Pain is a Moat: How Australia’s Top AI Founders Are Building Global Giants

I just really like building things. As soon as I started uni, I spent majority of my time not going to uni, but building things with my co-founder. And that was just like enjoyable. And it's a little bit of that attic, like I kind of get lost in time. When we started, it probably took 40 to 50 hours of human time per patient per year. Today we do it in somewhere from three to five. I think by the end of the year we'll have that down to a few minutes. The A1B66Moz of 12 months time is gonna be helping people who aren't the nerds getting AI agents to everyone else. Whether it's like someone has built a great AI agent and they need to convert it to something that my mom could talk to. How do we make Australia more AI-friendly and adopt AI in a faster way? The difference between Sydney and San Francisco is quite large, I find. When I come back to Sydney, I feel like I'm like falling back about six to nine months in time. Engineers here, in their off time, won't necessarily talk about engineers. But in NSF, even non-engineers will talk about engineering. You'll see engineers with seven tabs open, already parallel tasks, all making features and bug fixes at the same time, where it's in Sydney and still one task, by one task. Should AI companies look to hit revenue early, should they build the platform in the community more? This is a controversial one. I don't know if I should say this but. And we're back! Welcome to another episode of Over Subscribed. Today I'm joined by three of the best technical and AI minds in Australia. A great, agentic flavour to the podcast today. We have Pasha from A1 Base. It's like Twilio, but for AI agents, he's spinning out more agents than Agent Smith and Matrix in 2003. We have Anshul from EverLab. He's building autonomous healthcare. Being a customer for 18 months, I've seen great returns on my health and on the investor side as well, which is fantastic. And finally, we have Jackie from Relevance AI. One of the very first horizontal agent orchestration platforms in Australia that also taking over the world, massive series B recently announced with Bessamar. And this podcast is also powered by Relevance and A1 Base as well, which is great. So guys, before we start mixing it up, thank you so much. I know three of the fastest growing companies in Australia really appreciate your time. We're right here on the ground level. Surrey Hills, we're all the magic in the start-up scene is happening in Australia. But Pasha, just talking before we hit record, you've been back and forth SF San Francisco, the home of AI. You're over there once a month. Tell us about the differences on the ground, what the talent is like, and what's the adoption of agents being like in the US versus Australia as well. The engineers and the culture and the energy in San Francisco is really about pushing what AI can do right now. Whenever I go there every month, I'm always learning about new techniques and ways to get more out of AI, building new products, serving costumes in a unique way. When I come back to Sydney, I feel like I'm falling back about six to nine months in time because I feel like the culture here on the tech and the engineering. It's aware of what's happening overseas, but they're not in the weeds building yet. Interesting. So you guys are hiring. Penny, your co-founders over in SF now, thinking of hiring the team there or a bit of a mix. We're probably going to do a mix. I think there's a different culture between engineering and technical talent in San Francisco and Sydney. That kind of has the quality of the product you're creating. We'll probably do both, but they both have different strengths and weaknesses as well. Is it going to be where are you going to hire us? Or is it depending on who you get first? Have you hired any engineers yet? We have hired engineers. On Australia? In Australia right now. So we started here. That's kind of a proximity thing, right? But what we do want is the really crazy ambitious engineers in San Francisco. I feel like the cultural thing of engineering, the story of SF is that the engineers in San Francisco will try to just go all the way. To be the best in the world. Whereas in Sydney, they're really reliable. The great engineer is a good around building culture. But there isn't a desire to create the best product or technical thing out there. There's various reasons in the last case, but I'm sure there's elements of seeking some friends. We really make a lot of money in San Francisco. If I said Sydney, we haven't had as many exits. But yeah, it has that type of thing. Is that what you see as well between the two? Yeah, I think what you're saying, we don't have any engineers will suffer engineers in SF. But we have a lot of solution engineers. They work directly with customers. They've been absolutely fantastic. And I can definitely get what you mean by like the crazy ambitious in terms of that side. But yeah, it is a very crucial time. And I think it's really crucial to kind of educate your team about how is AI impacting engineering? Quite only, it's one of the big areas that it is impacting. And there is very much. What are investors or stock traders call it? Like alpha? There's an arbitrary. I've been trying to. There's an alpha. There's an arbitrator at the moment where it's like if you're using AI in terms of coding, you're going to get like honestly, probably 10 to 20x productivity, especially if you know how to use it right. And that is still not even like fully figured out by the world. Like how do you use like cursor or clawed code the best way? How do you even like write the clawed MD files appropriately? Or how do you even get like the agent to write the MD files so that you don't have to write it, etc. So even on top of that, like I walk around like some like technical spaces here in Geneva, San Francisco. And like you'll see like engineers with like seven tabs open or running parallel tasks, all making features and bug fixes at the same time, whereas in Sydney it's still one task by after one task. Which hopefully will change. Yeah, it's like the best practices is still being made. And it actually comes out from San Francisco to the rest of the world if I had. Yeah, yeah. I think, yeah, you definitely see that probably more from maybe like individual engineers in San Francisco. But I will say it's not like every single engineer in Sydney is like that. We have people on our team who like to push that forward. Like I've been talking to one of our engineers because we both do this similar flow. Where it's like we would make a bash file to essentially create multiple instances of Claude and multiple Git work trees so that it's very easy to now have like, you know, six to 10 like PR is running at the same time. And then all I find is just I will just like talk to it. And then just even test on like local host 3000 3000 and one and 2002. And eventually I'm like, oh, my computer's actually not able to handle. This video instances, but yeah, it's it there's definitely that like, you know, alpha slash arbitrage at the moment. How are you guys pushing that through the org? Honestly, a lot of education. I think as a founder being hands on and you know, telling people to do one thing is one thing. But like, if you're not, if you haven't done it, I'm very much a person like that. I like to get hands on. I really believe in the term like operate at all levels, meaning like if I can't operate as an engineer, I don't have a really good like taste slash instinct on what's going wrong. What's going going right, et cetera. So I just use it like, yeah, I started really pushing prod code with it about I'd say like two to three weeks ago. And yeah, pushing about like 12 PRs every week and that's why Brendan you you message me out like, oh, you saw I was online at three. Yeah, I'm like, yeah, I'm like doing work and it just like, you know, coding on the side. It also helps shore up a lot of weaknesses I would have like I suck at writing front end code, meaning I don't actually like, like never bothered burning like react and view. I'm great in terms of backing code, but like, yeah, so this helps me shore up a lot of that. And then, especially if there's like intricate like CSS magic that I need to do. It's like, I just ask it. I don't and then I still can read the code to verify that, you know, it's not doing anything like super on a fish or crazy. So yeah, do you get those PRs into production or do you then hand them over to someone else to clean up? No, no, I get it into production. I even, it's funny because I told the engineers because they would like put reviews on it. Yeah. And I'm like, great, just make sure that it's like a bit more, you know, contextual because I'm just going to plug the straight back into the, like I have a pretty good setup in terms of how I pause the comments. Yeah, the comments and everything. What do you have in the toolkit apart from Claude code? I actually mainly just used that one. We use our own platform a lot. So like, if I wanted to run Claude code just in the cloud and just for like a very small PR. So like, let's say I want to change this button from like, I don't know, on hover it changes color to like no, no hover change color. I would just actually just talk to one of our agents. It would spin up an instance where Claude codes is also spun up. It's also connected to GitHub. So then it would actually make that PR request and then I just look at that and I'm like, okay, great. Yeah, let's push that. Yeah. Before we get back to the conversation, I want to tell you about the sponsor who made today possible, Vanta. As a startup founder, you're juggling multiple priorities from the expected like finding product market fit to the unexpected like customer requests for SOC to ISO 27001 certification. But achieving compliance is time consuming and time spent on it is time away from needs of the business. That's where Vanta comes in. Vanta is the all-in-one solution for startups to become compliant quickly and build a secure system. foundation with ease. With the combination of automation and extensive partner network and a security marketplace, Vanta provides the necessary tools and expertise for startups to achieve compliance seamlessly no matter how urgent your needs are and at every phase of growth. Over 10,000 leading companies, just like the ones who have been on the Oversus Ripe podcast like EverLab, relevance AI, instant, alloy and Atlassian, trust Vanta to automate their compliance so they can focus on growing their startup. Startup customers get $1,000 off at Vanta.com/ over subscribed. And we're all A1 base customers here, I believe. Can you tell us a bit more like what does the Twilio for AI agents actually mean? So like our thesis at A1 base is that everyone's interacting with AI agents right now as if it was a tool. But ultimately, everyday people will talk to AI agents as if you're talking to a friend or a co-worker. And what we're building when we say the Twilio for AI agents is that we'll think communication layer so that when someone at work needs to message someone for, you know, like whether they're not the company by SMS, whether sending an email, whether they're sending like a team's message, they plug into our infrastructure which is built for AI agents to help make those messages get out to these other players really, really well. And I think for us it's like, does whole infrastructure or technical component to it which is like how do we actually enable that better for AI agents than just standard spam marketing? But yeah, like what we do is we want to help AI agents say, no, AI agents that are relevant so are never allowed kind of get to people in a way that they want to be communicated with, right? And you know, we're in this tech bubble here that we're talking about really cool things around like Claude code and stuff. But my mom still finds it really scary to talk to an AI. And like, you know, when you walk down the street, at least in like Sydney, most people aren't engaging with AI every like day at the moment. So we want to really change that in a way that's friendly and trustworthy for people. And do you think that's making it more relatable to people like Jackie, I remember I went to the relevance series B party of sorts and the out of your case studies, you had safety culture, you know, the great companies and all of those companies give individual names to their agents, which I thought was interesting. You know, they give them performance reviews as if they're a real person. Like, how do we make Australia, I guess, more AI friendly and adopt AI in a faster way? Yeah, I mean, hopefully we'll help accelerate that. We've been working with a lot of Australian companies. But yeah, I think the personalization does help a lot. It makes the person building it. It's like building, let's say a leg of piece yourself versus like someone else building it. If you start bringing a lot of these personified like actors, you just grow like more attachment to it. And as well as when it doesn't like, you know, work in certain instances, you're more like invested to want to improve on it, et cetera. Especially if, you know, there's a lot of like users for it. So that's what like one of our customer send payments at the other one that does the performance reviews for agents because they built a lot of agents and they're like, okay, which one is actually good because we want to like, you know, potentially keep some of these, but close some of these as well because they do use like tokens, et cetera. So a lot of these things help make it more comfortable for the rest of people. But it also is a matter of like distribution, meaning like, how do you distribute this? Like right now, you know, people, people would have to go into like, um, all these like interfaces. But what if they just want to talk to an agent, just via like an SMS, just via like WhatsApp, just via Slack, which, you know, why we decided now we want to use a one base. Um, so yeah, by having it more, you know, be communicating to you in, I guess, human-like channels, um, it allows you to be more comfortable with it. Another good example is like voice agents. We recently rolled that out. And that actually also, uh, it's quite fascinating how often people would actually talk to them, uh, even for personal use cases. Um, so yeah, what are the best agent names that you guys have at relevance? Like, if you haven't checked out the relevance marketplace, it's massive now. They're all names, most of them. Like, who comes up with the names? Do you like name the, you know, the shit agents after people you don't like? Yeah. So, um, we, I mean, the best one is probably the first one, which is Boch, which does, uh, sales, you know, BDR, um, business development. Uh, it's named after our head of sales, Roch, and then we just had it be in front of it because business development. Yeah, I am. Boch, um, and then, uh, that was actually an internal name and just somehow stuck. Um, that was a great one. But on our marketplace, uh, now we've, uh, slowly open it up to more people. So you can even charge, um, for those agents. So if you wanted to build an agent and it's really valuable, you wanted to charge it for like a thousand dollars, um, you could do that. Um, I've seen some, uh, really interesting names. Uh, some have actually also deviated away from like naming it like, um, a person. Um, some have like said, like, uh, use this agent to hire me. So it's, it's been interesting how they're trying to now do the, uh, you know, the whole YouTube, yeah, yeah, the whole, uh, YouTube thumbnail, like, tactics. Um, so yeah. But, um, I, I heavily recommend, like, personifying the names. I think what's interesting as well as like, on the, one point about Australia was San Francisco, what I find of founders and, and the teams here in Australia is that we care about the people more like, when we're creating these experiences, we kind of create a thing about the end user, where, whereas like, I feel like Silicon Valley is still in the weeds of, like, you know, better models and fast, which is still, they're both first year required. Um, but I love, what I love about being back here is like, Oh, how do we actually get agents out to everyday people, um, or in context that make more sense rather than just like as another, like, coding tool? Yeah, I understand. I find like, um, you, like, for example, uh, engineers here in their off time won't necessarily talk about engineering, but in SF engineers, even non engineers will talk about engineering. Um, like, uh, I just remember, I think, uh, when I was lost at SF, uh, it was probably around a half a year ago or something like that. I'll need to go back soon, but I was like talking to this marketing person. And he was talking to, to me about RLHF and like, going very deep in it. I'm like, wait, you're a marketing person. It's like, yeah, yeah. Um, I, I find this stuff fascinating because like, that's what that area all talks about. And I think that has a lot of like, it starts to fester where, you know, crazy radical ideas can essentially keep, you know, not be discriminated upon and people enjoy talking about it versus here, like, you know, um, there's a lot more of that. I guess like outside of work, you don't really talk about these things as much. So yeah, do you think that goes to work life balance as well? So like, you know, like, you just did by comment at the last cohort. There's a lot of videos of, you know, teams. I think I saw one from Den, another Australian founded team, the, you know, time lapse of them in their dorm working 18 hour days, seven days a week. Like Australian investors have somewhat found upon that approach, whereas in the US, it's more celebrated and stuff like that. What's your take on that? Well, I think it's just a spiritual thing, right? Like a student spiritually don't do the that type of 24, 70 grind in Americans tend to, I mean, it really is up to like, you kind of up into it, right? Like, you know, I think most founders quite transparently work a lot of how was we do it for the love of the game? I'd like to think, um, I don't wake up going, oh, I have to do some work. This isn't like being employed in an annoying company. This is like stuff that we love. But I will say that their Australian culture does have more of a work life balance thing. And it has its advantages than it has its disadvantages when it comes to like building a startup, but it's really up to the team that you want to build. Like, I mean, it depends on the industry as well, right? Yeah. Like the example I always gave is look at consulting or investment banking here. Right? Like, no matter how hard we work. Yeah. Like, even if you're doing the US crazy tech hours, maybe some of them, but it's probably not, not, not going to be comparable to like New York banker. Yeah. And so a New York lawyer or, yeah, those are exactly right. Like, Australian still work hard. It's just for whatever reason. And maybe it's the history of like the tech companies here. There's only been a few that've got into that corporate stage really quickly. Um, like by default tech here doesn't work super hard, but you still find, I call them the addicts. Like, I'm an addict. I think we all are. But you still find the people who, who love the game. Yes. The outliers. Yes, exciting. Well, we'll see there's less social pressure to do the the facetime in tech here, which is nice. So most people are more generally people are more efficient. Um, when it comes to getting worked on. And then we're talking about making agents more relatable, more personal, more, I know you guys are doing some very exciting AI stuff at EverLab. Like being a customer 18 months, it's an app I use every day. Like you're ingesting my last 30 years of unstructured raw health data, putting into a very beautiful UX that you can see on the screen now. So I mean, how do you think about, I guess 10 years into the future, like you're creating a new category called autonomous healthcare? Like are we going to see AI doctors? Are they going to triage? Like, we'd love to hear what you guys are thinking about here. Yeah, for sure. So when people think about AI doctor, for whatever reason, they imagine like either a humanoid kind of AI doctor or somebody on a screen, but it's actually AI. But if you think about what happens when you get a regular GP today, GP's got two jobs, job number one, give you some advice. Everyone thinks that's their main job. For the most part, their job is actually provisioning the party health care. services for you. You go to the GP, here's a referral for the chemist, a referral for a scan, for a blood test, for a specialist, and they're just orchestrating third parties for you. They get keeping Medicare and deciding, "Oh, okay, based on the information I have about you, usually that information is 10 minutes of verbal conversation and a tiny bit of historical data." And then they're saying, "All right, well, you should go over here." And maybe those guys will say something, "You'll come back to me and we'll catch up about it." And if you look at any of the digital health brands, any of the eucalyptus brands, updogkinson scripts, whatever, they've made businesses around taking one health care vertical and making them more convenient. You go to eucalyptus brand, each of them will sell you one drug. There's one for a financial aid, for a viagra, for a exempt, for the contraceptive pill, and pretty much what you're doing is you're saying, "All right, I will pay a premium so that I don't have to go to the waiting room, sit around, chat to the doctor, go to the chemist, pick up my script, and then do the whole thing again next month." We actually think that the doctor should just do the advice giving bit. Chances are that's what they're going to be best at. They probably don't want to be doing all of the admin work of keeping Medicare and digging through 30s of data and deciding where you should go and getting the results back. There's a human element here of judgment and empathy and kind of care navigation, which is really important for a human to do. Actually, what an AI doctor looks like is not necessarily maybe some part of it is, an AI-generated person on a screen, but far less than that, it's how do you automate taking all of the real-time data about someone, historical and into the future, deciding what their risk profile is based on that, deciding where the opportunities are, and then off the back of that saying, "Okay, we've got the ability, we've got the platform to provision the things which you might need. We're cutting all of that admin work out." When we started, it probably took 40 to 50 hours of human time per patient per year. Today, we do it in somewhere from three to five. I think by the end of the year, we'll have that down to a few minutes. We're doing that in a way where it's safer, more delightful for the customer, more reliable, and we're just shifting where the human time goes towards a higher impact activity. No, it's super interesting. I caught up with DST. They're a big investor early in Facebook. They're invested in Canva, Airwalex. I caught up with them last night in Sydney. Big fans of the show shout out to KP, Jonathan and Josef. They asked me, "What are the craziest pictures you've heard in Australia this year?" There were maybe lamenting a bit. They're not hearing these very ambitious founders as you're saying, "Pasha, maybe they see that more in San Francisco." I mentioned relevance, EverLab and A1 base. They agreed. It's like a crazy pitch. They took it back to be first heard, even something like Uber. If you heard the first Uber pitch by Travis Kalinag, sounded crazy that someone would drive up in a black car or their own car, pick you up, take you from point A to B. Why did you guys choose something so ambitious? You know it's hard. What was the term you use? Your hardcore or something like that? Adict. Adict. Not just love the game. You love the game. You know it's going all-consuming, your life's work or whatever you want to call it. How are you guys, besides being on this podcast, how are you going to inspire the next generation to go for those big goals? It's a big question. It's a big question. Break it down. I think it's what said, there's a bit about that, addict. I wouldn't push someone who it's outside of their comfort to run a startup. I think it's not for everyone. If you're chasing for the status of it, which some people, honestly, especially at SF, I've noticed that a lot. People will build startups just for the sake of it. I think it is a bit of that addict where you like building things. I think that was really cool for me. As soon as I, or even before uni, but as soon as I started uni, I spent my majority of my time not going to uni, but building things with my co-founder. That was just enjoyable. It's a little bit of that addict. I kind of get lost in time. I remember in fact one of our investors recently told me the story that I actually had forgotten about, but he said, "Oh, we were at the office at Friday." Then when he came back on Monday, we were still there. I had recalled that I actually didn't sleep much, but that was because I just purely enjoyed it. It's really, really fun to just build things. Even more fun when you build things that people like and enjoy and giving you good feedback or even bad feedback. Then with the bad feedback, I'm going to prove you wrong. I'm going to make this even better, etc. I just find that really, really enjoyable. I feel like as you build more and more things, your ambition of what kind of things you want to build gets bigger and bigger. It's when you build Legos. The first one is very enjoyable. That probably was just a mini car. Now you want to build a castle. Then maybe next, you want to build a whole city or town. There's that bit of mentality. If you don't find that enjoyable, I don't think you should force it. I think just because you're not in the same level of ambition as some other founders, that's not a bad thing in itself. If you do decide to go down that ambition route, I think having fun is kind of the key thing. Two questions about how to inspire that next generation. Specifically for Australia, I think one of the worst things about toll-poppy is that it's not in itself bad or it's kind of good to cut down people who get too big of a head and think they're better than everyone else. What I love about Australia's successful founders and people is that they're not going to be going to a cafe and being rude to the way they're just because they've made it unlike some other cultures that we're obviously all thinking about here. But I think one of the things that I think just for this next generation, at least in Australia, around ambition is that I think it's okay to give yourself the permission to try to build something that is well-changing. I think a lot of Australians actually, I mean, it's different for us. Before we're all at this phase, we're like, we've got some momentum and it's kind of cool. We've seen it a bit. But for the younger founders, I find that a lot of first-time founders here are like, oh, you know, I just want to be a small business. I want to do this. I want to make profit. And that's kind of cool in itself as goals. But a lot of the time they're saying it as cope, I think, because they're not able to tell themselves that, you know what, I want to show you something that's worth $100 billion. I want to make something that's like the first thing in the world that does this. And I think it's important to kind of tell that younger generation to be like, hey, it's okay to have those crazy ideas. If you succeed, you're not like, you're not going to instantly become an asshole, which is like, generally what people, I think in Australians, this is underlying assumption, you're a successful, you're a terrible person. Which is funny. When you talk to someone in corporate and they think about the bosses or the partners of these firms, they generally mean. And I think, I'd like to think, you know, like us three as well, you know, like we're able to show everyone that, hey, it's okay to be revolutionizing healthcare, revolutionizing the workforce, revolutionizing the revolutionizing infrastructure and playing for big. We're playing for the hundreds of billions of dollars of exits and have fun doing it. And like not, I don't think any of us will be dickheads. I hope it's a mega. I know these two guys from my age, everyone's like, we're just here for the game and we just love people and we love what we're building. So I think it's just like giving that next generation permission to dream big and it's not going to make you a bad person. Because once you dream bigger, you actually drive for more interesting insights, at least a better product decisions and business decisions as well. Yeah. I have an interesting take because it kind of, you know, comes into like even like how our engineers are, etc. But like, I think one of the things that makes a lot of these things harder is that in Australia, there's more enjoy, there's more random enjoyable things than just work. Like, I just really like building things. So that's like what really drives me. But like there is just other random things that, you know, in Australia that we have the luxury of. Whereas if in your own, I'm not not to crap on SF, but like NSF, there's not that many enjoyable things you can do outside of work, maybe like go to the mountains, but if you're not like an outdoor person, that's not that enjoyable. But in Australia, there's a lot of that like, you know, if you're really into running, there's a huge like run club scene here, like the city to SF. If you're really into bouncing climbing, there's a whole bunch of things like that. So it's like even just like, as I grew up here, society kind of commends you to explore a lot of these hobbies more. So you start realizing you start to enjoy things outside of just education and work. And I think that plays a part into it. Whereas in some of the more other cultures, you don't necessarily get the luxury to enjoy things outside of work, outside of education. I say this like I studied in China for like five years. All I thought about was like education. How do I ace my next test, et cetera. And this was in like primary school. I said I came here. It was like a whole cultural shift. It was like, but like education is not the primary thing I think I care about. It's like, oh yeah, there's you're like hell. There's like, oh, you can think think about like hobbies, sports, et cetera. It's like, oh, it's not just like the mask or that being said, I realized I really enjoy masks. But I think that's one of the things as well that kind of dissuades it a bit. So yeah. But what I think quickly to add is like what I love about a string culture is it's not just those marketing stuff, right? But like in sports and these other things, like we have a really good culture of working hard in reality. Like, you know, there's none of the, there's actually not that not that many status games in Australian culture, which is quite nice. Like you do good stuff in your good person. Like, you know, you're you're winner generally. And we have such a good sports culture. I actually feel like some of the best Australian people I had found has done a little bit of sports growing up Because we really like winning and you kind of see it in cricket and like in the football and everything else that we do Yeah, I'm in back to your back to original question I think all of us have also built a few things before and I think part of the part of the way I view going more complex is It's like when you win certain levels in the game You don't really want to go back and play that level again And so there's just something about like Mark and I one of my co-founders when we were we had this period of time We were just like looking for the next thing and we'd find these ideas and we'd be like this could definitely be big And we're like man, I don't know if I could do this for five years Right? Like like it's technically like another crud app like we know it'll make money We know how to do the playbook, but it's not exciting enough You kind of want to do something bigger and harder and Evolabs 100% that for us like it's It's complex in ways. I don't think anyone's had to solve before right? We have to deal with the real world people go get a blood test done at some random place Somebody loses a violent blood you have to figure out with the lab how to use the remaining blood to recalculate whatever biomarkers you're missing it's a It's a far more complex game and I think same for both of you guys and so I think there's also just like an element of You don't need to jump ahead, right? Like you can also you can also level up Do the do the easy things and then the slightly harder ones and at some point you'll feel like it's It's time to do something incredibly complex. Yeah, it's interesting like you're all I guess building in the infrastructure space to a certain degree I do think that's a massive Defensibility slash moat aspect now like real world things losing vials of blood You know going for a dexascar and going to do a physical in person. It's something that you know It's not easily easy to replicate on code I always say pain as moats Yeah, like whenever we see pain like people people always think When there's discomfort that there's something wrong like that's not comfortable things are breaking something's wrong But they don't have the mentality about the gym right like imagine you went to the gym and you're like I'm sore There's something wrong and there's obviously there's like good pain and bad pain like you can also go to the gym and Oh shit like a ninja. That's very different But in a business sometimes you're going through pain and you're like oh, that's good pain right because when we do this Nobody else is going through that pain and like the number of people around who are willing to go through this pain to build this moat Is probably going to be very few and so the more pain you endure the harder it is for someone to come after the things that you built Hmm, like I remember congratulations on your series a that was recently announced led by left lane A grade investor out in New York So I met with Laura from left lane when she came to Australia to do the DD and juice like Was blown away by the complexity like there's problems that you're tackling head-on like integration into the legacy health system Like I guess like can you tell us the story of like you sit down with Mark, Dr. Steven and Sam When you're tackling so many hard problems at one time and I know like you said a bit of pain is good But this would just be like you know as they say waking up every day and getting punched in the face with many different problems Like how do you handle that and as a team and I guess build that culture of you know finding people that want to come in and and Solve these super hard problems. Yeah, for sure. I mean one unique thing about us is we're a four person co-founding team Which is you know obviously a bit rare? We've got a doctor on the co-founding team which you'd be surprised this also rare and healthier And so I think we do a few things really well, right? All of us have very clearly defined areas were all got superpowers We're very good like we actually don't need to talk that much week to week because we all trust each other to go off and Own those areas really really well and we definitely need all four of us like this would be infinitely harder If one of us was missing just because there's so many different areas right like do I see even's been a doctor for 15 years He ran essentially like you know a private boutique practice. That was the beginnings of ever lab for six years before we found him So I think that's a huge part of it. I think the other part is we Interestingly are always more or less aligned like whenever there's a problem and we sit down It's never like any real big structural differences and what we need to do or we we need where we need to go And that's very unique about this business The nature of the business is such that we're incredibly close to our customers We have doctors jumping on calls with them. We have doctors as part of the team. We We interact with them just in such a deep way so often that we're never debating What should we build next? It's always a debate of should we build it in week two or week three? The the pool of what the customers want is very very clear And so it's usually just figuring out the details and the sequencing of what to build but You know, I think and I've done this before I'm sure you guys have as well But in the early days of the startup you often have this product market fit like are we building the right thing? Is this what people want? Let's go here. I was like gonna let's go over here and you're it's a very stressful time because you're You're questioning everything you do and for whatever reason we just never had that maybe because we started with What was fundamentally a services business in the beginning right six years of Dr. Steven doing this himself We had product market fit off the back of that and if it came far more a question of how do we scale this? How do we automate it? How do we make it better? I've been angel investing since 2017 made over 80 investments Many of them through 10 13 So 10 13 is Australia's largest network of angel investors and each month we send out a curated list of one or two deals That you can choose to join in on the journey These are companies like EverLab and instant that you've seen on the over subscribed podcast and other companies like go one Mr. Yarm and autograph if you're interested in finding out more about the world of angel investing I'm happy to jump on a call Share some war stories and tell you about the exciting companies that we're currently looking at Simply go to 10 13 dot VC forward slash over subscribed and you can book a call with me for next week And we can talk about all things start up and angel investing And I guess talking about hard problems going back to that and passion you're talking about Founders aiming for that 100 billion dollar outcome or mission doing a good mission. What are some of the crazy Pictures you guys have seen this year AI pictures You know, I told DST you three were through the crazy crazy so seen Working on hard things Anything interesting in your YC cohort that you went through recently? I mean, there's a lot of interesting things you've got AI for like pesticide discovery which is kind of cool AI for nuclear energy. I think there's like the thing is like What's interesting right now is that it's almost better to pick a Hard and vicious problem and put something like AI and and put that the talent to it now But yeah, I'm trying to do the crazy I mean I've seen some crazy pictures aren't AI which I won't share But no, I think what's what's really inspiring is that this wasn't like the B2B sass hero But everyone would be like I want to build a sass just to make something easier for another business where people are like Let me find the biggest opportunities in the world whether it's like chemicals Nuclear it's like communication infrastructure. It's like security and um apply The best and fastest technology to it and yeah, I think it's gonna pay off for a lot of people down the line You know um It is actually probably easier right now to take AI make Relatively easy money, you know like in terms of like Sell an app to a few businesses yet to like 100k turn a coin revenue. It's a little great milestones and everything Um, but it's inspiring to see a lot of people out there who are trying to solve nuclear energy, you know like like health Workforce is in everything and I think uh, yeah crazy pitches. I can get back on to some of the funny ones um But yeah, yeah, how about you guys? I always find crazy pitches are the ones that It's it's not my area of expertise so therefore I'm like, oh, that's crazy because I don't understand it um Say anything with like nuclear or anything with like um quantum actually quantum community I've understood more but like anything to do with like nuclear. I'm just like, yeah, no no idea But this sounds crazy and this sounds really cool. Uh, I think the in the AI space it's I feel like all of the ideas are probably ambitions to a certain extent I think you start to get like used to um just hearing about like, oh, this is just like um Etc. I think what I find more interesting is like how Specific it becomes like I find like for example, um, does this company that uh, well, it's a new startup um Ran by Joe Harris. Uh, alloy. Yeah, alloy Um, and they're doing some very very specific robotics thing and I find that really ambitious because that's he's he's speaking of space Where it hasn't really taken off but he believes the puck is pretty much going to move there like he's not he's not operating on where the puck is At the moment, but where the puck is going he believes like like robotics is going to be like absolutely You know, and I believe the same like it's gonna be absolutely the future in like two to three years or even less than that And he's building infra for that. I find that ambitious because it's ambitious in that the market isn't necessarily there yet But you're you're going to build for that market so you can be the first mover We had to do a similar thing AI agents was not common thing like I'd say like a year or two ago And we decided hey, we're just gonna build like the right info for it. We're gonna make people be able to like uh, you know Use it. There's a lot of challenges with that because If you're building for something that people don't really know how to build for how do you know what's the workflow etc like for example, GP, that workflow is super clear. Before agents, people don't know, didn't know what was best practices. Even now it's still a bit blurry. We've had to conglomerate like our own set of best practices. I find that pretty ambitious because it's like you're really operating in a lot of unknowns. Like how do you know like robotics data is going to look like X instead of Y? So what does that mean? You probably have to build in for us so that it can handle like either side or you're just gonna pick, hey, I just have to enough foresight or enough of a vision to know that it's gonna be like X. So I find those are the things where it's like, there's so many unknowns to how to solve this problem. I find those ideas really ambitious because it's like to execute on that, you're just operating on like pure instincts, slash knowing how to get the data. It's also really fun, but yeah. - That's quite interesting 'cause all three of you building in an area where you're one of the first to market, you're sort of predicting the future. I always like to say angel investing or the startup space, it's like the closest thing to time travel. Like you have to look at the founder now and the market they're building in and in the future sort of concurrently. Any favorite time travel movies going up? So back to the future. Expand days of future past. Back to the future. - Back to the future, the classic. - Back to future two. - Here's what Steinscape went. Is that going back? No, that's not. - Dark is really good if anyone's thinned out. - German time travel show. - Does that mark one of Mark's favorites? - No, that's surprisingly not. - But yeah, like time travel, like how did you guys come up with the foresight to identify these areas? Like, and Jackie started as a vector database. We're doing okay and then pivoted to chaining and then agents, I guess. Like how did you identify that healthcare? Was moving that way and yeah, passed away. - If you need to project the future, like one of my favorite examples is airpods. For years, for decades, I dealt, and everyone dealt with tangled earphones, right? And somehow when airpods came out, people like, oh, this looks stupid, that'll never work. Right? Yet this pain just became invisible to everyone. Like the entire world somehow became invisible to this pain of airpods. And like usually you're not making bets about, you don't need to see into the future about the problems people will have. Because the problems people have are fairly consistent. Like they want to do less work, they want to be more entertained, they want to feel better. Like that's never going to change. And so usually the only bet you need to make is that I can find a way to do this thing better for these people. And every now and then you might, you might build a business around like regulation or compliance, like you might say, oh, I think GDPR is going to happen and I can build a business around that and that's more of a future looking bet. But usually most things, like there's a lot of markets out there. We can just look at the market and say, all right, well, this kind of sucks in this way at the moment. We don't necessarily have a better way to do it yet, but at some point we'll have the technology to do airpods. We'll have the technology of AI to make the workforce easier. Like it's surprisingly not that hard if you reframe it that way where you just need to say, all right, how do I take a problem? I'm convinced about. And there's something coming up very soon and that'll make it a lot easier to do. I think I think the difference between like a founder and an investor, right? Well, always when articles get written about founders and startups, it's always after we figured out how to pitch it to investors and pitch it so there's something compelling about it. But like that's kind of like people think the market opportunity was the cause of the company or the startup, but it's really that most of the time founders find something interesting that we can fix that you were saying, right? It's just, oh, we can, that could be a little better. You know, like before we started A1 based, I was doing art school and I was like, actually this could be a little better. And this is like, we were doing it generally more for fun or for interest, I find. And if we pivot to something else, it's because as we were trying to solve that problem, we pivot into something more bigger, there's a bigger opportunity that we see. But I think like the cause of a good idea that ends up being big is actually more like having fun and having that instinct about what could be a bit better for everyone else. And then we learned to pitch it in a way that makes people like Brendan happy, which is like, oh, the market is huge. And sometimes, you know, like it's always a bet, right? Like I think a lot of the times when people talk about markets being big, like there's almost this assumption that it's like a contract you're making about the market being big. But really it's more like a poker bet, you know, like, oh, I think there's like a 12% chance that this market's going to be huge. But it's like for me as a person, as a founder, it's more worthwhile for me to put that time in because it's fun and interesting and it might pay off. Where as for investors, you know, it does shift where it's like, okay, like we've proven enough of the bet and this is how big it can be. So let me try to get you some returns. But I think that that difference is found is that like usually finding something that could be made a little better and whether it's exploration or something else, that's probably like why the next, like the core cause of a good company is more of that kind of instinct. And then the story of how it becomes a big market comes afterwards. - That's interesting. Speaking of investment, one of my favorite topics, very fortunate investor in all three of these amazing companies here today, not financial advice. But as you can see, fantastic founders, tell us about your latest fundraising rounds because they're all very recent, all very fresh. Might start with Jackie up the end. So Bessima lead series B, the same partner that actually led the first ever Shopify round. Absolute goat. Talk us through how the rounds came about, how quickly it moved and what it's been like working with one of the goat investors. - Yeah, no, Jeremy is amazing. I mean our whole board is really amazing, but one of the things that Jeremy does really well is that he will read the situation very quietly and then he'll just drop almost like a very pointed, specific question about the situation that pretty much like helps you almost like summarize the whole situation in a nice way and almost like get straight to the key point. And he always provides examples of like how other companies have done it that he've seen. Without being like, hey, this is the exact advice, but like just giving you as an example so you have more confidence in the decision. So he's really, really good in that. And also Katie and Libby, who's also on the best team, they're absolutely fantastic. They just, they just hustle so hard. They share us like how market trends are looking and everything just absolute fantastic group to work with. To be honest, I don't really like to give advice about a series B because a series B was, it was pretty straightforward. Like we got that round via an inbound template. Like it was preempted, we weren't looking to raise. We, in fact, I shouldn't say this, but we also didn't really shop it either. We're just like, we're so focused in just executing. That way, okay, this is a nice offer. Jeremy and Katie and Libby are just great people. They've been like, you know, giving us referrals to customers, sharing us really interesting inside. They interviewed a lot of our customers and they just did all of the necessary due diligence upfront so that we didn't have to spend that much time operationally. So it was just a no-brainer for us at that point. So, and I think it came for just like, you know, just pure like hardcore execution. The company was growing well. Customers were really speaking about us. And I think that was a really key one. And probably we'll be really key for our next round whenever we decide that's the thing that happened. But like, I think the customer part is really, really crucial. And I think one thing I've noticed specifically about SF investors, your customer DD is incredibly important. Like, what your customers say about your product outside of like just you know, you speaking to them is really, really crucial to how your product is represented. That's why you notice it's almost like a group of these like, companies as well that get fund raised simultaneously because they really happily like to use each other as well. But yeah, that was how our series B was preempted. We negotiated and then we executed. And then we moved on. So yeah, our series A was a lot more like true seeking learning about how to do the fundraising process. But yeah, awesome, awesome. And before we move on, I remember you used to have a thesis about AI startups generating revenue too quickly, which can sound a bit counterintuitive. But can you expand on that? Like should AI companies look to hit revenue early? Should they build the platform in the community more? I don't know if I should say this but. So what were you for? I, you know, I semi-regreted us not doing this. But we, you know, it's been a much longer journey. But I believe that if you're building an even specifically in AI but not really. But like, I think you should stay for on pre-revenue for a long, longer. Revenue is a good indicator, but small amounts of revenue is not a strong indicator. I believe like if you don't have a really, really clear path to go from zero to one to five, like within a year or two, it's not really that worth getting from zero to 300. And then like, you know, consistently like that. Because you're just going to get a lot of weird signals in the path of going from zero to 300 versus if you go from zero to one mill. I say this because we actually had to reset our revenue. Like, not recently but like for our CZ, Ceres A, we reset it our revenue. We had a pretty good revenue number, but we knew we wanted to reset it because we, you know, we pivoted. And going back to pre-revenue, we were a lot more diligent about how we were going to get revenue. In fact, I feel like we should have been even more diligent. Like, we were a lot more about like, okay, we're going to be very careful. careful about how we get the first customer, the second customer, and then from there being able to scale quickly. Because if you're doing it a bit too ad hoc, you start to create bottlenecks in how you serve customer, etc. And if you can just, and if not too much revenue is involved, it's actually better to remove those bottlenecks. So yeah. I got some advice in that space as well. We moved Michael Siebel for why I see this batch and he was like, the two things that it's actually with those two points, which is like, one, if you need to create the magic, create the magic first, because like, you know, that's, that in itself will always take time. But the first, the advice was, is very Michael Siebelism is like, the first million you make should be the same revenue you make at a hundred million. Yeah. And I think that's like, such, it's like, it's kind of like, it really simplifies the way you even think about your business when you kind of like, get to that point where it's like, that revenue I'm making, I'll be the same revenue I'll be making as we make it. A hundred percent. And I think it's even more, I think maybe, yeah, even specific to, I bet like in the generation of the pace of revenue growth is like, incredible. And yeah, like how we made our first million is very similar to how we made like our recent, you know, many millions. So yeah, it's, I think that's something to just hone it on a bit. And also, maybe this is more for the V, more for the founders that know about VCs, but like, it's very mathematically driven. So as soon as you make revenue, it's not on high pace, you're applied against that. Like it's a formula. It's a four, if the number to multiply from is zero, then they have to come up with a number. But if there's a number that they can attach to, then it's like, oh, you're at a hundred K. Well, if I do a hundred extra multiple, then that's, maybe that's not great for your evaluation. But if that number is zero, then they can think, okay, well, how likely is this company to make like a million in a year, two million in a year? So yeah. And then Pasha, you had a experience going through YC for a second time. How are you thinking about, I mean, you just close a great round. How are you thinking about the next round? As Jackie said, are you going to push for revenue early? What's the plan here? Yeah. So I mean, like we just did YCM and did our post YCM around. What we're thinking about right now is like, that's given us the cash to really build the magic that will enable the next wave of AI communication of humans overall. And I think one of the things is like, during my common year, we did make sales. We sold the companies. We get people on board. We do have customers. We're always learning from them and trying to improve for them as well. But we're still in the phase where we're not rushing revenue. We're trying to consistently grow, so we keep on learning. But yeah, we're not at the phase where we're like, okay, we're locking in to get to 1 to 10 to 100 of this period because really, I think for our particular company right now, we need a bit more time to create that magic. But we started at YC. So we started in the January of 2025. And that's still like, that we're still cooking as the way I look at it. I understand. And actually to be a bit more specific, the advice is not to not get customers. Just don't hone in revenue. So hard because it's actually quite interesting. Because even if you look at some of these insane growth stories, they also, what's not talked about a lot. And I'm sure everyone's seen the curve of like, oh, cursor 100,000 fastest end of the next, like another company. But usually those company actually had spent like two to three years actually building the product. And often they'll actually remove it from their LinkedIn history, etc. And it's like, yeah, just exploring random things almost something like that. The company was found out. Yeah, stealth founder. Well, because it was like a CAD company first or something, right? Yeah, it's like, yeah. Like even they, they had to, you know, many years of like actually just like pivoting and trying different ideas. And then they found the one. I mean, even lovable, right? Like that's the recent, uh, quickest like 100 mil story. They were GPT engineer for, I think a solid a year and a half or something like that. So they had actually spent time on that problem and that is a bit of like what's in the media versus like reality. I say media, even though it's very specific to the start of space, but like, yeah, I will say like, it's important to get your first customers as quickly as possible, I think because to just really build up that magic, you need a feedback loop of if that magic is good. But to get revenue scaled, I think that matters less in these days. And it doesn't help you as much. If you, if you're not in a belief that I can get to like zero to 100 was the same strategy. No, no, all of us are as lucky as an angel here at Phil Vellab and I think they're the right thing from day one. Yeah. Yeah. It took us to the FLAB seed round. One of the biggest ever seed rounds in Australia was left lane. It's very first Australian investment. You've had, yeah, revenue from day one, the FLAB standard protocol. It's $3,000 for a year. There's many hundreds of people on that now. Yeah, if you, if you're interested to hear your thoughts around, or a, how did that come about? Why did you decide to raise and, yeah, pretty interested to hear about your plans? Yeah, for sure. So I mean, we, we did have revenue from day one, but we, we kind of went through, I'd say, a similar eight, nine month learning period. Like we had a wait list on the website for a long time. It was in private beta. You were one of the very early customers. We spent a lot of time learning. Every time somebody got on a consult with a doctor, we used to spend two hours preparing a personalized flight deck. We had a lot about the narrative, the breakdown of the data we found, all the historic data, digging through all of it, presenting a really clear what's wrong. Why do we think it's wrong? What do we think you should do about it? That was two hours of every consult, and every member would have six, seven, eight consults every year. So it was very time intensive, and we, like the guy said, we're very deliberate about taking that time to learn and not trying to ramp up revenue too quickly. But then I'd say over that first year, we got really good at identifying, okay, we understand this problem, we understand what people like, here's how we peel away the humans, here's how we automate it and a way where we can continue to scale. And so I think we came into this year knowing that we definitely had product market fit, that we could go really quickly. We definitely didn't need to raise from a, like we're need the runway or something, but it was very much a, okay, like there's a massive opportunity here. We've got the beginnings of something really great. It's not just here, it's actually all over the world. How should we, like there's a real point continuing to display it slow? Let's start to ramp and let's see what raising another round could look like. And so yeah, shopped it around the usual, you know, went all over the world really. And I think what these guys were saying before, there's something a bit different about the ambition in the US. I think one, they've obviously got a few more leading data points, you know, all of the earliest furthest ahead cutting edge players are from there. And so I think some of their conviction comes from seeing detraction that comes in the US first. But then I think part of it's just cultural, right? They're willing to take bigger bets. They're willing to see further into the future, understand how the market's going to evolve, the consumer mentality is going to evolve. And so we had a lot of interest in the round. Ultimately, in the end, we needed to pick a lead. Went with Lifeline. They've got a ton of not just kind of consumer experience, but also healthcare experience in that space. And they just kind of saw the vision immediately from day one. They just got it straight away. They flew down. They spent a bit of time in person. And yet like it was just kind of the right fit. And then we obviously had a lot of the guys from the first round, like yourself, and a few other strategic people who we brought into this round. And it actually got wrapped up pretty quickly from our end. We thought it might take longer, but in the end, we got to go back to building a lot quicker than it otherwise could have taken. Nice. And so the main problems constraints on EverLab at the moment, you need more engineers, more GPs. Would you say that capacity? Yeah. And say capacity, to be honest. We play this game of, again, we're a very real world business. And so we have real world constraints when it comes to supply. How many doctors do we have? We definitely always need more doctors. And we find doctors really enjoy working with us. If you think about the typical GP life, you're often in a room with no windows and you're dealing with coughs and colds all day and they're these five minute consults. And you're not getting to do what you became a doctor to do in the first place. And so they come to EverLab. They get to work in a team. They get to have really long consults, have a ton of data work. A bit more at the cutting edge. I'd say we're not necessarily doing things which are like fringe or frontier science, but we're still very much kind of foundational medical just solving the basics really, really well and then going above and beyond where we can in that area. And so, yeah, doctors love working with us and so we're doing a big push at the moment of higher, a lot more doctors. And then all the other things which you would have gone through, whether it's physios, radiology providers, how do we build the supply more broadly to service all these different needs for our members? It's exciting. And how you guys think about international go to market. Obviously Australia is a massive market. I think Eucalyptus Proof, you can build a unicorn just in Australia, even though they're in a few other countries now. Like how you think about, I guess, Europe, US or is that some of the secret source still to come? Yeah, I don't know if I'd commit to where we're going next on here. But, um, Internationals definitely on the books. Like where we're gonna dip our toes in the water a little bit this year and next year look at it a bit more seriously. And the way we've set things up, it's not insanely hard for us to go international with the model. Like we've always kind of known from day one again, not because Australia's not a big enough market, healthcare and any market is huge. But just because, like we were saying before, we're gonna play the game again. Why not go bigger? And also I think just from an impact point of view, we do care a lot about the mission with all kind of as a founding team, even a team more broadly. I think once you get to a certain age everyone sees people go through healthcare problems that were entirely preventable. And so if we're gonna play this game why just constrain it to Australia? Why not go global, see how big that impact can be? - That's awesome. Autonomous healthcare going global. Very excited to see what that looks like. Speaking of going global, so it usually takes a startup on average, I've seen maybe three or four times, to get the US go-to-market example, right? Jackie you shipped Dan V, your co-founder over to San Francisco, Pasha, you've got Penny in San Francisco, you guys are back in Sydney. You seem to have got out of it pretty lightly, Jackie, six months, you said, since you're last. - Yes, that's true. - But do you think you guys have got it right the first time potentially in San Francisco Dan on the ground? - Yeah, it feels so, because majority of our revenue comes from the US and yeah, Dan's done a fantastic job and the team over there, like we have some of the best solution engineers, salespeople over there that just represents relevance and just are like doing an absolutely amazing job to make customers super happy over there. So yeah, I think we've done well and I think the cool thing was just making sure we have full time boots on the ground and not treating it as like, oh, we're going to expand there. Let's see it by a hiring one person over there. We said we did that, but it was already with the intention that Dan Vass was gonna move there. And I think there's still some things we've yet to crack. So for example, marketing over there, there is a bit of a very specific way to get into, I'll say like the Twitter crowd over there that we haven't really explored slash cracked to be honest explored, but yeah, otherwise, I think at least from the sales motion side, it's been going pretty well. And I think what was really crucial was just not just treating it, someone say our companies would expand to any countries, just not just treating it as like, oh, this is like another revenue point, but let's learn about the culture, let's learn about what the customers needs are and everything. And I do think Australia does both well into it 'cause there is actually a lot of similarities in both sides. So yeah. - Yeah, that's, and what if you pass your flying back once a month, I believe? - At the moment flying back once a month. I think a lot of it's really just to talk to our customers and see where the frontier is on AI. It changes very quickly. So that's why we're like, and every trip, I actually we have a good technical and product learning that will kind of embed into our product and infrastructure as time goes on. But yeah, I know we treat America as a real serious market. Like it's like you can't half, and even America, there's like submarkets. I feel like Australia is very homogenous. Like Sydney and Melbourne don't really feel like different cities in the scheme of things. Whereas you know San Francisco and like Houston and New York and even Oakland, which is right across the root from San Francisco, they feel like different cities and countries to be quite honest. So all of these subtle things. And then that's not even talking about the industries. But we do, no way, we, I fly back and forth to really set us up for that foundation. 'Cause we're not at that full revenue stage as you mentioned before. But like it takes a lot of up to even lay the groundwork, have the leads understand the different nuances between industries and cultures and subcultures and all of that type of stuff that is probably the required word to crack America. - Yeah. It's also helped us actually a lot in Australia because of the fact that a lot of American companies use us because Australian companies don't love to be the first adopter of technology. I wouldn't say like startups because I think startups that ecosystem has really, really improved a lot more recently. But like, you know, back in the days, like, if you're talking to, I don't know, like one of the large, like let's just say the banks, they're not gonna adopt you if like no one in America has adopted you. They're not gonna take that risk. And even for more of like, let's say the technology companies that aren't like a bank, let's say, I don't know, like a e-commerce store, they also don't have that strong preference to go local unless a bigger company have tried it. So that has actually helped us a lot because now that we have those logos, there's customer stories we can talk to, the customers, the leads here have actually accelerated a lot more quicker. - Yeah, I'd like to one thing about a store that does annoy me is that because like, I think we've, someone had this interesting point about the Australia's legal system. Like we have a very strong anti-tech legal, like sort of laws in Australia, but that's because what you realize is all of those laws are built for like these giant, well-funded American companies entering Australia are fully baked, already hundreds of millions of R&D in. So by the time like a startup comes here, like you're up against like regulatory hurdles, every big enterprise buyer has already only bought these giant American companies. So it's changed recently, but it is a tougher market to crack because of that. And it's almost easier sometimes I find to crack, at least in the end of our sales space, America, because there are a lot more cowboys sometimes if new users, sorry, if a new startups to buy from. - 100%, but hey, if that's the problem, then the solution is just to go to US, make sales and then come back to us earlier. (laughing) - I mean, how often, like a trend I'm seeing is a lot of younger founders are going to the US a lot earlier. There's a very robust WhatsApp group, Aussies in SF, I believe it's called very active and yeah, people give me an each other advice on how to get the visas, where to stay, et cetera. Like how do you regret not going to the US earlier? Do you think people should go to the US straight away, a hybrid approach like A1 base? I guess it's different for everyone, but have you changed your opinion on that since your last startup? - I think for people that have never been exposed to looking valley, that want to do tech startups, it's important to get some exposure first. You don't have to live there or anything, but there is something when you go there and you see a founder who is like 17 and wants to make the next 100 billion dollar company and they have a chance to do it, right? Like it is actually, for someone who grows up in Australia, it's just important to see that I think that shift and also the speed. I think what was Australian young founders as well, if they want to work hard, they want to put the effort in, but they have no one to model that against. And I mean our company is like, we're a bit older relative to some of these young founders, so they don't see it as much in their friend groups as well. So I think it's important to go and see in model against just to see, you don't have to adopt at all, but it's like, oh okay, like I see how hard they're working to look in valley, I see how ambitious they are, and then I might adapt that to what I'm doing and what I think is important. I think that's important for a lot of first time Australian founders, but there, it depends. Once you have that, then it's kind of a tactical decision for your company. - Yeah, it's also a lifestyle, right? Like if I really can't live in SF, but like having gone there, it's every trip has been very helpful. Like for example, to be honest, the first trip, I think it wasn't very helpful because of the fact that like COVID was happening, so the city was quite dead, I'll be honest. And then I think the third trip where we went there, that was actually really, really helpful because the city was back alive, and then now that's when I started to understand, okay, this is why it's called Silicon Valley. Like it's all these meetups, it's all people just like, you know, interacting in person, talking about all these things. Now what's really, really interesting, I think specifically for us going there, we were expecting a certain kind of trend happening over there. For example, back in I think what it was 2023 or 2022, we thought everyone would be talking about AI agents. - Hmm. - No one was really talking about it. It was only on the Twitter crowd for a little sub-segment that we were subscribed to, that felt like this was all they were talking about, but they weren't really talking about it. They were just talking about, oh yeah, AutoGBT tried it, cool, doesn't work, live on next thing. And I found that actually pretty interesting because it's like, you almost get like a reality check of like what is like all these smart people really, really think about. Like right now, RLHF is talked about a lot more recently, but I think it was like six to nine months ago, I thought they would talk about it a lot more than, you know, a back then. And it's like, it's interesting every time you go there to check in online, okay, what is like everyone talk about? And a good thing about not being as part of that bubble sometimes is that like you don't get as much of the noise and you're kind of more tuned into reality of like, this is how you solve the problem because it's very, things travel fast over there, where it's like one second, you know, I think we got that package where it's like, MPM really quick, then it was called a bun. For a while, everyone was talking about bun. We used it, we were like, oh, this is all right. But, and then just like quickly dissipates. But it's like, it's interesting to go to just to see like what is like the trend and how long does it last? Is it still the same? I say, months ago or are they why they not even talking about this new trend that they should be talking about and you come back and then next time you go and now they start talking about it so it's like it's not like they always are on top of it as like a whole community there's little assumptions sections but it helps like make sure that your thinking is not too over tuned but I strongly advise like if you're if you like living there you should live there like I think the Australian mentality actually does help a lot being there because I think you're able to like sift through what is noise and what is reality a lot better because of that like just more kind of skeptic we just have a lot more built in skeptic illness so yeah and talking about looking at trends into the future what are the latest sort of agentec trends even on the workflow side that where are we going to be in six months or no you're at the cold face yeah we I don't want to drop it too much but um work yeah we're going to this what we found this new protocol um actually no I shouldn't say we found this new protocol we found so one of the big things that's talked about a lot is that MCPs right now we've been extensively testing this new way of doing it and it is wildly better like wildly better without and that's the only thing you need to change you don't have to change like necessarily what how the protocol works from an architectural standpoint but usually what people put into this protocol is not talked about enough and we found a way that like actually just performs way way better and we're we're going to probably do a bit of a like um you know thought flow leadership stuff around that um but that's been a really really interesting thing and I think that's going to keep it advancing and one of my like kind of takes is that actually one of the things that people talk about is like prompt engineering is dead I actually think prompt engineering is very much even more alive the the the kind of advantage you can get from being able to know how to prompt agents write or even using LMS to figure out how to come up with the best prompts as well is even more important than now especially because these models are not getting wildly wildly better every iteration uh and that's because of how like data uh works like the data is becoming more and more limited um so now you have to think about how this data is formed how can you um almost like prompt uh your agent such that in a way is your you're almost like hitting that specific sub segment of how the data was like the models were trained to be able to make sure that it's in that domain versus like another domain um etc so like a good example is like maybe in like healthcare uh in order for you to get the best agent you actually almost need to have it be talking in like uh the prompt needs to be very healthcare-oriented but also you have to give it the ability to talk in code as well um so it's like does these weird interesting mixes that I find really really fascinating about like prompt engineering and I think it's actually going to be really really hard to fully automate that away it would there will be like optimization techniques so then it like you know RLHF that it really matters on how you capture the user feedback data but um yeah that's one of the trends I'm really starting to see it right now yeah nice and sure what does autonomous healthcare look like in six to twelve months oh you're just asking for our roadmap uh what can I what can I say I think a lot of our early members and you would have gone through this they somewhat came for the diagnostics they came for a look under the hood I'll always say there's nothing more powerful than telling people about themselves like look at horoscopes Buzzfeed quizzes those if those do well we should be able to do well telling you what's going on with your body uh and I think that the big shift we're starting to move towards is uh just looking after you for the rest of your life um and it took a lot of work to get to the point where we could and do the diagnostics the data understand people really well understand what to uh give them next uh and uh I think it's kind of what I alluded to in the beginning but what does it look like to actually be your GP look after you for life but do it in a really scalable way how do you solve the economics of that if you think about GPs or doctors in general we spend like eight to twelve years training each one of them we spend millions of dollars training them and then we pay them a lot of money each time like it's uh it's actually a really high leverage point if you can do it in a safe automated high-value way to take more of that time away from areas where they're not delivering the most value and reallocate it to them doing more kind of human to human work awesome and finally patch up what does the a1 base of six to 12 months look like it's really interesting because like when I think a1 base since two to 12 months I kind of see a lot more consumer adoption of AI like every day people today aren't really using AI in their day-to-day lives and I think what we where we've seen like the models go and the AI go now is that they can now we always like to think about AI if the output like long output that it gives but actually what's happened today is you can give an AI a lot of context even to give a two sentence response and it's actually quite a magical experience for everyday people and it's funny right like a big context big amount of context injected doesn't actually make the reply that much slower it's usually the other way if you have a long reply then it feels slow for the consumer um so I think what what we're going to see is like a lot of our work going around um helping get AI agents out to people across different channels and different platforms even maybe our own channels as well because what we see is that um AI agents and the context that they can now hold about a person about a task about about memory and and all of this type of stuff is gotten good enough to feel like you're working with someone else or chatting to a friend and I also think like you know um the a1 base of six six months or 12 months time is going to be helping people who aren't the nerds getting AI agents to everyone else you know whether it's like someone has built a great AI agent and they need to convert it to something that is something that you know like my mom could talk to or you know someone who isn't doesn't love tech can talk to I think that's going to like be a big unlock for AI agents in the future and I actually think we're in a nice position both at the infrastructure layer and then the consumer layer to help like facilitate that and let everyone work with AI agents as if it was another human being in those kind of channels. No nice. We guys really appreciate you all coming in today super busy building amazing hard companies. Why should people listening now like we've got a lot of engineers for example why should they come and join a hard but good quest in relevance, overlap in a1 base what's the quick page. I think if you really want to if you really like passionate about AI there's no better place to be than relevance and specifically around agents or even just pushing the fold of like what's possible whether it's internal use of AI and agents or even just externally you know where it plays to be as well as if you really like our culture we're very fast hard working culture we have tons of I guess like people who are very varied in age so that has meant that our team really enjoys spending time with each other so yeah if you if you like our mission building the home of the AI workforce trying to solve really hard problems without any clear indication of necessarily how to solve them then then come join relevance yeah you're only going to find it really interesting so yeah there are probably three reasons or three things that our engineers say they like about us one for us the engineer is kind of the lead role that drives projects and so you find a lot of engineers out there a lot of engineering cultures the engineers at the end end of the assembly line right you know product manager comes up with something they give it to design goes to the engineer the engineer builds it what a lot of engineers actually one and they're incredibly good at you know I never really loved the idea of the product manager is accountable but then he's like I'm just waiting for the engineers like it works a lot better if the engineer owns the end project and they're delegating to hey I need help with writing the copy I need help somebody do up designs I need help from PM to help think through this problem or be my internal customer and so all the veteran engineers at the moment own the entire stack of their projects and they get to be really kind of product first and think through everything end to end I think the other thing everyone says is impact we genuinely like at the end of every week in retro we run through a list of the lives we saved we we find genuinely life saving things for people every day and so we're close to our customers in a way which actually like even me personally haven't experienced before you feel the impact of what you're doing but you also feel the everyone in the business has a very good intuition for what customers want because they are a customer right the the healthcare aspect of what we do is universal and then I'd say the third is we solve very unsolved problems and we're we're solving like I said before problems that don't just exist in in code but also in the real world and it allows you to spread the the kind of muscles which you're using not just into UX land or engineering land but also operations and third parties and incentives and the problems just become far more multi-dimensional and far more interesting if you're kind the kind of person who enjoys solving hard problems yeah so anyone bases much earlier than these two guys and their companies What I would say is this, as a technical company, we like going deeper than in understanding something that people don't want to touch anymore. We understand mail servers, we understand telco infrastructure, we're going deeper than I think is reasonable so we can figure out stuff that no one else has figured out. We're so early that right now it's just going to be a fun journey with a small team in the co-founders. If you want to join us thought about that super early phase, we're here on an adventure and we are incredibly ambitious because I think we are tackling an opportunity where if we get it right, it's not even just financially interesting, but we are able to reshape how people have fun, like interact with their friends, their workplace, their, basically, even all of their systems that they do every day. I think we're super excited to see that world happen and we want to be the ones that make it happen at an infrastructure level and at a consumer level. So, yeah, come on the journey. Awesome. Thanks so much for coming in. Guys, great to be on the journey with you as all. Thanks for the inspiring words today. Jackie from Relevance, go to relevance, AI.com. Check out the marketplace so you can build an agent in natural language, which is great for people like myself. evalab.com.augp's. If you want, there's never been a better place to work in my opinion, the GPs. B2B as well, big B2B push. We've got some great companies coming on looking after their employees and a1base.com. You have a1base mail. Just came on mail.com. A1mail.com. Check it out. Great tool. Thanks all so much for coming on today. And before we go, if you look over at our screen over here, we have our North Star guest, Cliff from Canva. It's been 14 days. He still hasn't replied to our emails to be our next guest. We're going to go in Cliff. We're going to film an episode in Canva. It's going to be fantastic. It's going to be big. It's going to be beautiful. Thank you so much for joining us on over-subscribed. And we'll see you guys next week.