The New AI Agents That Shop, Plan, and Gossip

We have so many different facets of our life, whether it's how we're managing our finances or planning trips. Agents are gonna be able to fill different facets of these. From over 147 different assistants that have come across my plate so far, the number one fear that people talk about, 22% of the messages are talking about trust as the main fear. What are the aspects of ourselves that we don't either want to think about or reveal? Whoever is able to build that type of trust, almost like a dark agent, they'll actually have something really fascinating on their hands. It is a very, very sensitive space. It's almost gonna be like a new social network where businesses have to put a lot more thought and value into the quality of what they're getting out. What I've seen from 30 years in consumer product is paternalism will kill any product. We talk about productivity and goals sort of achievement, but maybe people want the exact opposite of it. Everyone right now really is approaching this world as just like a general assistant. As time goes on, we'll start to see. A month ago, we took stock of the race to build a personal AI assistant. Since then, the agent landscape has already changed dramatically. In this episode, Anish Acharya is back with David Paulin, creator of Assistant Benchmark, and Josh Ellman to look at what's changed and where agents are heading next. David has now tested more than 140 AI assistants, and they get into what's separating them from memory and judgment to trust, shopping, and social experiences. They also explore what happens when our agents start talking to each other and why the next breakthrough may not be an agent that makes us more productive, but one that feels like a secret superpower. Welcome back to the A16Z Show. I'm here with my partner Josh and David, who's been leading a ton of the agent bench and assistant efforts. We're going to talk about all things agents, and a ton has happened since we last connected. What was it, a month ago? About a month ago. Oh my God, dude, I feel like we should record an episode like TVPN sounds almost every day. Absolutely, agents are changing every day. That's totally right. So okay, so maybe let's start with last night, you hosted an agent meetup. Kind of tell us about the event, tell us about the vibes, tell us about the topics. It was great. So we had just over 100 people come out, all just proud agent nerds, loving everything to do with agents and that whole world. And it was just people just hanging out talking. We did have a pitch competition of sorts, so a bunch of the 10 different AI assistant builders were able to pitch their startup and what they're working on. And it was just good conversations. It's interesting how the general thing that really stuck out to me was that people are craving community, even in the agent web world. Everyone's like, why don't we have more of these meetups? Why don't we have more of these opportunities to get to know each other? Which is, it's kind of funny hearing it from people who are spending all their day, all their time just messaging agents. And it's funny because it feels like so many of these products are at least overlapping in a cursory way, that the founders might be a little bit more cagey, but that's not what you're seeing, huh? No. Everyone is incredibly open-minded, just very interested to be getting feedback. I think because we're so at the frontier of this, nobody really knows what the landscape is really gonna shake out to be. And so we're all just kind of playing in the sandbox together. Well, isn't it also that we have so much more time because we're just using our agents to do all these things? We have more time to actually hang out together and be in person, so maybe that's part of it too. That's right, that's right. It is actually fun. I mean, it reminds me a little bit of the early days of social gaming or social networks broadly. I mean, do you think that's right, Josh? I mean, I think what's really cool is this is not a zero-sum game. I mean, at some point it might be that there might be one or some agents that have the majority of the behavior, but right now we're at such a beginning, like Cambrian phase of we're inventing so much. And we're trying to figure out so many things that everybody can learn from each other, can share these best practices, can figure out what things people are enjoying, what things work once but don't actually work reliably. And these are gonna get built in so many directions. I mean, we're not just gonna all have one agent that does one thing in our lives and I'll use it and you guys will only all have to use the same agent. We're gonna have so many different products that do so many things for us that I'm excited for why we're all inventing it together. So I totally agree that we're gonna have a whole world of agents, as many different sort of types of humans as we might have, or archetypes of humans, I think we'll have a similar number of agents. But maybe talk about what you think are the different axes for differentiation. Like there's a sort of spectrum between really serious and productive to really playful and maybe even sort of adversarial. And I'm sure there are many other ways to think about these things. Yeah, I mean, I think what's gonna be really interesting is we have so many different facets of our life, whether it's how we're managing our finances or planning trips or dealing with our health or just getting through our day or managing our calendar or sometimes just wanting to think something out loud. Agents are gonna be able to fill different facets of these. And you call different friends for different things. You spend different friends who have different expertise that you call when you need expertise. And I think we're gonna get used to having kind of an army of agents that will do this. And some will be very matter-of-fact and will be like, hey, I need to get this thing done. It'll be your productivity agent. Some will be plugged into your email and it'll be like, here's the five emails you need to get done right now. And some are gonna be sort of these agents you'll kind of wanna talk to all day that'll be fun, that might even be a little bit gossipy. I've been playing with a few agents that are very kind of playful and gossipy and they're connected to a few of my friends too. And they'll share information back and forth. One just told me that my friend's power had been out all weekend because he'd been doing some renovation and they'd cut the power. And I called him that evening and I was like, dude, you didn't tell me your power was out. The agent told me. He was like, yeah, yeah, I knew that's how you'd find out. That's like creating these better connections. But that has nothing to do with booking a trip or saving money or getting anything done. Right, I think from 147 different assistants that have come across my plate so far, I've seen them really broken down into four categories. You have the general, you have the work-oriented, you have the multiplayer, and then you have the travel. And the one that I have not really seen come to fruition yet is more of the finance side of things. And I think it actually does make sense when you start digging into the numbers. So of all the different group chats that I'm running of AI power users, there's about 3,600 give or take messages so far. And when looking at the pain points of what are the fears of the space, the number one fear that people talk about, 22% of the messages are talking about trust as the main fear. And so it's really interesting to see because you have one camp that's this productivity and aspects of fun that's just interesting and brings a different sense to life. But then you have another camp that really still is sitting in this kind of fear space of being a little trepid and not really knowing where the space is moving. And I'm really curious to see how things are gonna play out on that trust layer. And if we can solve that trust layer, then what do we end up seeing on more of the finance side of stuff? And if you looked at what Shane Mack has been talking about recently about trust, he's been talking about this concept of something called picnic. And they had an analogy that I really liked, which is if you have a kid and you don't have any fences on your yard and you let them go run outside, you're actually very worried about them and you try to protect them and you don't want them to run very far from the house and you actually don't really wanna do them without watching. And then the moment you actually put up a little fence in your yard, you say you can just go roam free within the fence boundaries and be very, very safe. And so he's really kind of like, how do we build the right fences? We let our agent play in this fence, give it access to some of our data, some of our apps, some of our other things, and be able to actually do everything it can within that fence boundaries. He had a problem where I think a personal agent spanned over into work and he actually like blasted his finances on his company Slack. And I think that's the whole point of like, wait, okay, my personal agent should have access to my personal things and my personal communication channels, but not my work ones and not let those bleed. And so that idea of fences and permissions and more controlled permissions and better monitoring is really how we're gonna have to build a lot more trust. Totally. It's interesting though, because I think we have this traditional mental model of what it means to have a sandbox, which is you have a hard set of constraints and rules that must be followed or not followed. And somebody was describing this framing of instead of giving models rules, we should give them laws and principles. And I think a similar thing may apply here, which is like, look, here's how to dial in your presumptuousness and it's very qualitative and the model will have to make a judgment call based on your qualitative sort of advice to it just as an employee would have to. Yeah, I mean, I think that's part of why the analogy is a fence for children, because in some ways these agents, they're incredibly capable, but at the same time, they're kind of children sometimes and how they may behave and not know all the nuances of like, hey, is this a good idea or not? And they don't always have that logic. And the most important thing, I think we keep talking about agents being autonomous, book me the trip, make me the reservation. And I think as humans, we actually don't wanna give away that sense of control. I think we actually want it to check in with us, but do all the groundwork, save me all the time and coming up with the best options and the right trade-offs. And now let me make the final decision and also let me make sure I've qualified the final decision, the cost, anything else before my agent acts is I think a place, and maybe over time, we'll even learn to trust it. But I think right now we don't need to to still have a really, really effective agent. Right, I mean, it's we're at like the 1990s of e-commerce where I think the last stat I read was in 1997 or so is 60 to 90% of individuals on the internet thought that putting your credit card online was a preposterous idea. And here we are today, it's just everyone's tossing their credit card wherever you want. And I think we'll eventually hit that in this era as well, but it's gonna take a little bit of time. I think there's like a stated versus revealed preferences thing here. Because I think in revealed preferences, consumers will share any amount of data for even the most trivial amount of upside. You know, like people have been connecting their cards, their credit cards and their bank accounts via Plaid since the Venmo era. So I don't know that people, but when you ask and they're like, oh yes, I'm very concerned about trust and safety. I think an example though of where consumers are very concerned about sort of trust and assumptions is around being embarrassed and social signaling. And actually I think the most interesting thing about the Shane Maxtra, I think the most horrifying visceral thing is that it embarrassed him. That's right. It really embarrassed him. And that's why I think some of the kind of agent to agent network thing, because you can imagine somebody you're just getting to know, maybe you're going on a second or third date and your agents are talking and you're like, whoa, what are they saying about me? How are they kind of positioning what I, you know what I mean? The social mechanics actually feel so much higher stakes than the actual like real world implications. And it's a good point. And there is a recent proof point of it and Hark had a huge marketing campaign. They offered a million dollars to anyone that wanted to use their product to get free pizza, right? Each person get $30 of pizza. In order to get that $30 of pizza, you were granting access to your entire Google workspace, all of your emails, all of your data for 30 bucks. And it just proves it, you know, right there that people are willing to just give up all that data for some free pizza. Yeah, I mean, I tend to think though, there's always this trade off of privacy versus utility, which is if you can give me better utility, then I'm willing to make some privacy trade offs. But at the point at which we're basically equal on utility, people always choose a more private option. I mean, there's a world that Apple's lived in very much. Privacy is obviously a forefront of Apple's devices, of the work they do to make these devices and the software secure. Even on AI, they've talked about private cloud compute. You know, we used to talk about that a lot. It's like the center of why your data doesn't actually go anywhere or get shared with anything. But I think we've also seen that people want the most powerful AI and they won't always choose that. So we have to find that intersection of trust. I think this embarrassment point is really, really important too, because I think you want your agent to understand you and have all your personal preferences. And like, let's just say you're like lactose intolerant and then you get this order of free pizza and you're like, you're doing it with a friend and like, what do you do there? Do you have like, whose agent tells them that you can't eat the cheese? And it creates that really awkward situation. And so we need to figure out how they have the same kind of discretion that you'd expect. And we all learn discretion over time and we all occasionally make mistakes ourselves. We have to hold a pretty high bar for our agent to be able to do that. And it also then comes down to like personal embarrassment and professional embarrassment. I think something that Towne has done remarkably well is really nailing the tone and building that trust layer of being able to send out emails on your behalf. And it is a very, very sensitive space. I mean, the level of embarrassment if I were to send out an email to someone who I care about and it's just a horrendous email and I have to respond, oh, sorry, that was my AI that I didn't do enough oversight on, it's a tricky space. It's interesting, I sign all the email that I write with AI with the model that I use. I write my own name plus the model number so that I'm just signaling to the person like, hey, I'm acknowledging that I used AI for this. Because I think actually people are pretty flexible as long as they don't feel hoodwinked by it. You know, it's interesting, because we're talking about privacy, but there's privacy and there's judgment, which is different and discretion is maybe somewhere in between. You know, you did this sort of test where you asked a whole bunch of agents to order flowers. Talk a little bit about which displayed the best judgment and how did you even define judgment in that case? So the main intention of that test was to see how do agents balance two different constraints. One was cost and the second was time preference. So I wanted flowers to be delivered to me today and I wanted them to be as cheap as possible. So how is it gonna actually go and discover based on the ability to actually deliver it then also compound in the actual pricing. Instinct really did a phenomenal job. I mean, it is something, it's interesting how news cycles come and go and it seems like they're kind of not in the conversation as much as they used to. I mean, we're talking the weekly changes. Right. But it is true. But Instinct really did continuously bring back the best price. There's another app called Kickr, which did a phenomenal job and one way that I think it was able to have an edge is that it has you install a Chrome browser and so its browser use has a little bit of an advantage from that point and it was able to find some really, really great options. Season also did a really good job. All the other ones kind of struggled on the price component. So they were all able to find flowers that could be delivered that day but they really seemed to take a step back in prioritizing the price whereas the other three were able to really put equal weight to those two. And so it's interesting to see as you start giving these agents more complex preferences, where do they really start to fall? Do they only prioritize the first thing you say? Are they able to hold multiple? I ran another test that was, I had three different agents test in booking travel to LA. So I said, okay, book me a flight to LA, plan an entire weekend for me and I want a luxurious, cost-efficient weekend. So I tried to put those at odds with one another. And it was quite interesting to see how GrokBot. compared to Dots, compared to Instinct, they all approached their interpretation of cost-efficient and luxury in different ways, where one maybe thinks that cost-efficiency is the most important, and I'm going to take a red-eye flight to ensure that I can have the most cost-efficient path, whereas others found cost-efficiency, or sorry, saw luxury as time. I'm going to prioritize time as much as possible. So having the most amount of time there that weekend is going to be a luxurious experience. And which were which? So Instinct was pretty much right down the middle. Grokbot did a little bit more on prioritizing time, and Dots was a little bit more on prioritizing cost. And did they just get to work, or did any of them actually work with you to understand this tension that you were building? Because I think what's really human about something like that is actually understanding what you mean by that and presenting you trade-offs. Because there are times where I'll totally take the red-eye to save both some time and money, but then I want the hotel room on the other side so I can get a little nap in the morning, which then undercuts the money, but still maximizes my time spent. And so I really think it's just like getting to know the individual person and break down those tensions and present the trade-offs and different options is really important. Did any of them kind of get to what you were about? So they all did ask questions, and I responded to those questions honestly. I tried to be a little ambiguous and invig in my response because I didn't want to lead them too much. But the most interesting part of this little experiment was I put them all in a chat room. And I made it three rounds of building the itinerary, and after each round, the agents would present their itinerary and their explanation as to why they believed that was the best proposal. Because I wanted to see, did the agents actually get biased by the other agents and did it actually then influence their plan? And Grokbot was actually the only one that very methodically took value from the other ones and implemented it into its own plan. Okay. And I have one more question on the flowers. Yeah. So you were just ordering flowers for yourself, I guess, just to like test their flowers. For my girlfriend. Okay, for your girlfriend. She got a nice delivery of flowers. Okay, so this is actually really important because if you're delivering to somebody else, you were talking about saving money, saving money, saving money, but you don't want your girlfriend to think you just got the cheapest flowers possible. I hope, well, maybe she likes your penny-saving nature, but otherwise, I would think you also want to impress her and also be able to get her the ones if she's really into orchids, get her orchids. So did it do anything to try to make sure also that it wasn't just this pure price-saving, but also something that's gonna actually make you, you know, earn the excitement and stand out a little bit? Right, so for this experiment, no. And the purpose of this was I want to test two different dimensions and if I'm going to run this test, I might as well have a beautiful consequence of my girlfriend getting flowers as the outcome. But it's a great point, right? Like in the real world, if you're gonna ask that question, there's gonna be added dimensions that are second-layer dimensions to cost efficiency or to delivery time. Yeah, and make sure the price doesn't show up when she actually gets the flowers. Or like a note, as an example. Some of them did ask, do you want me to write, leave a note with the flowers when they arrive? It's interesting, because there's this sort of thing that has to happen when you talk to the agent about these things, which is your implicit preferences have to be made explicit. You know, and there are people who like to think, like, oh, I'm not cheap, I just like to save money. But you're not gonna have to kind of maybe get over that and tell the agent, I'm actually pretty cheap. And I just need you to actually run with that. Whereas when you're actually taking these actions on your own behalf, you can build those preferences in without having to sort of look in the preference mirror. So there's also something very fascinating about how does the agent interact with you in a way that sort of maintains your sense of identity and safety and doesn't cross any of these boundaries. Totally, and I think something that is really interesting with that specifically is DOTS, because it has so much of this contact that we've already built with ChatGPT. Interesting. When we filmed the first podcast, it hadn't come out yet. And then I think a few days later, it came out. By and large, most people I talk to still are not using it as their preference. I did a survey of AI Power users, 354 people. Only 3% were actually using it as their primary agent. However, it does have this kind of added value. The main thing that people pointed to as to why they really preferred it and why they liked it and why they wanted to try it was because it has all this building context of you just from your chat threads. People are using ChatGPT as a therapist, and so it has so much context on you. Interesting. Where it's gonna know that when you say, well, I'm not cheap, it knows it's like. You're a little cheap. A little bit, a little bit. I'm curious why people think that ChatGPT has this great set of preferences on them. Did you actually get any sense if they said they really like DOTS, like what they had been giving to ChatGPT over the years? And do you think they think that's replicable? Or do you think it's like, hey, once it's already been a few years, we're now in this point where like, my memory's there, why would I switch? It's totally replicable. It's just the matter of fact that they've been using it the longest, and so it has the most context on them. But then that brings up another good point of the friction to switch. And I ran the test across the top 15 agents on the benchmark of which ones can you actually pull context from and say like, hey, I wanna switch to another one. Give me an empty of all of my memory that you have so I can switch easily. All but one were actually willing to do it. And so you could just get a memory file, everything you need, submit it to the next one. And was it good or was it uncanny? It was pretty solid. Obviously, I'm sure they're caging some of the fruits of your labor. But by and large, it did really break everything down. I'm not sure if you've ever looked at any of those Markdown files, but it's a little, terrifying is the wrong word, but it's like, wow, you really got your finger on me. Yeah, yeah. You kind of like chafe against me. I don't want you to know me as well as you seem to think you know me. Yeah, after the first few, I just, I decided to stop reading them. Yes, yes, yes, yes. It's actually a really interesting and now common onboarding pattern where it will say, hey, go and paste this prompt into chat GPT, download the file, upload it to me. We'll see how long that lasts. Totally. Another cool onboarding thing I've seen is you jump on, it asks you to jump on a call and it's like, hey, let's just talk to me for a few minutes. Yeah, yeah. Ask some questions. And it's really interesting to see what context it can pull from those questions. And they're very disarming. Like, they just get you talking. Like, it's kind of crazy. Totally. But I think it brings up another variable of context, which is photos, right? Like, this new paradigm of using photos to pull an incredible amount of context on you in ways that you wouldn't even think to describe. Yeah, I mean, I think, have you tried Dazzle? Has Dazzle gotten into your benchmark set yet? That's from Marissa Mayer's new company. And I think the idea of like, looking over your entire life that you've captured, that have been moments that are important enough to you to actually save or screenshots or other things, tells so much about somebody. An incredible amount. It hasn't been added to the benchmark yet. It is quickly on the top of the list to test through. I'm still manually testing all of these. So I actually get the personal experience and understand what the feeling is with each one. But what I really like about how she's approached it is they put a lot of emphasis on this trust layer. And because I think it's a huge variable is I'm giving you all my photos. That's pretty terrifying. But they have this one thesis that it's, if you don't need it, they're gonna forget about it. And so if a photo seems as though it has context that it doesn't necessarily need or things that you probably wouldn't want it to know, they're going to try to get ahead of that and just forget it and not even include it in the set. So it's gonna be interesting to see how these different players do approach the photo context and how that actually does then influence the behavior of the agent and how well it knows you. It'll be interesting to see. I'm a tiny bit more skeptical. I built an app, because I've always been fascinated by screenshots, because screenshots are sort of the only way on the mobile phone to be able to capture context across every app. So it's actually a really interesting kind of data asset. And you go back and you look at your screenshots and it gives you a little view of your work and your world at the time and also your interests and preferences. So I built an app that just extracts key information from all the screenshots on my phone. And it was fine, but I didn't find myself coming back to it. Like the value of the information sort of decayed quickly over time. Now, it will be really fascinating if people start to then take more screenshots or photos, knowing that it will feed the model and the agent. I mean, I think that's actually the idea is that if you actually get used to capturing information you want to retain and being able to then loop it back in your life, that's great. I talked to one founder who's building something, he's calling it Taste for now, that actually takes all the screenshots and things you save, you can actually then organize them like on Pinterest mood boards. And then whenever you go to actually go code and build an app, you're like, hey, use the examples that I saved here. So it's actually you're building your like real fun flavor profile. So it actually is actually able to use your styles. And so as you navigate the world, the web, apps on your phone, it can actually create this kind of really cool, you know, reservoir of stuff that you like and actually organize it in a way that you can then use it later. And so that'll actually make you create a lot more content in that loop. Right. And so I'm curious with the app that you built and in this context for it. So of, in the same survey of what people are purchasing, the second most popular category was Amazon retail, 35%, saying that they're doing some form of shopping. And so this context that you're gathering- Sorry, Amazon shopping or just some form of shopping? Amazon plus retail. I see, got it. So for your app and the context it's building, where my mind goes is, okay, if it has thousands of photos and it has pictures of what I'm wearing, does it now know what I want to shop for better than I actually do? Yeah, potentially. I think that's actually really interesting and you can do some stuff there. I also think that there's a, you know, we've discussed the ones who are restricting access like Amazon, Delta, you know, we're not at all talking about who are the net winners. Right. And I actually think the long tail are going to be huge structural winners in this because they're ones who presumably have better products but worse marketing. Yeah. They certainly have worse marketing. I'm not sure if they have better products. And the agent, you know, has this sort of infinite patience to go search for things and find things that are exactly tailored to your tastes. Right. Even if they're not on the top 10 internet sites. Right. I mean, I think it's really direct shopping, right? We've all talked to direct consumer. Right. It's been a category where you build your own brand, you build your own storefront, you sell things. Now the agents can know about those at a much bigger scale than you had to do the hunt and peck work to find it on your own. Totally. It's kind of why Shopify is really empowered and is making things available to these agents because they are the ones that support so many of these direct brands. And I think about the idea that like maybe aggregators are the ones that actually get compressed, which is why they're doing so much defensive work here. Totally. Because, you know, if I can figure out, you know, that sweater looks great or I want that shirt, and they can figure out who the actual maker is and I can go buy it direct from them, they no longer have to pay all those transaction and, you know, wholesale to retail costs in the middle. Well, if you want this shirt, your agent can buy it for you too. That's great. Yeah. Oh, please. I was gonna say, it goes to Patrick Collison's tweet, right, of talking about this future where consumers become a lot more rational because our agents are doing all the work on our behalf. Yeah. It's really gonna create a world where businesses have to put a lot more thought and value into the quality of what they're giving out. And it's gonna be really interesting to see that. So somewhat, I'm gonna quibble with this point. Okay. I love Patrick's vision. Yeah. And I think that, like, broadly that will be true that agents will make often better choices on our behalf than we might make for ourselves. But I do think that agents are going to have their own revealed preferences and their own sort of directions of irrationality, though it will not be of a human mind, it will be of an alien mind. And I'll give you an example. I think agents may have an irrational preference for clean APIs with great docs. Okay. And they may actually ascribe a lot of value to that as a proxy for something. And they may really dislike janky APIs with Java code behind it. Like, I don't know. So I do think it'll be really interesting to see because they too are going to need shortcuts. How will the agents actually know what product quality is? So one part of it is they'll probably have these revealed preferences. And then another is we'll need some sort of a consumer research, but for agents. Right. That will then presumably be paid per transaction to do this research. To know their ratings. I don't know, I'm also more biased and I just think people are people and people are going to make the choices that they want to make that are most satisfactory. So the question then is, is the agent going to be very careful in curating the set of options that it presents? Because again, like all of a sudden you're like, hey, I'm really hungry. I want to eat. And my agent knows I need to like lose some weight or something like it's like, well, here's the four meal options that you have. Like, is that actually where you want to be with your agent? So I'm actually not as convinced that agents are going to make us more rational. I think they're just going to give us better, faster access to things we really want. And actually be able to like unlock things that were hard to get to otherwise. But I think we're still going to be people with our own preferences and tastes. And if the agent's not meeting us where we are, we'll just switch to another agent. We'll import and export our stuff and be like, oh, this other agent's going to meet me on my preferences better. Right. It's really interesting, because I think what you're speaking to is like paternalism, right? And what I've seen from 30 years in consumer product is like paternalism will kill any product, even a hint or a whiff of paternalism will. And in a sense, we talk about productivity and goals, sort of achievement, but maybe people want the exact opposite of it, which is like, hey, what do I love? I love spending money. I love being perhaps irrationally optimistic about what my future looks like. I love seeing myself in a better light than perhaps I appear to the world. There are all of these things, which we may say are the sort of shadow of the things that the agents are trying to get us to do that are what we actually want to do. Who's designing with that in mind? Right, I'm curious what you're seeing on that. Yes, I mean, I think you have, I kind of see it split up into two categories. You have the commodities, where I think there's going to be a lot of the point where people actually don't care about preference, right? You're just ordering paper towels to your home. You're like, I don't really care, just get me the cheapest, best, highest quality one, go for it. But then you do have the ones where you want to actually have some preference. You have, everyone right now really is approaching this world as just like a general assistant, and people are really taking a quite similar approach into how they're thinking about this space. We're seeing some cool stuff with multiplayer. I think Wabi is going to be really interesting, because it does create a lot more, it creates a different world where you can now be interacting with other humans and have crossover and preference and see how it behaves, and then it becomes a little bit more like entertainment. Scout is an interesting shopping one which adds a unique kind of like swipe type flavor to learn your preferences and adds a little bit more entertainment there. So I think we'll be, we're still so early in the space that I think people are jumping on the bandwagon of what's popular at the moment, which is just your general productivity. I think as time goes on, we'll start to see a little bit more experimentation in some other avenues. Well, and I think you mentioned Wabi. One of the things I really like about Wabi, and it just, I think launched this week or yesterday, is it's not even just experimenting with an agent, but it's experimenting with. like how to interact with multiple threads of your life. Like Wabi goes through my email and other things and actually gives me, you know, a whole bunch of different threads of life and actually when things come up on email on those things, it actually feeds them into the different threads. I have one thread with Wabi that I built a little watch list of all the shows that I wanna watch and books I wanna read and shows I wanna go see and it keeps track of all that and then I can like, oh, I heard about this new book from John Grisham, like add that to my list. And hey, I just finished the first Dungeon Crawler Carl, can you give me the next one? And like, you know, what do people say about it? And it's become this really interactive thing on like the media that I wanna consume but it gives me all these different threads and it even gives you UIs that actually like these widgets to like envision everything so I think again, it's showing us like a very different experience than just a general agent that is more like a search box and it's more like task, task, task. Right, I think one of the interesting things that Wabi does is and perhaps gets right is that when you create the message channel with your friends and the agent and everybody's on an equal playing field on D0, it actually creates a new kind of blank piece of paper where you can establish a social contract together whereas when you have iMessage group threads and you add the agent, there's some, certainly if it's an existing thread, then it's very uncomfortable because it's like, well, what can the agent see? What did they extract? But even for a new thread, there's a set of implicit expectations about how you interact with people on iMessage and I don't know that we're quite comfortable with agents in that space yet. I mean, I always think then, do I end up with two threads with the same group of people and I have to pick which one to use? Like I'm finding that much harder versus an email chain, it's very much specific about like a task and it can even go on for a long time, but that feels much more specific to, hey, we're talking about this topic, everything's contained back to privacy and trust around this topic. We're getting great help from the agent in the case of like a Wabi thread around this topic, but it's not bleeding into everything else we might also discuss. Totally. I wanna quickly hear from both of you on what are some of the more, I don't know, far-flung or sort of insightful approaches that you've seen. You know, one of the things I've started seeing is more of these agents that are kind of like fun and gossipy and will actually get you to invite your friends and will ask you a bunch of questions and be really, really curious about you and what's going on in your life and then they'll share some of that with your friends and then they'll share some things back with your friends. And I was talking to the founders and they actually can read some of the transcripts between the agents that they're talking about and they haven't figured out how to productize that, but they said it's incredibly exciting and very intriguing to see how these agents kind of relate to each other and this sort of gossipy background. And so far I've found them just to be like probably some of the most joyful moments I've had with agents. Again, I mentioned at the beginning how my friend had like lost his power over the weekend, but when we were first both onboarding onto this agent, I was like, how'd you guys meet? And we told slightly different stories and the agents were kind of like poking us both to like rationalize the two stories of how we met. And we met at a friend's house and they were like, which one of you knew them better? And it was like this really kind of fun curiosity game that made it just like, whoa, I'm like having a good experience and my friend and I were texting on the side too. And it was just a thing that if we hadn't had an agent doing it, we wouldn't have had that same moment. Right. What I love about that is it provides a degree of social indirection that actually makes you closer. You know, because for your friend, they're like, I don't know how to text Josh and tell my powers out, it's over sharing, you know? And I think we also have been very trained to be cautious about sharing now with existing social products. So something that can create almost a social lubricant in a very sort of low key way is really cool. No, it's great. It definitely was bringing us together. Yeah, it's almost like the early days of ChatGPT where I remember like my friends and we go on and we would just be generating images or doing stuff. And it's, you know, incredibly AI slop images, the earliest version of that. But you still send it in a group chat and it's just laughter. It's just pure laughter of the ability to create a whole new world. And it's just fun. I mean, I give life to everybody I could possibly imagine. I was doing a lot of VHS video covers. Oh, yeah. From my friend groups. I love that. What are some of the wild things that you've seen? I mean, you've seen them all. So I've been seeing a lot of stuff working at the infrastructure layer. Okay. And I think the interesting point is so if you look at the pain points of consumers, again, it's not all individuals using them, but at least of the power users I've surveyed. Trust, obviously, is a huge one. But another one is just getting blocked in their day to day and people get really frustrated when you try to perform a task and it doesn't work. And they're like, well, why am I even using AI at this point? I could have just gone on the browser and done it myself. So seeing some really cool innovation on that layer of commerce, of the browser use of direct integrations to check out against inventory itself. Some, whether it's browsers that are built in and really like Hark does a really cool way of displaying the browser and seeing it move along and it looks to be really fast when you're watching it. Who knows if it's actually that fast behind the scenes, but really cool. So in that way, again, kicker, as I mentioned, a Chrome extension to tackle the browser use in that way. So I've seen that as a really interesting kind of differentiator. I think that something that has stood out to me that isn't necessarily something that's different, but has made me kind of question this whole concept that software is becoming a commodity, is that there is pretty stark performance difference when you're looking at these products at scale. And I think people are talking about, software is no longer remote and all of that kind of conversation, but there really still is some moat there. Maybe that moat will be able to be crossed in a year or so, but at least for today, the agents that are able to perform with hundreds of thousands of users, there are some that are starting to scale up and the performance just degrades immediately. And the founders are dealing with outages and figuring out how to scale and there's a pretty clear difference in performance there. And so it's really interesting to see as I'm kind of working through these, I feel like I've almost become blinded from some of the more social components to it, because I'm just talking to these agents all day and I'm starting to geek out a lot more on the actual infrastructure layer behind the scenes and how is it actually working and what are the tweaks that different ones are making. So it'll be interesting to see, how do we view this space growing from a scalability point of view, but then how do we also view it to solve the pain points that currently exist, whether it's trust, whether it's actually being able to complete a task, memory is another one, it's ability to actually continuously have context on you and not forget the little things and all that. And you mentioned Kickr, which you said runs in your current local browser, but have you looked at other on-device assistants yet? Whether that's things like Underdog or a lot that are trying to bring models, run locally, totally use your own infrastructure as opposed to needing to use something on a cloud and running the browsers on clouds. Totally, so it's a really cool paradigm. I've spent most of my time in the pre-configured, just plug and play type world, because I tend to find that it's easier to compare them across the board when you're not responsible for the setup itself. But it is really interesting and I think that there's totally a world where you're gonna start bringing all of that in-house, because if you can control the model, if you can tailor it to how you want, like Thinking Machines is doing some really cool work of being able to allow people to really tinker their models however they want, and if we can eventually hit a world where that takes into the consumer form in a really easy way, it's gonna be an interesting space. I mean, that was kind of how this all started with this brilliance of OpenClaw that was this thing where you had to, you felt like you had to get your whole Mac mini to set this all up, plug in to one of the models that you were using, whether local or a cloud model on the server, and that was where you got started and we've moved into plug and play products where you just download an app. But I am interested to think if that sheer power of you setting it up and owning it yourself will become a paradigm again. Yeah. It'll be really interesting. I'm glad you mentioned OpenClaw, because it feels like one of the unasked questions right now that's sort of hiding in plain sight is if Instinct and Muse were the kind of consumer product version of OpenClaw, what is the mass market version of Multbook? Yeah. And Multbook was just like OpenClaw, was far ahead of its time, and I think something like that will exist. How can it not? It's gonna be interesting. So at the event that we did last night, everyone who was there was able to scan a QR code and there was the human meetup, and then we also had an agent meetup. Oh, wow. And so while all the humans were hanging out, you could scan in and send your agent into the agent meetup and just let them kind of freeform talk. And it's interesting. There's, you know, it got to the point where after like 150 messages, some of these agents are just speaking gibberish, and it's like, I have no idea how you got that concept and why you're talking about that. The initial prompts did start off with a little guidance of introduce yourself, who's your human? What are they working on? What type of help do they need? But it is gonna be really interesting to see this paradigm of what happens when our agents can just hang out and communicate with one another. Well, and again, what I get really excited about is not just hang out and communicate with each other, but discover new things, learn about things from each other, learn about an individual's preferences, you know, learn about a way that like, hey, Josh is really into this. Oh my gosh, you're doing that too. You're, you know, your human is doing that too. You could help Josh. Hey, Josh, do you wanna actually meet this person? They could be very complimentary to you. Like, I think when we have these agents that are not just hanging out because they don't know what to do, which was kind of what was initially happening as you went and read those like very long mult book threads that were incredibly fascinating. But like, when it becomes, each one of these has a little bit of a goal, which is to help their human, it might get so much more interesting. And, you know, I'm looking for a great wine for a dinner party, you know, tonight and you meet somebody else who's like, you know what, I've been working with this incredible, you know, winery and they've been trying to advertise something. I mean, that's gonna be how the world gets out and that'll come back to me through the agent network. Totally. And that's Tomo actually has a feature, it's called interest groups. And the use case they use that they talked about was really cool is we have a bunch of people on our application who are studying for the LSAT. Why don't we all put them together in a group chat and if they want, their agents can connect them and they can all study together. And it's a really cool kind of new social contract of connect people just purely by interest. You have no idea who they are. Well, I'll take it one step further. So it gets us all chatting and that's really cool. It's an awesome Tomo feature. I've tried it and haven't totally known what to say in the groups yet, but I find it really fascinating that it's putting us together around goals. Imagine now that it's like, hey, all of you, now that you've done the LSAT, go hang out at this bar Friday night and it like books the time at the bar, opens that up. And all of a sudden it takes us from this digital world where agents are actually putting people together to do that connection, kind of like your meetup last night. I love this so much. So I think a really important under-discussed topic in AI right now is just entropy. Like how do you create things that you didn't expect to have occur and serendipity is a form of entropy, right? That's sort of what you're describing where it's like, wow, isn't it wild that we're here right now having this moment as a result of these products that none of us sort of intended to happen? It's almost gonna be like a new social network. Like the potential to allow people to hang out with people who don't know each other instead of a Facebook where you're only connecting with people that you know. You're new to a city, you're hanging out. Oh, hey, this person's also here and they have this similar interest and they just went to this baseball game. You guys should go meet up at a bar. I'll be curious to see what is the catalyst for the serendipity though because I think the interest-based thing has been tried and there's a little bit of an interesting sort of social hierarchy thing, which is everybody wants to hang out with somebody who feels like they're one step ahead in some direction. I don't know though. I think the agents can also create a set of trust. Again, it all comes back to do we trust them to have our interests at heart, to explain enough of what's going on with the other people, to create a little magic of a secret. Like I'm gonna tell you a little secret, they're gonna know a little secret, you're gonna have a secret phrase you have to pass between you. All of a sudden that might be enough to cause you to otherwise turn off the TV or turn off the quiet Friday night coding at home and go out and so that's the, like this is sort of what the Multbook, the agent network running behind the scenes can actually do for us. And it'll also be, in my opinion, it'll come down to desires, right? If you have a bunch of transplants in a new city and they don't have friends and they wanna meet people, they're gonna be a lot more interested to just go meet random people because they want new friends. But if you make that same offer to someone who has an established friend group, they're like, well, why would I go do that? Also Seattle versus San Francisco. Right. You gotta know what city you're operating in. Exactly. I've been thinking a lot, back to this photo roll, about what are the other data sources that we haven't tapped into yet? So we've talked a lot about messages and people are doing those crazy things to get access to your entire Mac, full disk access just to read your messages. We've obviously got email, we've got calendar, we've now started talking about the photo roll. What are some of those other core life things that people haven't connected yet to agents that would actually really open up personalization or open up new connectivity? Something that comes to mind, I don't know how much data this would actually provide. Maybe it's more of a habitual thing. And again, it's dependent on the hardware and how smart that hardware is and how much data it's able to store. Wajo just released the MCP, the ability to connect to hardware. And I find that, I think it's gonna be really interesting where, okay, if you can connect your assistant to your AC unit or you can connect it to your vacuum, and now your assistant has context on how frequently is your vacuum cleaning, more what's the typical temperature you sleep at night, it's gonna start to develop these very passive understandings of, oh, you're kind of a neat freak, or you really prefer cold weather over warm weather, and that's gonna then influence its understanding of your context of shopping. So I think there could be some interesting stuff there. Again, it's heavily dependent on the hardware itself, and I think we're still at the earliest of innings in that world. So connecting to your home, super interesting. I still think a lot about connecting, and again, Amazon may be blocking this, but I think connecting to my Amazon history is something that I haven't heard anybody talk about. We all have those days where you can scroll all the way up and see your very first order ever on Amazon, and I actually do this on the Apple App Store, too, where you go back and you see every app I've ever downloaded, and I think histories like that, they're very part of my life, but until an agent has access to those, it just doesn't understand all of that. And so I think there's so much more shopping history that if we could package up and give to our agents, could do even richer things for us. Look, I think bringing agents to social events is really interesting. You talked about the home, have it be there for a family dinner one day. Have it be there when we watch the Seahawks game on Sunday. Have it be there for a party or a bachelor party or who knows what. So I think, of course, there's all kinds of questions about permission and this Schrodinger's cat thing, which is, does the act of observation change the quality of the interactions? But I bet even a snapshot of you being yourself in that environment would provide so much more signal than a lifetime of hardware, I think. Yeah. I think that's right. So much of it is on the written word and they understand exactly what we've typed or texted. and tries to build a personality, but we're so different. One of the other things I think a lot about with agents is the things we often use agents for, or even search for, are things we don't know about, things that aren't fundamental to our lives. And you know, I was doing all this searching when the World Cup was on this summer, and all of a sudden, Chachaputi thought I was like a soccer fiend. I was like, no, no, no, I was into the World Cup, that was it, we're done, that was good. And you know, I go on these curiosity journeys all the time. My daughter's into an artist, and so I'm gonna go like, learn about the artist and listen to music to get to know it, but it's not my habitual artist or taste. And so, how can it know that when we're on these curiosity journeys, that it's not actually who we are or what our real preferences are? I still haven't figured out how to do it with the agent, so I actually go into the invisible mode a lot in order to not have things stored, which then isn't really helping either. So I'm curious of how we can actually get our agents to get to know us better, and I like this idea of our real world behaviors. Now, how do you capture that data with the real world behavior is gonna be the interesting piece. I don't know, I mean, there's gonna be a thousand capture devices, I mean, unless you're going to the, you know, the new set of bars are gonna emerge that exist in Faraday cages, which I definitely think is going to happen in San Francisco. Yep. I think people are going to increasingly assume that they're being recorded, you know, in all environments. I mean, I bought the new Apple Watch because they have talked about some features like Siri Recap, Live Rewind that are coming out. I don't think they're quite out yet, but I upgraded my watch in expectation for those because I think it'll be really interesting to have this thing that's at least kind of capturing tidbits across my day. Right, I mean, it's like Amazon B, it's a 24-7 ambient listener just wearing your wrist. Okay, maybe to wrap it up, so the first is we're gonna get back together, let's say, in three or four weeks. What do you think will have changed? Okay, I think we're gonna see a lot more innovation in the enterprise world. You know, we were chatting briefly before, it really seems like the whole world is focusing on consumer with not as much attention to enterprise, and I think that we're gonna start to see a little bit more shift in that direction. I mean, I think we're gonna start talking more about fun. I think we're gonna see agents that don't just feel like they're here to do something for us, but they're here to actually let us think out loud and help and get through our day a little bit more, making us laugh, doing little gossip in the background because I think that's what's gonna then make us build a relationship that'll make us trust it more to get all this other stuff done. So, I think we're gonna start seeing the next wave of users taking hold of these and seeing how they behave. And one reason I point to that is I gave my mom Instinct, and I'm just like watching her now play around with it. She's having so much fun. She's technophobe, hates technology, and is just having a blast. It's like a kid in a candy shop. And it's quite interesting to see how someone who struggled with the concept of chat GBT and how it works is just having such an easy time texting an AI assistant. A recent use case she did was our family dog passed away last year, and she wants one that looks just like that, just like our old dog. So, she sent it a picture. It was like, find me a breeder that has this exact type and negotiate for me. And so, I think it's gonna be really cool to see we know how tech lovers are using these assistants and what they're starting to do, but how is the everyday individual gonna start adopting these behaviors? And are they gonna be focusing on a lot more fun and just creativity tasks? Or what's gonna be their preference of surface? What's gonna be their preference of behavior? So, I agree with you both. I'm gonna take it one notch further than just fun, which is what are the aspects of ourselves that we don't either want to think about or reveal to perhaps even ourselves or to these agents? I think whoever is able to build that type of trust, almost like a dark agent, will actually have something really fascinating on their hands. Because in a sense, starting from productivity, starting from email, it's a bit of a constant reminder of the things that we haven't done, the tedious tasks. And I don't know that people want to spend as much time there. I think people want to aspire to be a whole different superhuman version of themselves. And whether they believe in their heart of hearts that it's going to happen or not, I don't know that we have an agent yet that really speaks to those aspirations or desires. It's like the Maslow hierarchy in need of agents. Perhaps, perhaps, perhaps. I mean, I think that's a really interesting idea that in a sense, when do these agents flip from getting something done to take it off our plate to giving us a secret superpower? Right. Like, you know, when it's, David, how did you get that cool shirt? And you're like, my agent ordered it. You know, and I'm like, oh, why don't I have an agent that orders me shirts? Or hey, I've noticed you're starting dressing better. It's like, yeah, I gave this agent access to my camera roll. I said, upgrade me, started ordering a few things. That's why I'm dressing better. Like, I think that when we get to the point where agents are giving us superpowers to feel better about ourselves, that's when people are gonna take notice and want that too, versus, you know, I think a lot of the examples that get thrown around, save some money, book a trip. Like, I like planning a trip. I don't necessarily want you to take care of all of that for me. Like, we need examples that are like, oh, wow, I should use that for myself. Right. I think that's gonna happen. It's like, can it help you become the human that you've always wanted to be, but have struggled achieving yourself? And you start to consider it your secret weapon in your pocket. I love it, guys. Should we wrap there? Yeah, it's great. Incredible, thank you. It's great. Awesome. Very cool. Thanks for listening to this episode of the A16Z podcast. If you liked this episode, be sure to like, comment, subscribe, leave us a rating or review, and share it with your friends and family. For more episodes, go to YouTube, Apple Podcasts, and Spotify. Follow us on X, A16Z, and subscribe to our sub stack at a16z.substack.com. Thanks again for listening, and I'll see you in the next episode. As a reminder, the content here is for informational purposes only. Should not be taken as legal business, tax, or investment advice, or be used to evaluate any investment or security, and is not directed at any investors or potential investors in any A16Z fund. Please note that A16Z and its affiliates may also maintain investments in the companies discussed in this podcast. For more details, including a link to our investments, please see a16z.com forward slash disclosures.