these questions come back to what's the mole. Trust is the mole. Will you give your power of attorney to the agent if you trust it? My bet is 95%, 5% of users will do anything. 95% will not trust giving your power of attorney to your agent today. Now, in four years will that start to happen? I suspect forty or fifty percent of people will give a power of attorney to their agent by 2030 because they've had enough interaction and developed enough trust. Mhmm. So trust becomes the key issue, and this is why startups may have an advantage. Startups with great transparency will have an advantage over people who already have a brand of lack of trust.

Speaker 1: Yeah. What's been your journey with Wajo and Pho? Did you invest early and then this is just the right product at the right moment? Did they pitch you on this day one? Has this been years in the making? What's been the journey?

Speaker 4: Well, we invested in it day one. In fact, we brainstormed with Shivani, who's just a great technologist and a real understanding of the human factors. Mhmm. Those were two things that are key. Mhmm. So we invested and we even brainstormed the idea. So we're really happy with Vajo, where she's, and she's of course deeply technical and has this other understanding of what consumers need, and I I call it empathy for consumers is really what you need Yeah. And say, how do I best represent them? That's the right mindset. So we've been very excited about Wodger.

Speaker 1: Do you need an access code? Is it in early access? How does how does actual onboarding work right now? Because it feels like if there's a human loop, you can't jump straight to general availability because you might get flooded. Right?

Speaker 4: You you also wanna grow slowly Yeah. Because you wanna of connectivity to lots of things, and that builds over time as you get more users. Today, it is not free access. Okay. It's limited access by invite only. Okay.

Speaker 1: And hopefully, you can give people an invite code from the company. Yeah. We'll share out some invites to anyone Drop them in the chat. In the by September 30, there'll be a free release. So by the end of the month, you'll see it pretty open to everybody. Awesome. Well, thank you so much for coming on, breaking it down. And Yeah. Great great to see you. Great to you. Tussling with all the different trade offs in the modern era. Have a great rest of your day, Vinod. We'll Great. Talk to you Cheers. Thank you. And send my best to the puppy. I hope he heals up quickly. Let me tell you about Cisco. Critical infrastructure for the AI era. Unlock seamless real time experiences and new value with Cisco. There's some breaking news before we bring in our next guest. The FAA, the Federal Aviation Authority, has restricted airspace up to 10,000 feet for the Tesla Roadster reveal. Remember, we have a bet going here. Elon has teased that the next Tesla Roadster, which will be announced October 2, I think, October 1, he has teased that the Tesla Roadster will leave the ground. He said it might fly. There's a SpaceX collab. There's four cold gas thrusters. We had a debate. Where did you you had six inches? You had six inches? No. I was at nine. You were at nine and I was at six feet? I was at half a foot. Half a foot? I thought you were at eight inches. No. I thought I was at, like, six feet. I wanna jump in. I want it flying. I'm going 10,000 They FAA shut down the airspace around the launch site. That's this. They shut down so the FAA has has issued a temporary flight restriction over SpaceX's MacGregor, Texas rocket development test facility ahead of Tesla's October 1 roadster reveal. The restriction covers a 1.5 nautical mile radius around the site and extends from the ground up 10,000 feet in the air. That altitude is unusually high for McGregor where restrictions tied to static engine fires and component testing have often topped out closer to 2,000 feet. So they're going five times higher. The larger block of airspace is feeding speculation that Tesla plans to demonstrate the long promised SpaceX package for the Roadster. Elon Musk has previously described a version of the car equipped with cold gas thrusters that could dramatically improve acceleration and potentially allow the vehicle to briefly lift off the ground. Reports earlier this year suggested Tesla was preparing a remotely operated demonstration at McGregor with the car unmanned and spectators positioned hundreds of yards away because of the expected noise, which is just a hilarious a hilarious thing. You're demoing something that if you drive it within hundreds of yards of someone's house, they will be annoyed or maybe not safe. Like, how is this going to work? It's a lot of it's a lot of showmanship for, what could wind up just being, like, a pretty fast sports car. But Tesla has not confirmed that the Roadster will actually leave the ground. So far, the company has only teased the event with language like go for launch and imagery showing the car on what appears to be a launch pad, but the FAA restrictions suggest Tesla and SpaceX are preparing for something beyond an ordinary static presentation. Whether that means the Roadster will actually hover on October 1 remains unknown, but the FAA has now reserved a substantial column of Texas airspace for whatever happens. I wanted like an f 35. I wanted to do full VTOL. If I wanna park, just bring me up, plot me down. I don't know. I'm excited. I think I think he's gotta deliver something. It's been nine years since he took the deposit. Are you gonna put it about deposit down?

Speaker 2: I think it's I'm not gonna put a deposit down.

Speaker 1: $5,000 due within, like, ten days. And it feels like it's kind of high optionality because if it's incredible that so many people are gonna flood in after the fact, But unlike waiting nine years, you can just wait until October 1. If it's a flop, you can just get a refund. So for locking up $50,000 for ten days Yeah. I think you should do it. I it's pretty tempting. It's pretty tempting. I need to really read the fine print or have AI read the fine print to understand if there's a chance I get hosed. But there's a chance that you get one of the first ones, not even, oh, it pops financially and it resells. Just the fact that it could be a really special and cool car that's really fun to have. And if they bring it in with the final price of $2.50, there's a lot that they can do to impress at that price range. If it comes in at 800, they're competing with a lot of No. I mean, if if it's just a

Speaker 2: if it's just a sports car that looks like a Cybertruck, it'll be a huge hit. If it's $2.50.

Speaker 1: If it's 800, now you're competing with new Vollari. And Nuvolari has a v 10, v eight, something like that. Like, it needs to be it needs to be Cybertruck, Cybericon, like Huracan Cybertruck, but also two fifty. Because then you're like, wow. This is this is crazy. This is competitive. And that's what Tesla has been for a long time. The Model s was was really, really performing, delivered a lot of value for a pretty decent price. Same thing with the Model y. It's still like a value play even though it is a relatively mid market expensive car. Well, yeah, I bet you

Speaker 2: Aaron

Speaker 1: has has his preorder down from like We're gonna find out. That's gonna be the first question. We're bringing in Aaron Levy from Box. He's the cofounder, CEO. Welcome back to the show. Did you put down deposit

Speaker 2: for a Roadster nine years ago? For a Roadster deposit?

Speaker 3: I I would I I didn't do nine years ago. I was very tempted this time around though. Right? I did I did go down the rabbit hole. There's these great YouTube, you know, kind of Speculative fiction? Yeah. Sleuth folks that Yeah. Did you see the thing if you No. If you white out the the the the video, it says where we are going.

Speaker 1: Okay.

Speaker 4: Do

Speaker 1: you see this? Is is that a reference to Back to the Future? Yes. So it might it might go 88 miles an hour. That would be huge. Either that or or So no

Speaker 3: So we're there's a whole I went I went and watched a twenty minute video. I'm very excited. I'm glad that you What's your what's your theory though? What's your theory? No. I don't know. The the I mean, obviously, Elon would make a flying car. So it's I think we're all we're all waiting for that. Now I guess the question is, like, practically, what is it what does that do if you literally can't fly it?

Speaker 1: But but but, like, if you're great, we can fly it if you're within 1.5

Speaker 2: nautical miles of Starbase, Texas and the FAA clears you, I guess. But I wonder I wonder if there's some loophole where, like, if you're on your own private property, can you can you get off the ground and and you could have, you know, those people have have The speed tall things. Yeah. No. No. No. I was gonna say, like, you know, you drive onto it and then it spins your car around. Oh, yeah. Yeah. Could make it yeah. Turntables. So you could make it, like, you know, if your your garage is kinda tight and you wanna turn around,

Speaker 3: a little indoor property. We may have reached the point though where Elon is building things for his personal life that exceed our

Speaker 1: normal zoning laws of of mere mortals. There's be a lot of people with deposits that are like, you're telling me I need a 100 acres of private land now? I was told that this would go zero to 60 in two seconds. And the craziest thing is the original the original specs were zero to 60 in two seconds. Cars already do that. The Model s Plaid did that. Not not crazy. 250 k. Okay. That's like a reasonable price point for a supercar these days. But six over 600 miles of range, which is which is actually not been done by any EV guys. And in a smaller vehicle, it's not like it's gonna be huge. So Yeah. Very, very uncertain how Yeah. Whatever they mean,

Speaker 3: seem to be doing things with batteries that I never thought would be possible with batteries. Like, I was always told, like, twenty years ago that, you know, we'll never get more than 30 miles in a in a car ride. So

Speaker 1: you just wake up one day, and I guess it's 600. So Yeah. I mean, I guess if you launch with the cold gas thrusters, put it in neutral,

Speaker 3: coast for a 100 miles of rain. Maybe that's exactly yeah. They'll get you the first 300 miles with

Speaker 1: Yeah. 600 downhill as long as you just cruise them. Anyway, how are things going at Box? Roller coaster ride. SaaSpocalypse fully canceled. We were very happy to see that. A bunch of we we were early to cancel. Yeah. See. We bottomed You guys you guys canceled it within like three hours. Yeah. No. We were like, we're in the business. It's not happening. No. But we did have we did have a a newsletter that went out like Yeah. During the depths. And I believe we are the first people Sasspocalypse is canceled.

Speaker 3: Yeah. Because Actually, you know, you almost though had to you almost couldn't cancel it too soon though because you you had to have the full trough. Yeah. And you had to have, like, full capitulation turns out that there there is actually software that is powering the world and these systems of of record are incredibly robust. And Yep. And actually, the more agents that you have running on these systems, in fact, the more important the backbone becomes. Yep. You know, I like, everybody always sort of made this claim that once you have all these agents, then actually all they really need is a database underneath. But the problem is is that database has to be coordinated against possibly thousands of end users and then possibly tens of thousands of agents. And so the the access controls, the business logic, being able to actually render the the experience, the security, obviously, you know, see OpenAI Hugging Face, like, the these core systems are actually gonna become more important in a world of of all of these agentic workloads. So it's good to see the market, I think, paying more attention to that reality, and it's obviously, you know, driven by, I think, the continued performance of these these software products doing well in the market. Yeah. I mean, even over the weekend, I was generating AI videos. I was trying to get them from my PC to my phone. And at no point was I like,

Speaker 1: build a new storage system for these videos. Even if it could do it, I'm like, I don't wanna waste a prompt. I don't wanna waste a second doing that. Let's use what I already use. Let's use all the tools that are already built out that are robust and I can share links and all these things. I don't need you to one shot some some shoddy clone. It's a waste of time. It's a waste of money. It's a waste of energy. Bunch of things. But in terms of in terms of resetting the narrative with the market, the street, investors, employees, everyone, how much of it is going back to, okay, you thought this thing was existential. You thought we were gonna be in so much trouble. We're not. It's business as usual versus there might be still some threats, but there's more opportunity, and that's actually outweighing the threats. And so that's why the company's performing well in the markets over the last couple months, for example.

Speaker 3: Yeah. Very much not business as usual, I think, for anybody in software. Okay. So so like like, this is if this is, like, sound I don't know if a single company is still being run the way that it was run maybe, you know, eighteen months ago. Mean, everybody is is, you know, running insanely fast Yeah. Into the future. And then I think we're we're you know, there there still will clearly be some parts of the market that don't capture as much of the value Mhmm. As they did previously. Mhmm. And then there are gonna be some parts of the market that capture actually a ton more value than they did previously because of of what agents can do. I mean, you look infrastructure providers as an example, like the cloud flares of the world, totally on fire. Because agents need sandboxes, they need compute, they need network, they need gateways, like great business. If you look at Databricks or Snowflake Yeah. Agents need to be able to triage against large sets of data that are that that that that you need to be able to pull together in one kinda common system. So doing fantastically well. For us, we have reaccelerated growth far past our internal plans because it turns out that enterprises need core systems to be able to manage their unstructured data. And agents need to be able to work with that unstructured data to make decisions or move information through a workflow, whether it's all of your contracts or your research materials or marketing assets or financial documents. And again, like, it might be easy in a personal productivity sense to be like, okay, well, I can just have that data in some cloud system or even in my local device. Now imagine an enterprise with 10,000 employees. Yeah. How do you ensure that those agents continue to go after the actual canonical records? Yeah. The actual documents that that are the real sources of truth and that are the authoritative versions of that data, that means you need platforms. And we're obviously competing for how do you become the platform for all of this unstructured data. But I think there's a lot of value being created really across the these ecosystems. And and I would say, right now, my my view would be it is unequivocally been a net positive for software, even with some of the spaces being under pressure, simply because how much more work agents are doing on these systems and the need to be able to provide the guardrails and be able to be the traffic cop for what those agents are doing, and to be able to deploy these agents natively within your own agent harness Yeah. In your own system of record. And and all of that is creating more value for these platforms. At at a system level, what is the

Speaker 1: what is the biggest, like, lever that you wanna pull? Is it, like, speed of interaction with agents? Because in the previous paradigm, search and and, you know, file retrieval, if it took a couple minutes, somebody's drinking coffee, they're checking their email. Maybe it's not as, like, crucial. But if it's like if my agent is hanging because I'm, you know, churning through a bunch of of records or files, that could actually be more annoying, like like or or or is it just the volume of content? Because even an employee, if they're not, like, totally slopping it up, they're probably producing more files, more contents, more video. Everything's just multiplying. So, like, where have you been, okay, this is the lever we need to really pull to make sure that agents, like, love the Box experience?

Speaker 3: Yeah. I I mean, I think realistically all of the above. First of there's an explosion in just content creation in general. Agents are producing content obviously at an unbelievable rate. You know, if you Stock look at grenades.

Speaker 1: Yeah.

Speaker 3: You gone through this? I wanna talk I wanna talk about that, but let's finish what you're saying. Sorry. There's just data more than ever, and and you're gonna need to you're gonna need to govern it. You're gonna need to store it. You need to manage it in just as if a person created it. And so all of that is obviously, like, the core backbone system. We're doing a lot of work at the second layer, which is which is we we we effectively think about it as sort of context retrieval for your unstructured data. But obviously, if you did a regular search, you know, two or three years ago, and you compared it to a search today when we actually can capture embeddings of those documents, we can our agents do re rankings of the search results. We search in parallel so we can run multiple versions of your query at once. Your chance of finding your document, not only faster, but just generally more accurately, is way higher today than it was a couple years ago because we not only have all these other signals, but we can parallelize the process Yeah. Much better. And so an agent from the outside coming coming in and saying, hey, I'm looking for this record, you're able to now much more accurately achieve that result, which is fantastic. And then I think the thing that we're also, we're spending a lot more time on is how do you better secure data in this world of, you know, kind of infinitely more agents than what we had, you know, three or five years ago? You know, the OpenAI Hugging Face incident has many interesting ramifications to how we build agents, how we ensure the safety and alignment of them, but has very real real world implications on the security of our of our system. So while this was more of like a production system from from Hugging Face that was under threat, you can imagine the internal environment of an enterprise. If you have, all of your critical contracts or financial documents in an environment that agents can eventually get their way into, enterprises are gonna need to care deeply about how do you protect that information. So we're imagining and and reimagining all the ways that you would protect data in a world where you had agent swarms that were being deployed against your data environment. And how do you go and defend and protect that type of risk? So lots of I mean, many, many layers to this problem, but really just through the lens of imagining of a 100 times more agents than people, how would you work with, store, access, search Mhmm. And secure your data in that kind of world? That's the platform we're building. Do you think we need to move away from the word sandbox?

Speaker 1: Because I don't know if you have kids, but if you've ever taken, like, a five year old and plopped them down in a sandbox, at no point should you as a parent believe that that kid will stay in that sandbox. The sand is gonna go everywhere. The kid's gonna be on the playground two seconds later on the swing set. They're gonna be running around crazy. Sandboxes have never been secure environments. They're not jails.

Speaker 3: Yeah. I think I don't know who's that terminology.

Speaker 1: It's outdate. It's running its course. Well, think it's running its course. Yes.

Speaker 2: On personal agents, how are you processing Oh, sure. The new the new race Mhmm. To be the most helpful assistant ever. And then, specifically, I wanna understand how you think all the sort of agents versus walled gardens war is gonna evolve? Just because it seems incredibly obvious that Amazon will fight to the death to maintain their, like, $75,000,000,000 ad business and and the health of it. But but what's your read?

Speaker 3: Yeah. Super super fun how fast moving the space is. I want to I I want to first apologize because I I honestly did not have the imagination on the consumer front. And so so so I I was, like, I was pretty skeptical of of what the breakthrough would look like on the consumer end. And it it's funny because it was obviously staring at us with with Open Claw. But the but just the form factor and the packaging and getting it all kind of in the right way that both Muse and Instinct have done, and obviously there's gonna be more more more, you know, entries into this space, has just been phenomenal. And and then the capability progress on computer use and browser use is just amazing. And I do think I have a I have a list of, like like I have a wish list of what we need architecturally as an industry. Like, you need to eventually be able to have some better password sharing dynamic. You know, if I can in one click be able to get into any of these products in a very secure way and there's a trusted handshake between the agentic system and and, you know, the tools that I'm using, we're we're gonna need that. Payments are largely solved because of obviously the work that that Stripe and others have done. So I think that's pretty seamless, but but really it's the authentication that we still need to work on as you're kind of ramping up the use of these agents. But I think it's gonna be an incredible market. I you know, the ability to now hand off the whole problem to an agent as opposed to just using it nearly as a research assistant, I think is gonna be a breakthrough for consumer use cases. I'm very excited about what this can mean for all new business models in the economy. I think there's a you know, the Suttrini post, you know, five months ago, four months ago, whatever, had had some of the more, you know, kind of, you know, risky to to capitalism elements and and the market. But you can easily look at the exact other ends, which is I might now be a consumer of five times the amount of local service providers. Sure. As a result of of these types of systems. So, you know, like like if I if I could lower the barrier to just like, I need flowers, like, right now delivered for this thing, and and it handles the whole thing. I I would I mean, at least in my personal experiences, I'm gonna I'm gonna grow my usage of the local economy as a result of this. And so Yeah. Very interesting to see, you know, you might have,

Speaker 1: I think, some experience Cutting out the middleman and installing an even more powerful middleman.

Speaker 3: Yeah. Yeah. Well, or layers and layers of middleman. It might be middleman all the way down. Mean Here you go. I like I I like Yelp as an example is like, how do you sort of triage the local economy as a Yeah.

Speaker 2: But there's something there's something even with even with how good the Internet has gotten in so many ways, like, there's still, like, specifically local providers where I think using flowers as an example, like, many the the place that I order flowers from, it's a good reminder to everyone to order order flowers for your

Speaker 1: significant It's five months until Valentine's Day. Valentine's Day starts now. No,

Speaker 2: but it's like the service is they make amazing bouquets, but the process of actually making an order is like so It's so painful. So painful. That that, yeah, just that little bit of friction. And so people say, okay,

Speaker 3: consumers don't want productivity. They don't care about productivity is what a lot of the bears are saying right now. But they also hate friction. And so what what's gonna what's gonna what's gonna win out? You know, I think I I think that that's the thing is I'm I'm so utilitarian and productivity pill that I I shared the same view of, like, consumers don't sort of value their time in the same way. But but I think it's just then not on that dimension. It's like the pitch is not to save you time. It's just to do things that you've probably, like, added to your remit. And and then there are lots of ways which are, like, that like, on one hand, they're time saving, but on the other hand, they are just, like, literally doing more as a consumer. Like, you know, we have three kids who are always signing up for more things. There's there's more camps. There's more, you I know

Speaker 1: don't know what what was that one for? The three kids. It's just very good. Congratulations. We have we

Speaker 2: have

Speaker 1: nine kids. We have nine kids in the

Speaker 3: future. So so so, you know, you have camps and museums and and and this thing and you're renting that thing. And so so it's not even like you're saving time. You like, if you could just have an agent do a lot of these activities for stress you reliever. Yeah. You're stress reliever, then you actually just do more with your kids. And it's like it's fantastic. We Totally. We we we went to there's these Japanese tea gardens in San Francisco, and I had an agent

Speaker 1: Oh my god. What is that one? That's for Japanese tea gardens. I'm just excited. I'm sorry. Is exciting. Bring castle. Bring me the castle. So

Speaker 3: And we had an agent book the whole thing. It was like we just like added that to our plate that day because it was that much easier to go and execute. So I like the idea of more of the kind of consumer abundance use cases. And so I'm very excited by this as just even more volume in the economy of of just buying and selling. I mean, you're you gonna can imagine the you can imagine the thumbtack, Yelp, TaskRabbit use cases of, hey. I need help right now to to go install this thing. And it's just like all new all new forms of of activity happening. So so I I'm I'm a big fan of the the personal agents. I think it's a it's just full net growth in in the consumer economy. Yeah. I completely agree with you. At the same time, I would assign some sort of low probability that in a few years, we're having this conversation

Speaker 1: and being like, oh, well, actually, what we saw in Q4 of twenty twenty six was prosumers, early adopters. It didn't cross the chasm. And Yep. Those agent personal agent companies are basically getting ready for the enterprise and doing, you know, enterprise level OAuth. And, like, we've seen this you've seen this play out before, and there's a Ben Thompson take about this and whatnot. So it it I don't think it's like a foregone conclusion that will cross the chasm, but it feels like we're definitely the closest we've ever been. You know, six months ago, people were, like, using command line coding agents to, like, order food and being like, wow. And this feels like we're definitely one click forward, but we're still in really good Here here with with a technology like this Yeah. You you have to you you have to be able do a couple of things. One, you have to extrapolate to

Speaker 3: to the ultimate sort of maybe the limit on technological progress. Sure. And so and so if you if you sort of somewhat just just imagine that the agent literally can, you know, with within 98%, kinda perfectly execute these tasks. Yeah. And everybody opens up their CLIs and their MCPs, and browser use is is sort of good enough for all the long tail. Let's just assume that for one second. Yeah. Then then you're you're, you know, you're not really left with the technological breakthrough that's required. The form factor itself is just your mo is just a mobile app on your phone. So we're not we don't need, like, societal level change of behavior, like all, like VR glasses or something. And so I I think this one this one kind of works because it's not If the technology works and the consumer kind of change of behavior isn't high, then it just literally just might be a new way to interact with your mobile device and the apps that you already would otherwise use. So I agree. We'll have to we'll have to see, but it's not like this is not sort of like Oculus VR level, you know, kind of, you know, in of

Speaker 1: fact, I I had a friend who who sort of framed it as more of like a return. He was like he was like, yeah. I'm just I just, you know, I'm just like a business person, but I feel like a gilded age CEO. I'm like, take a memo and then dictate a letter to someone and go and do this. And and it and it feels like you're in, like, great Gatsby or something as opposed to it is futuristic, but it's also like a return and a democratization of a certain style of working that is actually been the way it's happened throughout history many, many times as opposed to Yeah. Yeah. Being entirely Yeah. The point 0001%

Speaker 3: could do this in history, and now everybody can. So Now everybody can. Now everybody can.

Speaker 2: Jordy It's not just the fedora class.

Speaker 1: Yeah. What what is your what is your take on X rays? Do you have a PDUM? What's what's the internal house philosophy? PDOM my my PDOM still remains,

Speaker 3: you know, kind of very, very low. Yeah. So I'm but, you know, I think the the the progress on AI is is certainly, like, I'm, you know, I'm I'm I'm in a weird category because I have low PDOM, but I actually appreciate all of the safety work that's happening. And so and and so it And it kinda has to happen.

Speaker 1: Yeah. A lot of it's, like, cybersecurity and stuff. Like, these things are good to happen.

Speaker 3: And and and I think that that the because of the amount of attention on the safety, I think that's sort of what causes me to have a low PDOM ironically. Sure. Yeah. So so I I see the events that otherwise could be quite, you know, catastrophic or or problematic. Yeah. But because of the attention that I think the ecosystem

Speaker 1: Yeah. Incent it has an a strong incentive to go and attack Totally. That lowers my PDOM as a result. Yeah. So A lot of the stated PDOMs are conditional on doing nothing. So it's like Yeah. 10% if we don't fight it back. And it's like, okay. Well well, if we fight back, what is it then? Then they'll be like, oh, zero it percent then because that's the difference I wanna high PDOM if literally Yeah. If we just fight to yeah. Then

Speaker 2: funny that there's so much criticism there's so much criticism of any safety rhetoric or or talking about safety and criticism of the hugging face incident. But can you imagine like alternative reality where all the researchers are like, I don't care about safety at all. I'm building this as fast as possible. Totally totally. No. It's actually great that we have this powerful Yeah. Technology and the people building it Taking it very seriously. Cared cared almost

Speaker 1: as much about safety five years ago. Right? Yes. Twenty years ago, some of them. But been on this forever. Yeah.

Speaker 3: Every lab has found it like that. I mean, I I think that's a key point. I think the one thread that we still have to figure out, though, is the language will be co opted by parties that don't have the same kind of either level of either view or goals or incentives on the situation. And so so the the one the one risk is is is how does it like, once it once you get that language into the rest of the world, it will be turned against you no matter what. So do you have the right sort of fortification of the industry and of progress to be able to not just basically get the whole thing halted and and shut down? That that that sort of like is a very narrow line between the right way of elevating the risk and and continue to fund all of AI, you know, safety and and alignment research. And then one click above that where basically we will just vote for the data center data centers to never get built again. Yeah. And navigating that is a really tricky kind of terrain. I'm not sure that I've seen a perfect sort of anybody perfectly finesse that. But that's the that's the one kind of, you know, challenge here is is do you end up in a spot where where enough people say, you're telling me it's extremely scary. You're saying that the likelihood is actually, like, high enough that I should be very worried, and I could just press a button and vote this thing to not exist. Yep. I probably am just gonna press that button because, again, so far, you know, your best you're the best thing you've presented to me is that is that you're gonna go shopping for me on my mobile phone. So so, like, that's like, like, we're gonna have to work on that that sort of, know, kind of,

Speaker 1: you know, trade to consumers Yeah. A There's a lot of, like, odd alliances between people that just they I mean, they don't even use the the frontier models. They, like, they they wanna go back to a world where there's no GPT three. So they're like, oh, cool. These people speak for us. And it's like, I don't know that they're actually going to try and go that far. And so a lot of them are saying like, we're not even going to pause training. You know? So Yeah. Yeah. I think are some like, we'll see how things get papered out.

Speaker 3: Yeah. I I I'm I'm still, unfortunately, on the more nervous camp on can we can any country, realistically, but but can we kind of navigate through this where all parties are satisfied? I I think it's Yeah. I think it's just extremely unlikely. Yeah. So I I expect this to be quite messy for for the coming years. I think it's gonna become the dominant 2028 conversation Totally. From a from a presidential election standpoint. Like, it has to be because it's gonna deal with jobs and and GDP Yeah. And safety and and the next sort of major risk that that people start to think about. So and and we as an industry, as a tech industry, I don't think we know where we want things to land. You have this sort of roller coaster of, you know, one day everybody signs the open point. Open source thing, yeah. And then the next day everybody signs the pacing letter. Yep. And like, I think there are some people that can kind of figure out how the two letters are the same, you know, could deliver the same outcome, but but they basically pull you in two different directions. And And and so I I think it's gonna be interesting to see how that how how we kinda thread all of that. Well,

Speaker 1: it's a hundred and forty six days till Valentine's Day, so fire up that personal agent. This is your notice. We're only a hundred and forty six days away, so get ready. Make sure you're onboarded to some Frontier Intelligence personal agent that can make sure your Valentine's Day is fantastic in the next hundred and forty six days. Thank you so much for coming on the show. Nice to see you, Aaron. Always been ahead. Too long. We'll see you soon. Alright. Good to see you. Have a good one. Goodbye. Let me tell you about Railway. Railway is the all in one intelligent cloud provider. Use your favorite agent to deploy web app servers, databases, and more while Railway automatically takes care