E55: AI Native Teams with Jacob Posel from HQ

Messaging with Jack was the worst thing that ever happened to you? No. Um, I like, I went out, I went out one night and I left my backpack at someone's house, a friend of mine called Jack. And then I woke up the next morning super hungover and I was like, shit, my backpack's in at my friend's house called Jack. So I text him and I was like, yo, like I need to pick up my backpack. And he wasn't answering me. So I like call him, he like wasn't answering. So I ended up like calling his girlfriend and then getting my backpack.. And then like 3 hours later, Jack sends me a video of a baby, like a little baby. And I was like, I was like, what the fuck? Like, why is he sending me a, like, why is my friend sending me a video of a baby? And then I like, I just didn't, I didn't understand at all. And Jack was like, I just had a kid, so I'm showing you my baby. And I was like, okay. Like I just said like, okay. Like I just didn't, I just didn't understand. And then like, oh, like I think like a month later I go and I actually go to message my friend Jack again. And I realized that I had messaged the wrong Jack and I had messaged this Jack about his like real baby and I just felt terrible. He had like just had a child and sent me a video and my reaction was just like, okay, like I guess. So yeah, I felt terrible. I remember this exchange. Yeah, I didn't, I didn't realize that you thought you were messaging a different Jack. That's funny. Yeah. Anyway, so I'm so embarrassed about that. Ah, don't be embarrassed. No harm, FL. Dude, Jacob, you're finally building SaaS for everyone. For everyone. I am. For everyone. Yeah. Yeah. Yeah. You don't have a small town problem, as they say. No. Um, HQ fortunately works for everyone, for every type of business, which is really nice. Yeah. Tell, tell us about it. Cause I, I, so I, I've learned not that much, but I didn't realize you were, uh, I thought this was just you building it. I didn't realize you had some partners in this as well, or I, I don't even know how that's structured, but anyway, I'd love to learn more about all of it. The story is actually pretty cool about how I, um, how I got, how I like joined and got involved with the team. So I met Corey, Corey Epstein about a year ago, actually in, I was in Rishabh's office and we had a call and I remember it like vividly because it was like one of my first calls when I got to San Francisco. I did it out in the hallway. Anyway, he like showed me this product that he was building called Indigo and it was like very ahead of its time. But it was like at a point where AI hadn't like quite caught up to what they wanted to build. It was like this personal AI that lives on top of your computer and just allows and just like helps you do work embedded directly into it. It was like, it was very cool design, very well. And I was pretty impressed, but like, They just, AI hadn't caught up yet. But like Corey was, Corey's like a visionary. He's like, he's exceptionally smart and good at understanding where AI goes, where AI is going. And I could like tell that pretty early on. Anyway, him and I like went through Twitter, went back and forth over the years. Eventually he reached out to me and he was like, you should try our new product or try this new tool that we've built internally. It was just started out as this internal tool called HQ, which Corey runs LiveRecover, Voyage, like, um, and Indigo. So he has to operate at a super high level of efficiency. And so it was this tool that allowed him to use AI, use AI more effectively. And anyway, he, he was, he onboarded me onto it and I was like, holy shit, this is really cool. Like I thought I was a good AI user, but it made me so much more effective as like a personal tool, right? It was like for me to be better at AI and I was highly incentivized to help make this product better because it was helping me in my work, in my consulting work. I was working with multiple companies building things for them. And so I was incentivized to help with the product. So I got pretty close to the team, started working with them more directly, was consulting at the same time. And the number one issue I was seeing across all these businesses was it's just really hard to get team adoption of AI. Like all these products are inherently designed to work, um, single use. Like there is some organization sharing, but it's not robust and it's not a focus. So everyone was just working in silos. Everyone was quite, um, everyone was just repeating work. Um, people weren't kind of moving in the same direction, working in the same place. And HQ at the same time was building up all the context of my business. Like inherently the product becomes better as you use it more because it builds up context on how you work and how you operate. So I pitched the team, I was like, you guys should just build a version of HQ for multiple people, for like a team version of HQ where you can, where people can use it, it makes them better and more effective, but you share these capabilities and solves a real problem. For the businesses and they, um, they basically agreed and asked if I wanted to ask if I wanted to join the company. So I was like, yeah, I was like, I'm already, I already want this to be better and I'm already going to be incentivized to make this better. So I was like, yeah. So at the same time, Ramp launched an inter— they like wrote an article about this internal product they built called Glass. And that was really cool for— that was really great for us because it validated it and the— and like the market, meaning like Twitter really liked it. And so I tweeted like, we built RAM Glass for everyone. And, um, I just got insane demand. And so I like pushed the team. I was like, let's fucking, like, we have to, we have to get this to market. We have to push this out. Um, and we ended up building something really cool. And then we launched it, um, about 2 days ago. And yeah, does that, I don't even remember what the original question was, but I just yapped about it for like 5 minutes. No, that was it. Just wanted to get the context behind everything. Yeah. So wait, what does it do? I mean, I get the general idea that it, you know, doesn't, uh, it brings teams together to work more collaboratively with AI, but like, what does it actually do? It's at its core a cloud-based file system that you interact with, that you use when you use a tool like Claude Code or Codex or OpenCode or any kind of agent harness. It syncs files to the cloud and helps you resolve conflicts. So it abstracts away Git-level complexity. GitHub-level complexity for knowledge users. And then it has additional features built on top to actually work for bigger businesses. So it has access control built into the file system so you can control who sees what. It has secrets sharing because most skills interact with external services. So it has secure encrypted secret sharing. And then it also has app deployment. So easily deploy applications from directly inside of HQ. It's really a super simple level, super simple level, super simple thing. Like I have a go-to-market team. They all have built their own separate janky tools for how they automate outbound, like a super common problem. This allows them to just create one all inside of HQ. They all interact with the same thing. Not only do we have, do we know what they're interacting with, but it gets better as they use it because they're all compounding on top of the same skills and knowledge system. So Jacob, I mean, you and I know each other, but people listening may not know that you and I know each other. But when I first heard that you were doing this, My thinking was you're finally working on something that you should be working on. Yeah. Which I won't go into all of the reasons for that, but there was a series of other things that you worked on, which I did not think were a good use of your time, which I have shared with you separately. Okay. Do you feel that way? Yeah, I do. I think my like calling, or at least where I get excited, is getting people excited and getting people using AI. And like, I've always had this tension of I felt like AI was super powerful and I was just, and I always had such great experiences with it and Not everyone has that. Not everyone gets up to speed. And I feel like this really does, like, it takes the 0 to 1, but it also takes the 1 to 10 kind of thing. So I feel excited about it. And I've had this like really great experience where, because I've been just like faffing around with so many different types of businesses and business models over the last 2 years, like SaaS and agency and like consumer and brands and stuff like that, um, I feel like when I speak to people, I just have a really good sense of where the value is for them in building with AI. Like I can directly tell them, this is where I think HQ is valuable for you. And that's been, that's been quite nice. So it does feel like a good product for me to work on and I'm super excited about it. And also finally, I'm working with people with like really, really smart, cracked people, which is like really exciting to me because like just working with it, like I'm like a, like not a good engineer. I'm like not a good engineer. Like maybe I have good product insight, but like I'm not a great engineer. Now I'm working with great engineers and that's just so, it gives me so much energy because I don't have to focus on the things that I'm not great at. Like I can just focus on what I'm great at. And they, and everyone else in the team can focus on what they're great at. So that's really nice. Yeah. So there's this, there's this thing that I wonder that we debate forever on this pod about like SMB, mid-market enterprise. Okay. We like, we have this conversation. I don't know. Every, is it fair to say every episode at this point? It's gotta be at least once a month, right? Once or twice a month. So roughly all the time. Okay. So, so, so I wonder as you're building it, where you think it will have the most value. Over time, like in which segment of the market and, and, uh, you, you don't need to share the segments that you're currently working with, but, uh, you can share what segments you're currently working with, what segments are coming inbound. And I just, I just, uh, or if you don't want to, that's also fine, but I just wonder, ultimately what I'm curious about is where should the end state be? And given where the end state should be, how does that influence how you think about the go-to-market today? And my, and my push is basically. When you post something on X, you are only going to get one variety of those customers. And I am wondering if that's who you want. So in terms of who's come in, in terms of demo requests and in the pipeline, we have very big businesses, like thousands of employees. And we also have like two people startups. So there's been a big— From X? Yeah. I'll share them with you afterwards, but yeah, we've had both. Yeah, yeah, no, no, don't share it here. I'm— so, so, so you think that posting on X can bring in enterprises? I mean, it did for me. Okay. But I'm just built different. I love that. No, skill issue. Do it. Um, I mean, I, I guess I posted on LinkedIn and at the time we didn't have UTM tracking set up, so I don't know exactly where they come in, but I think it's from Twitter. But like, I know from experience, like I have in the past gotten enterprise inbound from Twitter from a business that like we both worked with. I think you probably know who I'm talking about. I know, I know, I know. But two others or three or four others. To, from, I think, like, large— yesterday, really large businesses came in. Um, and I haven't even looked through like the rest, but yeah, we've definitely had enterprises. And then like, part of it may be that your content is interesting to those people, like the things that you talk about and post about, uh, in terms of like AI usage and stuff, like the big companies obviously have these problems too and are trying to figure out how to get adoption. So like, obviously those people are there on those platforms. And if you are speaking to the things that they care about, you might be more likely to reach them. I think it'll be interesting for you, like the stakeholders that come in, um, haven't always been executives though. Like some of them are, I'm a head of AI or I'm leading AI enablement at this large, um, business. And so I'm interested in learning more. And I feel like the persona of that person at these businesses can now tend towards younger, um, and tend towards more online. So, um, I think that makes sense. I think, um, those people coming in through Twitter definitely makes sense. Who do you want to serve with the product? So, so now, so now actually you have the prototypical problem. So, okay. So actually this makes the question even harder, right? So like, 'cause a lot of businesses, they can only attract a certain type of customer. And so then they just end up saying, okay, if I attract this customer, I serve them. But now you're saying I have everybody from thousands, tens of thousands of people to 2 people. Mm-hmm. And so now prioritization of where you want to serve first becomes even more important. And so how do you think, like, forget you, Jacob, the person. Like the job is what is HQ's place in the world, right? Like your job at HQ is to make sure HQ gets its place in the world. So where, where does, where is HQ's place in the world? I don't know the answer. I don't know it exactly. And that's the reason why we launched how we did kind of like go through demo because I want, I want to have high touch in rolling these out and see how they land. Uh, my intuition is that it works best for larger enterprises, like a bigger business where a lean tech team could build HQ, like a lean startup could build HQ or their own version of HQ, like obviously because we did it, you know, like they could build it, um, a super lean or even like a relatively lean with a technical person could build it. But, and they're already pretty strong AI users. Like they could figure out the issues that they could figure out ways to solve the issues, or at least they, at this stage, if they're, if they're early, if they're young, if they're small and lean, then they're going to be AI first and AI forward. I think the best company that would benefit the most from HQ is you have a top-down initiative. We want to implement AI inside of our businesses, and I'm just really struggling to get everyone using it. Like, I'm struggling to get the person who has been part of the company for 10 years, who fills some position that I don't know, that, like, I'm really struggling to get this 40-year-old member of my, like, accounting team or marketing team to use AI. And I don't want to put them in a terminal. I don't want them to, I don't want them to, they don't need to understand what a CLI means, but I really want them, but I see other people doing things that are so super powerful and I want that for everyone else. And I think that happens in a bigger business because kind of by nature, average skill level will tend towards the middle or like decrease. So that's, that's where my intuition goes. Can I ask a question? Yeah. I can't remember if this was in the video. I feel like it was not. Does the, does the tool also give you reporting in terms of how people are using it? Yeah, I was about to ask this. It does. And this is huge. It does. And this is, I think this is really big, like being able to see what people are working on, how much they're working on it, like the token usage, stuff like that. I think that's going to be really exciting for bigger businesses. I don't even think you have to be that big of a business to have this problem. I mean, basically any business, I think, where people have been doing something for an extended period of time in a specific way, when you have more than 10 people, I think this is actually a problem. I do it because you have multiple levels between the person who's trying to execute and implement AI and the person who's doing the work. You have this problem. And so I think, I think this starts at, I don't know, 15, 20, 30 people, um, as being a problem. So yeah, I just don't, I just don't think those people are good customers, Jeremiah. Like that's you and me basically, right? You're describing you and me and I just, I'm like thinking to myself, like, why would, like, like it just feels like bad business for Jacob to spend his time talking to you and me versus spending his time talking to workday. Yeah. Well, let's talk about the economics of it though, because I think that's actually interesting, right? Like, I mean, Jacob, I just had, we had a guest on, this isn't out yet, but a company called Adapt. So we're using something that's similar, at least in the way you're describing it at our company. And it's a, it's a basically like the, the AI brain of our business. And so it's a really interesting concept. And I think it's a, it's a conceptually very similar to what you're building, but I'm finding us spending a massive amount of money on it. And basically their model, we, uh, we paid them and they basically run everything through a bunch of different models depending on like what the task is. It basically picks the right model. So it's not, it's not just Claude in this context. Yeah. So in that context though, I don't know how their economics work, but I know that basically there's like effectively like we're paying we're paying a percentage of the token usage, if that works out. And so in that case, that's actually like quite lucrative, I would think, for them. But I'm not sure how your model is. So I'm curious what that looks like. Well, because we built on top of Claude Code and Codex and whatever you're using, we don't charge per token. We don't charge for tokens. So like a customer, I think costs a dollar per month for us. So our goal, um, and this is like something I think is worth discussing with you guys. I just feel, I feel like we live in a time right now where there's like almost an arbitrage opportunity based on distribution to get insane valuations, um, or to get insane opportunities based on, um, distribution. And I just want to land HQ in as many businesses as possible. Uh, I, we are offering to charge for service. So because we have this, we have this like lens into your business and we, it's like super powerful. Like imagine I was working with you, Jeremiah, and you wanted some skill across your business for some team to use. Like if I'm part of your HQ, I can just push it in and everyone else on your team can use it and I can teach them how to use it or I can create a handoff skill or tutorial for them to use. Right. So like I have a lot of access. So, um, we are offering service for an implementation, which there's been a fair amount of demand for, you know, like set it up for us. And teach and build some stuff for us. But in terms of the actual like pricing, I just want to get it in as many businesses as I can. I want as many people running their businesses on top of HQ as possible. And obviously like there's insane ways to monetize on top of that. If like, if you, if I'm the primary interface for the ways that people are interacting with AI, like, oh, I'm the primary workspace or I'm one level away from Claude or OpenAI and I'm the agnostic open source provider to it, then like I can sell, I can sell or provide a lot of value on top of that. What's wild too is that, so one of the things I think about a lot with this with any business built on top of token usage effectively is that there's a— in my business, I see a ton of waste in the token usage because of the fact that like people are doing the same things multiple times. There's like, there's probably cheaper ways you could do the same thing. There's all the, there's different models, all this kind of stuff you could like. I am very certain that the things that we're doing today could be done with a third the number of tokens if we actually knew what we were doing with that. And so if you sit there and you have all that visibility, like I actually think that will be a major business in the future is how do we people coming in and saying, I'm going to help you optimize your token usage, whatever that looks like. And I don't know if that's kind of what you're thinking about, but I just think there's so much opportunity in that context of, if you are in all these businesses and somebody's spending $10,000 a month on token usage and you can get them to cut that in half, they're going to, while delivering the same amount of value, there's an opportunity for you in that context. Yeah. And incentives are aligned here. Like if you are using a provider like Adapter, whatever, where they charge per token. Like they may say that they're going to optimize your tokens. They may say they're routing you through a lot like small models, but incentives inherently are not aligned. Like they want you to do more and you want to do, you want to use less, but our incentives are the same. Like we use HQ, we use it inside of our Claude Code and Codex subscriptions. We rolled out a skill called Caveman Bro, which we found on X that reduces the number of tokens because we want to save money because we're all blowing through our subscriptions because we're also like because we're all so efficient now, you know? So it's just different. So building on top of those things and working in the same way that you guys do, it just changes the economics significantly. Yeah, I like this distribution play. I think it's great. I think it's really strong. One of the things I wonder, and this is, I guess, more of a product-based question, with teams in AI no longer being solo, you know, we talked about one of the problems of solving that everyone is working on the same thing over and over again, that the work is getting redone. Let's say by the sales reps doing their outreach thing. Your sales reps are all building separate outreach methods that could consolidate into one skill. But then I guess, and I just want to battle test this one example, you lose the personalization. Like, I'm sure every organization has seen their sales reps say, I actually don't want to do what the other sales reps are doing. I don't believe in what they're doing. And the small amount of variation, like, that's getting standardized, uh, in, like, let's say, a skill would actually frustrate them, right? Um, and you, you decrease your exposure to, um, people who do incredible net new work, right? Like, um, One of the things that I've done pretty recently that I think is pretty nifty is I managed to get my Mac Mini to handwrite emails, like total perfect human simulation using Gmail for my cold outbound as opposed to something like Instantly, right? So that then it's just sending actual emails as if it was a real user. I like that. I think it's, I don't know if it actually does that much versus sending email via the API. But I like it. Like there's something there that feels good to me as if it might, have some benefit. If I had to, I guess, comply with the way that the rest of the team were doing outbound, I might like find that frustrating. And then the team as a whole might lose the ability to do things, you know, individually, right? Differently, depending on the preferences of the user. Yeah, I just, yeah, would love to hear your thoughts on that. Like, is there ability to fork off and say, actually, yeah, I want to use these other people's skills. I want to do my own thing here and build something net new, completely different. And not be a team player, actually be a solo builder? Uh, there is. So like HQ has a personal, there's like a personal version of it and a company version of it. And not every, you're not locked into the company cloud. Like I don't have to push everything I work to the cloud. I don't have to pull everything I work from the cloud. I can have personal things that I don't share with the rest of the team. So yes, like the capability is 100% there and my HQ could become personalized and I could say, I want to use this version of the skill, but I want to put my own twist on it. But I would also challenge the premise of the question. Like if one SDR is doing something better, than the rest of the team, then surely as like a head of sales, I should just, that should become the default way that the team works. And like, yes, maybe like, but yes, I can also roll personalization into that process or like a little bit of, a little bit of flexibility on top of that. But like, I just think that if like, if I have a team skill and one SDR is doing something that's better, then it should like, that should just become the team skill. So like in that example, yeah, 100%. It's chicken and egg, right? If everyone's standardized and you've gone in a direction that is becoming concretized, like how do you fork out? And so I could imagine in a sales team, you know, small sales teams, just 5 people, forget big sales teams, people just getting constantly frustrated at what other people have built. And they're like, no, I want to do things my own way, right? The way that Dan is doing it is really stupid. Like, you know, yeah, the metrics look good on a one week, but what will they look like on a month, right? When they're statistically significant, they'll have a bunch of, you know, explanations for why. What Dan is doing and is the prevailing skill right now is not what they want to do. Right. And if they're commission incentivized, they should have, I guess, the freedom to do that, right? To fork off. I mean, I think this is like a short-term problem. Like, let's just take the sales example. In, in, in like an ideal world, you wouldn't have random humans doing what they think is best in some competing way. Like, you just will have criteria for what is correct. You'll build skills around that or you'll build capabilities around that. Maybe you have one person operating at far more efficiency and they can basically eat the rest of the team. Like there's no bottleneck in output with AI. Like if I have 5 sales reps and I have one person that's consistently outperforming, then they should do all the outbound or I could build agents on top of it. And that's a product direction that we want to go into HQ in the future is like, you've built this company workspace, let's now deploy agents on top of that knowledge base and skill base. And I think like that would be an ideal direction. But like to answer the main core of the question, it is It is customizable, it is personalizable, and I think the way that you manage what is at a company level versus a personal level is like a matter of preference of the executive or the leader or the CEO. Cool. It sounds great. It sounds like a great product. For time, who do you think is the best buyer of this product inside the company? So you, you, you actually, or maybe this is a question we should ask you in 6 months, but I, I'm just curious, like, what is the title of the person who you think would buy this? Two, in my experience so far, it's the CEO. It's like, it's whoever has the most incentive to enable AI in their business. So it's the CEO has been a lot of the, um, a lot of the inbound because like they're putting all the pressure down onto, into, under the AI rollout. Or for larger businesses, there's this like head of AI enablement role, which has recently popped up that I've seen coming up more and more. Uh, and they are responsible for AI rollouts. And so they are they're the one buying it. But, um, the way that we've been landing it is like we'll land it with your strongest power user and then distribute it outwards from there. So honestly, that's like usually the CEO. And like, I'm curious why you think that is the case, that in so many businesses the CEO— I don't think it's— you think— is it the CEO or is it the founder? Founder. Sorry. Why the founder? I guess now I've just answered by reframing it. That answer. That makes sense. That makes sense. I mean, Okay. You don't think it's CEO as well? Like, I think the, uh, you don't think that's true of CEOs that like take over businesses? No, I think depends. I think there's a small minority of CEOs who have the mindset that it takes to do, to do this. The average, like the average CEO is an optimizer, not a transformer. Um, and the average founder is a transformer and not an optimizer. Actually. And so you find like founder CEOs have to augment themselves with optimizers and, um, like vice versa, right? Like that's a, that's a very common, um, or like systems builders. And I think that, I think that the reason the founder is the most active user of AI is because they're happy to go back to a team of 1 or 2. Yeah. And that's what they're doing with AI. Like, you know, like, uh, I've always been pretty impressed with, uh, Cory at Jones Road. But recently, much more so, much more. There's Cody. Cody. Cody. Sorry. It's because I said Cory. Yeah, you said Cory because of Cory Epstein and now I'm like getting confused, but sorry, Cody at Jones Road. Okay, so recently I've been very impressed with Cody, much more so than I've been in the past because I think he has like decided to go founder mode on a bunch of things. And I'm like seeing him do things. I'm like, dude, other people are not going to do this. They're just not. And it's like so evident because he'll, the way it's evident is he'll post something and the responses from people is, can you share that? And that is the tell. The tell is they want someone else to do the work and hand it to them. Have you ever seen Cody say, can you share that? Cody is just doing it himself. And that is, that's the, that's the tell. The tell is like, no, I don't see in the comments someone asking like, dude, what inspired, like, what was the first step you took that got you on the journey to being able to do these things on your own? If someone asked that question, that would be like, okay, this is actually someone who is likely to end up in the same spot as Cody. But that's not what people are asking. And so I just, I just think that like, yeah, I don't think, I don't think the average CEO is going to do it. I think the total number of force of will that's required is just monumental. The interesting thing there as well is that it's not a technical barrier. There's no technical barrier to entry. There was in the past, but this is the one thing where there isn't a technical barrier to entry and you could just get stuck in. I've seen Cody's tweets as well. And yeah, you see people in the comments saying, hey, can you basically hand this to me? And he is looking to hand it to someone. He wants to hand it off to someone and that person's going to have a job role. And I actually think that job roles are very similar to the person that's in the comments saying, can you hand it to me? Because they're given a title of a thing that they do, and then they're given a system that they have to follow and repeat that's set by someone, either by standards of the industry or by the CEO or the founder. And then they've got to go and do that thing. And what you might find with employees in certain organizations is an unwillingness to adapt, to update their methods. So let's take, I guess, the typical example of someone who might be at Jones Road Beauty with Cody, someone who sends email in Klaviyo, right? That person has a system that they've been going through for weeks and weeks and months and months and years and years. And it could actually be really hard to make them AI native. Same thing with, I imagine they weren't building all their landing pages with the CEO in the past, right? But now they are, right? So who was doing it before? They have their processes. Maybe they were using Raplo. Maybe they were natively in Shopify. But now he's creating this new role. He's shaping out what it's going to look like. And someone's going to go in on train tracks and do the thing that Cody trailblazed, right? I think in a lot of ways, this might be the reason the founder is the person that does it, is because they have that high agency to just go and get the thing done, right? Even without a technical barrier to entry, they're just prompting, right? Jack, that— can I push back? I think that's a massive mistake. If what ends up happening is he hires someone to go do that thing, I don't know that that's what he's doing, but if that's what he ends up doing, the reason it's a mistake is that that thing that he built today is going to be different in 3 months. And this is actually what separates Cody from everybody in the comments. What separates Cody from everybody in the comments is that what he was doing 3 months ago is different than what he is doing today. Right. And actually he chose to continuously update what he's doing. And that's exactly why you cannot be handed something because this is, we are not settled. Like AI is not a settled technology. Like cloud is a settled technology. AI is not a settled technology. So the, the, the, the, you have to be, you have to like not hand someone something. You have to say like, here's how you have to think. Are you gonna get on the treadmill and keep running? You can't say like, hey, this thing is done and hand it to someone. And that shift in mentality is exactly the problem. And I think that's why, Jacob, to your point on like the CEO versus the founder of like who is the adopter of HQ, by the way, HQ also, nothing static about the output. You build a skill, you push it back up to the parent organization, how, or a parent organization. It can, you can choose how many hierarchies you want in the parent and the, and the, and the permissions, but the point is actually that it is by definition evolving and it is a learning system. And most people want to be told, here's how you use the tool, now operate the tool. And that's literally exactly not true. And that's why the person who is going to be most likely to buy is a founder and/or is someone in the organization who has that mentality, but they're super hard to identify outside in. Super hard. Yeah. Does that mean in your mind that most people, while we, I agree with you that like this technology is not finished, like it probably won't be for years, right? In terms of like what the final form is. So in your mind, does that mean that most people are going to be perpetually behind until we get to some sort of steady state? If that ever happens, I don't even know what that looks like years from now. But, um, cause like what I'm thinking about here is somebody could go and enable HQ., but unless everybody in the org is actively working on improving it, or you don't need everybody, but there has to be a driver of that. Otherwise it's going to stagnate as well. Correct? The single— you, you, if I were a salesperson at HQ, you know how I would sell it? Tell me. Madam or Mr. Founder, you have a distribution of talent on your team today in terms of how AI forward and AI native they are, and you're sitting to yourselves, and your only tool is fire the people who are not AI native and promote the people who are. That's the, that's the only like choice you have today. What if I gave you a tool that took your top users and automatically propagated what the skills are to the rest of the organization? You can still do that sort, but at least now you have a second lever to take the distribution of talent on your team and through a tool force the issue on at least some larger portion of your organization. That's how I would sell it. And I would sell it as an organizational issue. Um, and not, not a product issue. Our tagline is one person's breakthrough becomes everyone's baseline, which is similar to what you're saying, right? Like you, you distribute down the capability. Yeah. My suggestion is that you directly provoke the question that they're already interacting with, which is like, what portion of my team do I have to fire? Because it's not a mistake that Coinbase and like these massively profitable companies, why do you think they're firing people? It's not to save money. I mean, like, no one in their right mind thinks it's to save money, right? I mean, can we— I assume we agree that it's not to save money, right? Despite the impact on the stock price, I think that Brian Armstrong articulated it really well, exactly why they're doing it. Actually, I think Dorsey as well did the same thing. They're describing it's not just to save money, right? Why is it? Tell me, I missed the Coinbase— It's a fundamental restructuring. They're like fundamentally restructuring everyone's roles, basically. To have a different expectation from people. And Coinbase has been doing it for a long time. They're the reason I first started using Cursor. They've been at it since like early 2025, but they're cutting now and it looks like they're changing the expectation for every single employee at the business. That might be a little bit too much of an absolute, but like for a lot of people, right? Yeah. I just can't stand crypto. I stay, I try to stay away from all the Coinbase. Okay, fine. I'm kidding. I'm kidding. I'm kidding. I'm kidding. Cloudflare. I'm kidding. Okay, whatever. Cloudflare. Okay. Cloudflare yesterday did a 20% layoff. You know what I mean? Why do you think Cloudflare did a 20% layoff? Dude, these people are printing money right now. They're printing money. So like the premise is if they could, if they could have N+1 AI-enabled employees and they would choose to keep them and this allows them to do that, but they just can't, they can't get that cohort to work. This is the craziest labor market in history. Okay. You can't keep the people who are not AI-native. There are for sure not enough AI-native people in existence. For you to actually fill the roles in your company. So you have more job postings than you've ever had before because everyone is trying to find that subset of people who belong to that cohort. But the total number of actually employed in tech people is going down. It's not because, and none of it is for financial reasons, which is the crazy part. They just don't like, what, why are they firing them? You can't have those people in the company. Why can't you have them? You can't have them. I don't, I want to push back a little bit on that, that it's not at all for financial reasons. I do think one of the things we're seeing is that the software market has seen really big loss in stock price. And so you've seen, what is it, the top 25 software companies are down 50% over the last 6 months. I do know that has pressure and ultimately the way to expand your price is better revenues and better margin and basically giving confidence that you're going to be able to succeed in the future. And it's very hard for those businesses to show that right now. But I think part of what we're seeing is that you have, um, things like debt covenants and stuff that start to come into play when stock prices drop to certain levels and you've got profit margin concerns and all these kinds of things. So I think there's a little bit there that maybe goes beyond the just trying to save money where we're seeing companies making some of these decisions based on the future outlook of the industry, not necessarily just the whether these people are AI native or not. So I, I don't know if that, uh, made sense, but anyway, that's, I think there's, there's some stuff there with the, the banking system ultimately that plays into this as well, not just the AI versus non-AI native people. I think that's true if you have already signed yourselves up to believing that you are not going to become an AI native company. So under that context, that's true. But in the contexts of companies where that's not true, Cloudflare being one of them, because what's a big part of what Cloudflare is doing right now, they're making, they're helping companies make decisions on whether or not a particular bot is allowed to crawl the page. They're like, I mean, like, again, we don't talk about them like an AI company, but they're for sure an AI beneficiary because the total number of pages that you have to render and how you decide whether or not you allow an interaction on your pages, all of it is being arbitrated by Cloudflare, right? This like random company that controls like what, 90% of the internet basically. And I think that, and it's a massively profitable company. And so, yeah, I don't know, man. I think it's like, I think the total number of people who are going to be effective is today not large. And so you need to figure out a way to get that total number of people who are going to be effective to become much larger. And there's, I think there's way more pressure to doing that than there is financial pressures. Uh, when, when you are, when you are cash flowing, if you're not cash flowing, it's a totally different issue. Yeah. But it, in this context, even just, just work on the like Cloudflare. So I just looked it up. They've got, I don't know enough here, but I'm just, as an example, they've got $3.5 billion in debt. Sometimes cash flow with these companies isn't actually the thing that matters. It's your debt covenants and what the stipulations are that have been put in place. And so when your stock price is down and the industry's down and you've got like margin pressure and the bank can come to you and say, okay, I have a friend who worked at a large public company that went bankrupt recently. It wasn't that they couldn't pay their loans. It was that they tripped their debt covenants and had to then pay back $100 million that they didn't have at that moment. And so the bank said, You've got 60 days and like, well, we don't have, uh, they went and tried to get $100 million and they couldn't get it. And then you basically just do an asset sale and that's it. So there's some of that kind of stuff. I think that's happening too. I don't think it's all like for Cloudflare, their debt, their stock is down. It could also be that, uh, this isn't a good example because I don't know enough about Cloudflare, but there's businesses like Cloudflare that are doing layoffs in part just because of the fact that a bank may be coming to them saying, hey, you're, you're at risk here. You need to be careful. And I think there's a lot of that kind of stuff that's happening right now too. Yeah, I think to give like a really concrete example of what Rashad was saying about those people cannot exist at the business. If you take a look at just like AI solving customer service, right? That decreases risk at companies like the ones we were just talking about, the ones that we were just talking about. Uh, customer service sometimes sell customer information. That's like a real risk that you completely mitigate by having AI solve it. There's also human error, right? There's some rate of human error, um, from these people. And then there's the economic value of the service actually forcing compromised, right? Bad English in a lot of cases. No one has a good experience with human customer service for the most part, right? For the most part, people are frustrated with their experience with customer service. I don't get frustrated with AI customer service. It just solves the problem. I'm like, okay, cool, bye. I don't feel like I have to be friendly. It's like, okay, how do I do this? Here's it, here it is. I took you to the page. I'm like, oh wow, that's incredible. Or like I was in customer service for ElevenLabs the other day. I was actually speaking to it and it was taking me to different pages. I was like, this is incredible, right? This is what customer service should be. So I don't think that Humans can exist in that future, right? Personally, right? And it seems like these businesses are choosing for humans not to exist in that same capacity that they used to. But I'd love to hear what you guys think about this, and especially you, Jacob, you have so much surface area on these different types of businesses. But everyone knows the romantic view of product-like growth, right? No marketing line item, adoption through frictionless problem solving. People get into the app, they love it, they use it, it solves a problem for them, like Claude. Right. But we're still seeing AI implementation experts headcount for the purpose of bringing the team along. Because even with frictionless access to software that solves problems with like an aggressive velocity, right, at this unprecedented ease, some people still don't adopt, right? And they stay in their old ways. So you need people to come in and basically usher them. So yeah, I'm interested from you guys that are leading teams. And I know that Rashab, Jeremiah, you guys are actually confronting this problem right now in your business. How do we bring everyone along?, right? And make them update their ways. What is the role of, you know, software in this? Is there gamification? Is there a way that we can, uh, design the tools to help people come on board? Like what you're doing, Jacob, with HQ. Uh, what's the role of AI natives as headcount coming into the business and bringing people along? And then I guess the last one would be, what's the role of the expectation from the founder and the CEO to say, hey, actually you've got to do this. This is the new expectation. Just like we expected you to use Google Docs. And no one was particularly passionate about Google Docs, but you use it because it's part of your job and that's what you do. The founder or the CEO or the team lead or whatever it happens to be now has the expectation that you're going to be AI native. Yeah, that was kind of like a really big loaded question. Yeah. Founder, software, designing the software in a way that's frictionless and brings everyone along, like maybe gamification or making it fun. And then, um, and then new headcount, right? Uh, that's part of what we're doing with my business. Uh, it is free. We're trying to get as much adoption as possible. But on the service side, we're helping businesses solve real problems in their business with AI, right? That otherwise required people who had opinions and didn't want to update their process, right? What do you guys think? I think the coaching piece is so important. I think, like Jacob, you're talking about how you have the service aspect. I think that's so critical right now. Like Rishabh, you're talking about the, like constantly being behind the things that Cody's doing today are not things he was doing 3 months ago. I mean, that's true for me as well. And that's, it's just constantly changing. And I think the, the other thing that's interesting here, Jacob, the amount of demand you had so quickly with this launch, I think is just indicative of the fact that very few people actually know what to do right now. And, um, I, I'd actually be curious for your take on that too, but it just seems like there's this, we see a few examples of people doing things really well, but then the gap between that and actually figuring out how to make that work for yourself, for your teams, for your company, whatever, seems to be very large right now. Um, and Jack, I just, I, to directly answer your question. Like, I don't think gamifying software fixes that. I think it's really like coaching people ultimately. Like, yeah, AI is a new tool. Um, not everyone has time to spend their entire life on Twitter and using AI like I do, you know, like there's just real people with real jobs doing their real job and they actually don't have time to like get really good at using a tool like Claude Code. And it's overwhelming. Like the product experience is overwhelming. You know, like you do have, I think there's a reason why it's quite barbeldust because you have this chat box. There's like Magic Chatbox, which is really easy to, it gives you a lot of flexibility. It's really easy to use. But also it's like, what do I do with a Magic Chatbox? You know, like I put myself in the mind of some random employee and they're like, what do I do with this thing? You know, they don't understand the, and it is a technical tool. Like it just is a tech, like if I'm going to hook it up to my CRM, there's a lot of things that maybe I understand or like you guys understand about how that mechanically works because we work in software businesses, but like it is inherently technical. So like it's not, it's not necessarily easy. It's not necessarily intuitive when you get into it. And so I think it's fair to assume that it requires coaching. Like, and you can't compare it to the Google suite of tools. Like if I open a Google Doc, it's like very obvious what I do, right? It's a page for me to write on. If I open a Gmail, it's like very obvious I'm sending an email. But when I open like Pandora's box and I can literally do anything with this thing, it's like, well, what does that mean? So I do, I think it requires coaching. Um, but I do think once you've, once you've kind of shifted your mindset and once you get an intuition for it, then you do become really good at it. And that's why implementers are also valuable because, um, they are able to take their expertise and what they've learned and they're able to do that for people who don't know and able to teach them. And then there's also a mismatch of expectations. And this is something that I've shifted my mindset on recently, which is I actually don't think everything needs to be done with an autonomous agent. Like a fully autonomous agent that does everything end to end for you. And that's what people sell on Twitter and LinkedIn because it gets the views and it gets the, it gets the retweets. Like if I'm optimizing for retweets, then yeah, I'm going to say I built a fully autonomous agent that automated my entire business for me. But if I'm a realistic person, it's, I'm most effective when I'm inside of a tool and I'm driving it and it's super, um, and it's a very efficient assistant for me and it goes off and it does work sometimes, but I'm driving it forward. And so you have a mismatch of expectations. A CEO goes in, they say to their employee, I want you to automate my process of creating ads for me. And they just can't do that because of what I said before. And it's probably not the right product to build. And ultimately they should be building something a little bit more assistant-based. And so they should be working inside of a tool. They shouldn't be building, and they try and build agents. So I think it's a combination of a lot of things, but, um, but ultimately it goes to like why I built this, which is it puts people in the right place, which is they should be in Claude Code. They shouldn't be magically summoning open-core agents to do all their work for them. It teaches them how to do it along the way. It's designed in a way to abstract away the complexity. And I understand that not everyone's going to be an expert straight away. So I try and teach them and train them along the way. And I think the future of it, sorry, I know this is like a really long answer. The future of it, I think is, I do think it's possible for AI to teach. Like we do have a tutorial command. We do have a setup command. I think AI is a good teacher and that can help self-enable and self-heal the product.. And so that is a future product direction when it goes to product-led growth, when it goes to kind of self-enablement and setup. Uh, but in the meantime, I think humans can fill that gap. Dude, what a great note to end on. We should not cut you off. I didn't hear your thoughts on— I gotta go. On what? On what? Okay, I'll do it. I'll do it really fast. I'll do it really fast, but I really do gotta go. Yeah. On thoughts on what? On the expectation from the founder, uh, the role of the founder, the software, and Um, I should add in the individual because ultimately at the end of the day, the individual actually needs to want to use these tools. But yeah, the role of these different pieces in driving the business to become AI native, bringing those people who aren't using AI yet into using AI in every part of their role. Yeah, I, I have a very simplified view of this, which is this technology is evolving at a rate that no other technology in the past has ever evolved at. There's a small subset of people who want to move close to that rate or at that rate., and it is their job to get everyone else to move as quickly as we are able to make them. And that's it. And I think that if we do that, like, everybody wins. Uh, and if we don't do that, we will experience pain in all sorts of ways that I think are unpleasant, from capitalistic ones to societal ones to like corporate ones, whatever. Um, but that's the TL;DR. And, and Yeah, I think what Jacob described is like, hopefully AI is a way that we enable that, but regardless of whether or not it is, there's some set of people who are moving at the pace and you gotta, that's the job. Um, yeah, we're going to run. Yeah. Yeah. And you got sales calls until 8 PM. So I'll speak to you guys soon. Thank you guys. See ya.