The dawn of surgery bots + buy a home for $250 (w/ Andromeda & Mogul) | E2313

0:00Syndicate, something that's worked incredibly well, allowing a group of people to act as one unit to buy an asset. People want to buy into single-family rental properties. Real estate, as we know, is the world's largest wealth generator. 90 % of millionaires became such through real estate investing. The idea is buying shares in a single-family rental home because it's completely headache-free. What is the goal of your investors? They want to get 10, 15 % back every year in a dividend, or they want to get the appreciation. The assets typically yield between 8 % to 12 % per year for dividend payout.

0:36That's dollars that you can withdraw tomorrow. How do you make money? How do we at Mogul make money, we're saying? This Week in Startups is brought to you by MongoDB. AI-assisted and agentic coding is helping you build faster than ever. Start building at mongodb.com slash AI. CLA, innovation takes balance. CLA's CPAs, consultants, and wealth advisors can help you get from startup to where you want to end up. Get started now at claconnect.com slash with you. And Northwest Registered Agent. Get more when you start your business with Northwest. In 10 clicks in 10 minutes, you can form your company and walk away with a real business identity.

1:16Learn more at northwestregisteredagent.com slash twist. Hello and welcome back to Twist. This is Alex. Now, have you read the books in the Expanse series or seen the show? One of my favorite parts of it is whenever they get hurt, they get shot, they get irradiated, whatever it is, they go back to their ship and they put their hand and arm into a little device and it fixes them. I've always thought that was a beautiful vision of the future. No matter what's wrong with you, there's this machine that can put you back together again. Now, we're a long way from that today. If you go into the doctor today, it's a person.

1:46If you get and surgery done, probably also a person, and on down the healthcare chain. But there are companies right now working on building us towards a more robotic and more autonomous healthcare future, including in the realm of surgery. Now, that will either delight you or terrify you, but I wanted to learn more about where the technology is today and how quickly it's advancing. So please join me in welcoming to the show. It's Nick Damiano, the CEO and co-founder of Andromeda Surgical. Nick, how are you? Doing well. Great to be here. I forgot to ask before we started. Are you a science fiction fan?

2:16I am. That intro is perfect for you. Let's start by defining where the company is going. You said on LinkedIn that, quote, surgery is currently almost untouched by AI, but in 10 years, AI and autonomy will be ubiquitous. So paint that picture for me. Where are we going to be in a decade? And then I want to go back to where we are now. Yeah, I mean, like a lot of labor in the world, surgeons are going to be totally transformed by AI in the next decade. Right now, surgery is way behind where it really should be in terms of where we are in terms of tech. So we're somewhere between the, to use a driving analogy, the horse and buggy era and the car era now.

2:54There's not really much autonomy happening at all. Whereas for driving, of course, we have Tesla, Waymo and others deploying autonomous cars on roads now. So we're taking surgery to that point where it's going to be mostly autonomous. Surgeons will still have a job and a role to play, but a lot of it's going to be done by AI. So now you guys currently have a system that involves an iPad app, endoscopy, so we have a camera. We can look and see where we are. You work with surgeons, so it's kind of surgeon-mediated, and there's also a robotic arm doing some of the work. It might be better if you do explain kind of what the current Andromeda system is and then what procedures it can kind of currently tackle.

3:33Yeah, so it is a robot that's controlled by an iPad. The robot is fairly simple. It's an off the shelf arm. We don't want to be a company that builds lots of complex hardware. The exciting part is really in the software. And so what's happening on the iPad is the important part. We are starting with urology with this procedure called HOLEP. It's for BPH or enlarged prostate, something that almost every man gets at some point in life. And then we're expanding from there into other urology and then other, a lot of other surgical procedures to where this can be a general platform doing most types of surgery eventually.

4:10If I was thinking about the realm of surgery and, you know, as a non-surgeon, I'm clearly an expert. I don't know if I would have gone with urology and endoscopy because it involves putting things in sensitive places, which to me seems a little bit terrifying. So just because I'm such a lay person here, why was this the right procedure to pick as kind of the company's wedge point into our future, more autonomous surgical future. Yeah. One of the things you wouldn't expect in surgery is that urologists were the first adopters of robots and really one of the most tech forward types of surgeons. So like intuitive with the DaVinci robot started in urology initially.

4:46There are a lot of procedures though, like this whole procedure that are not done with robots now. So DaVinci covers some of it, but then we're covering basically everything else in urology. And we have these surgeons that love robots that really want new tech to do these other procedures. And this whole procedure specifically is one that's really hard to do. And you've got this huge gap between somebody who's just learning, who takes like five hours and struggles through a procedure, and then the best in the world who's doing it in 30 minutes easily. And so we want to elevate everybody from the novice on up, up to that level of the best surgeon who's just cranking through eight or 10 cases in a day with no problem.

5:26This begs the question of what in the current system is autonomous, what is educational, and what is assistive? I'm trying to figure out kind of today, what can a job to bring to a surgeon who hasn't done a whole lot procedure before to help get them from, oh God, this sounds scary to I can do three or four or eight of these in a day. So right now you can do it robotically with the iPad. And so that's like just kind of table stakes for a robot. You can do the procedure using our controls. And then we've built some early autonomy features. One is like Google Maps for the body, where it tracks the landmarks and what you've done in the surgery so far in real time throughout the procedure.

6:06So you always have that spatial sense and just know where you are and what you need to do next. And then there's also, we built this auto pivoting where instead of having to worry about all the degrees of freedom of the scope, it controls the pivot point and kind of auto adapts to force to minimize, as you can imagine, in the area that urologists operate, you don't want to have these huge forces applied where it's going to cause all kinds of damage. And so we automatically minimize that, which we think will reduce the incidence of things that are damaging in that space. And then we're releasing software every week.

6:43So we were in the cadaver lab, we're in the clinic doing live cases all the time, and then we're building stuff. And every week we release new features and are improving the autonomy level to where it's becoming more and more autonomous and then easier and easier, of course, for the surgeon. Yeah. To be clear, I was not trying to be rude by saying what autonomy exists yet. I think you're laying the correct foundation, the right technology to get there. I'm just trying to figure out kind of like, as we draw this arc forward, where are we starting from? Because currently, as far as I can tell, there's no autonomous surgery being done.

7:12So we're you have to build this. Now, data is an interesting question that I talked to you about with everyone in the broader AI space. And, you know, scraping websites does not apply to you, nor does bringing in enterprise B2B workflow data. How do you get the information to train systems about human tissue and surgeries and such? Because I would presume there's a lot less to get, to scrape. So how do you get the right stuff to teach these robots to be more autonomous? The data is hard to get. That's one of the things in this space that's really challenging, and it becomes a moat for any company that operates in this space once you get that data.

7:48You can get surgical videos from YouTube. We've actually talked to a lot of surgeons that are YouTube creators that have shared some of their videos. So we have a huge data set of that type. But then there are other things like the robot kinematics, the forces, anything else you can pick up in the procedure that is not available from just videos. So we have to get that data ourselves because no one's ever collected that. So then being out in the clinic, doing these cases is how we get it. Right now we've done, I think the only, we've done 45 cases and it's the only data set of 45 cases that exist.

8:21Like that's a hundred percent of the world's data. And it's really hard for anyone else to get that. Yeah. So then as you want to add procedures, you know, moving away from just hole up in urology and endo urology, I think is the phrase that I had to learn while prepping for our chat today. Are you going to have to go out and do that as well? Or are there data sets that exist that you could bring to bear on bringing more autonomy to robotics? Or is this going to be, you guys have to create it every time you want to add collarbone surgery or whatever it is? So it is scalable in a few ways where we're not just starting from zero.

8:55One thing is that you can break surgery down into a series of surgical tasks. And so once we can make tasks autonomous, as those might be repeated in different procedures. And so that's one way we're gonna be able to scale faster and faster as we build. And then we can also go back to the tactic of taking videos for different procedures. There's people even building databases of surgical videos. So to at least get started, we can access those or talk to the surgeons we work with, get videos and build some basic features based on that and then expand from there. Is there a flywheel effect? Like as you get a bunch of data from other parts of the body, like if you get really good at working on livers, does that apply anywhere else?

9:33Or is the body sufficiently specialized to that? What works for kidneys in terms of pressure or sensitivity or vision or whatever does not apply to, I'm going to name another thing, a gallbladder. Yeah. Because the set of surgical tasks people do now is finite and pretty small, then yes, you can generalize across procedures. If you do a, say a prostate enucleation versus an eye enucleation, there are similarities to that. I mean, so in different parts of the body, you can repeat some of the stuff you built. And that's how we're planning to scale from maybe one or two new procedures a year now, then up to four or five, hopefully 10 someday.

10:10Okay. So you've identified a real problem and you put together a solid solution and a business model that you believe in. So you're all set to launch your new company, right? Not so fast. If you want investors and potential customers to take your new business seriously, you need to consider forming a Delaware C Corp. And that's where Northwest Registered Agent comes in. They're going to give your new company a real identity. That means an address for your public filings, a domain, a custom website, a business email, and of course, a phone number. And that's going to take just 10 minutes and 10 clicks.

10:41They don't charge hidden fees. Customer service is available around the clock. They're not overwhelming your inbox with spam and they make it easy to cancel at any time. So get all the advantages of a Delaware C-Corp independent, regardless of where in the U.S. you're operating from. Visit NorthwestRegisteredAgent.com slash twist for more details. And the links are in the show notes. You just mentioned that there's kind of a finite number of surgeries that we do. That's a very interesting comment to me, because when I think about it, surgery does seem to be relatively broad, but it sounds like it's actually more narrow and more systematized than I might have thought.

11:18It is. So there's a finite number of surgeries, And it's the world in general is more bounded than, for example, driving. There are just fewer edge cases. Not that there aren't edge cases. Surgery is the body is more variable than you would think. But it's a lot less complex than the entire outside world that cars are operating in. The data is more scarce. So you don't have as much training data to go on. But then the world is less complex. And most importantly, you don't have these other adversarial actors like other cars or pedestrians. You have to model their behavior when you're in surgery.

11:48There's nobody else working against you, fortunately, when you're in the human body. No, that makes a lot of sense. I mean, because the body can't fight back because I presume everyone who's getting the surgery done in this way is unconscious. So they're not even moving. They're just pretty much static. No, I appreciate that. So when you decided to build the company, was it kind of made possible by the fact that there are now off-the-shelf robotic arms that are sufficiently finely controlled and workable that you can build this? Is that a new thing in the market? That is new. There's this arm from KUKA that we use right now.

12:23They're a German company that makes robot arms for different use cases. And there weren't that many until very recently you could use off the shelf. We could have possibly built one, though it would have been a lot more complexity for us to take on. I think also for the world being ready for this, there are two factors that had to exist before this for people to really embrace this concept. And one was surgical robots in general, mostly DaVinci from Intuitive. And then autonomous driving, because now people are accepting that cars can drive themselves with Waymo and Tesla and so on. So those two things coexisting make the world a lot more receptive to what we're building now.

12:59Why use off-the-shelf tools versus something that you designed for your own use? Now, I love the idea of getting to use whatever people have around loosely. But to me, it sounds like if you're building something for a robot, it might have a different shape, a different sharpness than something that a human might use. I'm curious about that. I mean, there are great tools for surgery already. And we want to focus on what we are, what we're good at as a company, which is building this autonomy layer. And so while it does help to have a robot, to have some physical incarnation where you can actually control the whole autonomy loop, it doesn't really help to build custom tools.

13:38So it's a lot faster for us to be able to focus on the software and not have to build a tool for every surgery that's done. Companies that choose that game, which is basically every other robotics company that exists so far, will find themselves moving a lot slower and a lot more inertia than what we have. Yeah, you've really kind of excised all the things that you don't need to do in-house, focused on data collection, software improvements through an iPad mediator. And that does seem to be a really fast way to build a better system. but that does kind of bring us into the what happens when it works and how doctors fit into this.

14:11Now, on one hand, every surgeon that I know is completely overloaded, exhausted, burned down, about to fall over. On the other hand, they're quite highly paid. And so there's kind of a bit of a trade-off there. But to me, the idea of letting surgeons do more work more quickly, you know, with the help of autonomous tools makes a lot of sense. On your website, you guys talk about this being a sous surgeon, which I really appreciated that. That's a good framing. But down the road in time, maybe not 10 years, maybe it's 15, I can also see this technology improving to the point to which it can handle routine stuff like taking out my appendix if it's going to explode, for example.

14:46And so how do doctors and surgeons feel about being accelerated and then possibly down the road made slightly archaic or anachronistic in certain procedures? I think like a lot of areas that AI is expanding into that the job of a surgeon will change fundamentally, but it's not necessarily going to go away. I mean, there may be a point in the future that all of our jobs go away and it's all just AI and we'll have to learn to reckon with that when the time comes. But for this one, I think it's going to be just a surgeon seeing their job in a different way than way before it becomes something that totally replaces surgeons.

15:21So what we envision in the long-term is a surgeon sitting at kind of an air traffic control center where there's a lot of different procedures happening. They can orchestrate that at a high level, but not have to do all the tedious, mundane tasks all the time. So maybe you're overseeing 10 surgeries instead of one right now, and your job in each surgery is different. It's more high level and less just routine, low level stuff than it is now. And most surgeons we found are really into that, that people, they want things to be autonomous. They want to have these unpleasant, mundane tasks taken away up to the point that they're totally cut out and don't get paid anymore.

15:59And that's a line we don't really worry about crossing. Well, I think we're far enough away from that that that's going to be a great problem to solve once we get there. But it does sound like, as we've seen software development get automated to some degree by AI in a very condensed time period, this is going to be similar-ish, but over a longer time period. For example, now a lot of developers orchestrate or observe and guide their agents versus writing the code themselves. This sounds analogous to that. Although I do think it's going to take a longer time to get the data we need to tackle more surgeries than it is to get more coding data, which is indexable, available, and online.

16:36How could you guys go faster? That's the question we ask ourselves every day. So we were really, besides, you know, the, I guess the goal of the company to just get to this, this like big autonomous surgery vision, we're really focused on being the fastest iterating company in the medical world. So we want to take the iteration cycle down from what it is now for most companies, it's years down to a month or weeks. And we've done that in the clinic so far. We've been basically doing cases, then building new stuff, testing on the bench, cadavers, then back into cases in less than a month. And it's been shocking to the surgeons that have seen that there are way better features this month than there were last month.

17:18That's not the way the medical world has ever gone. That's not med tech. That's the inverse of how medical technology is usually expanded. So of course they're blown away. They've never been treated like a first party customer before. They've always been treated as the thing that gets approved after the insurance company says yes. So yeah, I bet you're popular. Yeah, it's really just focus and automation that are the key principles behind that. When I talk about not making every device for surgery, like anything that goes into the patient and just looking at every step of the process and asking ourselves, how can we make this faster?

17:49How can we automate parts of this? that's the key. And I think most companies don't do that in medtech, especially people just are not into automation. They're, they're following a, an old playbook that is usually quite slow and it's really hard to, to scale. Well, that means you can hit them like a hurricane then, cause they're not even playing the same game. So that's quite, quite nice. Now going back to the car, the car analogy, self-driving as kind of a thing to be figured out that's going to help people get to understand this. One thing that self-driving companies do, and we've talked to Wabi and Wave and just everybody.

18:20They're very interested in world models and they have tinkered with synthetic data. And I'm just curious if those two things have any play in your realm, or if those are things that don't work quite as well instead of dense biological systems. You could describe it as a world model in a way. So I think that concept, and I've seen a few different ways to define that concept, but I think that is basically what we're building. In terms of synthetic data, we're open-minded to that. We haven't used it at all yet. There are definitely drawbacks to using that, and there's ways it could go wrong if we were to go down that route.

18:56So for now, we've been using an approach that's more, I guess, more down the middle and not taking too many liberties in that regard. Though we will try things out and see if that's something that produces good results as we move and make this more autonomous, go to more procedures and look for other ways to scale, maybe synthetic data generation capabilities get better and better to where we do want to use those. And one thing we have to just think about as we do this is that AI is advancing really fast and we want to be able to ride that wave and stay up with whatever the latest capabilities are.

19:31Literally stealing my next question. So Nick, I'm curious about how advances in general AI technology is improving or aiding the company because it seems like in the era of Fable, We've made a lot of progress since GPT 4.5. So tell me. I mean, we use AI for a lot of our operations. We have to make sure that like with things like with coding that doesn't introduce non-deterministic bugs that end up causing big problems. We have to put a few guardrails on that and make sure that we're not overusing it. Yeah. Definitely for all the documentation processes that bog down companies usually, I mean, AI is a core part of everything that we do in that regard.

20:05So in every process that we have, we look to use some kind of LLM or some AI to improve the process. So it's really been about optimization. And then, yeah, for all the paperwork, it's just massive in terms of how much effort is saved. I was more thinking about, as we consider the idea of a sue surgeon, again, using that phrase from the website, how much does an increase in generalized AI intelligence help you get to that point? Like, are the models smart enough today to accomplish your goals? Or do you still hope for some step functions and generalized intelligence you can bring to bear on your specific problems that will help you resolve, you know, other use cases more quickly?

20:46One of the themes we talk about over and over again on This Week in Startups is making sure you do your chores. I'm no expert on these things. I have some experience. Steven Estes from CLA is an expert. Let's talk about being cash efficient. Tell us about efficiency and what you see in the top tier startups in your practice. We're seeing kind of an interesting trend out there where companies aren't needing to raise quite as much as they had in the past. You really have to be careful as a founder to only take on as much money as you really need. You've got to do the forecasting, you've got to do the modeling, and you've got to dial it in and get it right.

21:22Otherwise, you're going to end up either not raising enough capital to get to where you're going and you're going to have to go get venture debt or go back, have an extender to the round, or you're going to give up too much of the company because you just didn't recognize how much money you actually needed. Yeah. Very important to get this stuff right, folks. And that's really a bummer when startups don't do things in a button-up way. Always have a great partner, a good partner to have on this adventure. While things change, my friend Steven over at CLA. So if you want a trusted advisor by your side who will navigate you through taxes, accounting, and everything in between, it's time to take action.

21:55visit claconnect.com slash with you and don't forget to drop a mention that your boy jacal sent you that's claconnect.com slash with you start today i think it would it would in the future that if you could have a better semantic understanding of what's happening in surgery i don't think it's quite there yet to be as reliable as we want we have to be careful in the beginning to not have things that are going to hallucinate in weird ways so we're not using as much generative the AI in the first version of this product, though, that is a thing that I think there's huge potential as it gets better and better, especially with a surgeon in the loop where they can have control over what's happening and prevent things from getting out of hand, that we can introduce that more and more.

22:38For now, though, it is more of the kinds of AI machine learning that's being used for autonomous vehicles that's been used in Tesla and Waymo. Okay. Now, two questions more. One about capital, one about commercialization. So So when I think about robotics, I think about catastrophic levels of expense. Nothing but love to everyone building robots. It just takes a lot of money and time, it seems. You guys are using off-the-shelf technology for the arms. So I presume that lowers your overall capital needs. But does the company have access to the funding that it needs? And how interested is the venture community in backing you?

23:09Well, we were about to announce a round that we raised recently that will come very soon. So stay tuned for more on that. Do you want to give us a hint? Like, is it tens of millions? Is it hundreds of millions? It's tens. There we go. It's between zero and 99 million, everybody. Breaking news. Yeah. So we've been both very capital and time efficient so far. We were the fastest robot from zero to launch in three years, and we've only spent about 15 million total in getting there, whereas the norm is hundreds of millions. And that's because we don't go crazy on hardware. We've moved fast. Time is really expensive.