The TechMed Movement Has Arrived | LSI Europe '24

Alan Cohen 00:06
Hello, ideas. Good morning, everybody. Thank you very much for the kind introduction. And I'd also like to thank Henry and Scott and the whole LSI team for having me today. Thank you. You've actually helped me by inviting me to the event. You've actually forced me, over the last couple of months, to push hard on my thesis of techmed. What I'd like to do today is explain a little bit about how I came to this thinking about the industry. I'll share a little bit of my own personal journey to it, talk about what I think it is, what it means, and share a couple of examples of how it's occurring, and hopefully leave some time at the end so we can make this a little bit more interactive. I'll also talk a little bit about deep tech. One of the things that was very striking to me about this week is when I think about what's going on in AI. I had the pleasure of working and funding a company by a man named Yoshua Bengio. Yoshua Bengio and Évelyne Kuhn were graduate students. They did their PhD under Geoffrey Hinton, who was recently of Google. They are effectively the parents of neural networking and AI. If you think about how neural networks work,

Alan Cohen 01:24
you inform multiple nodes. If I think about the LSI event, you meet somebody, they introduce you to somebody else, they bring in knowledge from a third party. That's how neural networks work. They collect information from the outer peripheries, and they feed that back, and that's actually what machine learning and neural networking does. So that's a little bit, I think, what I've been able to get out of this event. Just in the last couple of days, I've learned a lot about how the industry works. I'm not a traditional med tech investor or a med tech entrepreneur. I've met some amazing people, and I've seen some really great companies, several of which are super interesting to me and the firm.

Alan Cohen 02:06
So how did I get here? And then we'll talk about how techmed got here.

Alan Cohen 02:11
I come from the tech industry. I've spent a couple of decades, almost three decades, in tech. I started my tech career working for two little companies, one called IBM and the other Cisco. At IBM, I built e-commerce businesses in the mid-1990s.

Alan Cohen 02:32
People may remember June 5, 1995, was my first day at

Alan Cohen 02:38
IBM. It was also the launch of Windows 95, and it was also roughly the time frame of the IPO of the Netscape browser.

Alan Cohen 02:47
I worked there for several years. I launched several businesses. I'm not an engineer. I actually had to start this from scratch and make that journey. A couple of years later, I moved to Cisco during this internet wave, and Cisco at the time, and probably still for a good part of even today, was the underlying infrastructure for the internet. Over 95% of the traffic on the internet ran on switches and routers that were built by Cisco Systems. The two lessons I took away from IBM and Cisco in building businesses there are that it is much easier to build a business if you're satisfying demand versus creating demand. If you looked at the internet enablement of organizations and individuals, the growth of e-commerce, the growth of Amazon, and the ease of access to information as well as the ability to transact on the internet drove an enormous wave of investment and consumer and business behavior. That was kind of my orientation for getting started. In the early 2000s, I got very excited about the startup world, so I was where many of the companies here are today. My first startup, Tahoe Networks, which is conveniently not on my slides, was a gigantic fireball. It was a terrible miss. It was the only company backed by Jim Goetz, who's in Excel, who's now at Sequoia, and Jeff Yang at Redpoint, that actually didn't make money. I managed to work for that one.

Alan Cohen 04:19
Tahoe built infrastructure to allow mobile devices to connect to the internet. The problem is that there were no 3G devices, there was no iPhone, there were no tablets, there was no Android. So we were all dressed up with no place to go, and the company ultimately exited as kind of an aqua hire for Nokia because it had very good technology. So I kind of learned my lesson, and then I worked for three very pivotal companies in the technology industry as Chief Product Officer and COO. The first one was a company called Airspace. Airspace was the first company at scale to allow people to send

Alan Cohen 05:00
up wireless LAN systems. In fact, we won the largest customers in the world, but we didn't do anything that somebody who wasn't paying attention to at that time, Intel, put a chip on every laptop called Centrino, and pretty much everyone wanted to take their laptop and connect through Wi-Fi to the internet. So we did a great job at satisfying demand. The company was acquired by Cisco, and I had a brief stint back with my old employers. It was much better to be acquired than to take a job there. It was definitely a way to go. I then worked for a company called Nysera, which virtualized networking. We turned it from hardware into software. It was acquired by VMware for $1.3 billion; we had $400,000 in revenue.

Alan Cohen 05:43
We were acquired because we were actually the pivot to pressure the entire change in the industry and the growth of the cloud from a networking point of view. So it became a very strategic linchpin. And then my last company was a company called Illumio, which is a cybersecurity unicorn. It's a company currently worth $3 billion, and it turned network security into software. So I think you'll see two themes here that I think are very important as we think about the techmed thesis, which are that hardware and software are fusing and changing in very transformative ways. And number two, the drive from the end customer is always going to be the most important element. Then something happened in my career. In 2016,

Alan Cohen 06:32
my father was diagnosed with glioblastoma. He had surgery at Mount Sinai Hospital in New York, and during his cognizant year, he basically talked me out of working in tech. He asked me, "What are you doing? You've done this a couple of times. Do you really need to do this going forward?" I thought about it, and I called him up one day, and I said, "Thank you. On the bridge to mortality, you just put a bullet in my career." He really made me start to think about where I can put the remaining professional years that I would have

Alan Cohen 07:05
more productively into doing something that would actually have a more long-lasting impact on society. One thing I learned about my father as he was dying is that he was a Navy corpsman, and during the Korean War, he was called up as a reserve. He was too young for World War II, but he was in the Reserves, and he was parachuted from an aircraft carrier—not literally parachuted, but moved to Korea—where he was on the wrong side of the Chosin Reservoir when a very large Chinese army came across. He actually conducted field surgeries as a corpsman. I still have his surgical kit from 1950 wrapped up in the canvas, with all of the sutures and all the things that he had. It got me thinking about the practice of medicine.

Alan Cohen 07:57
I was recruited to DCVC. They were investors in my company, Evolve, and I joined them. Initially, I worked with three companies. Evolve Technology is a company we built and took public in the weapons detection space. It uses an array of sensors and computer vision to know whether you have a gun or an iPhone in your pocket. If you went to the Olympics this year, you walked through Evolve. If you walked through one-third of major league sports in the United States, you will walk through Evolve. If you go to the cricket stadium in the Australian cricket stadium in Melbourne, with 100,000 people having a good time, you'll walk through that. I also worked

Alan Cohen 08:37
with Caption Health, which might be a little bit more familiar. Caption Health satisfied a problem that we have: a huge gap in the number of stenographers and doctors that can complete EKGs using ultrasound. So we built a software platform using algorithms to actually provide guidance for non-trained medical staff. It could be a Navy corpsman, it could be a medical assistant—not a nurse, not a nurse practitioner, not a doctor, not a cardiologist—to actually follow the guidance and capture the EKG, including the ejection fraction, accurately. That was a company

Alan Cohen 09:17
that, two years ago, was acquired by GE Medical, who saw the competition with Butterfly and the movement from large cart-based systems to portable point-of-care ultrasound handheld devices and said, "We have to have it." At the end of the day, that's what happened. I spent time also, I mentioned, with Element AI, which was the OpenAI of Canada, led by Yoshua Bengio's team.

Alan Cohen 09:43
When I make investments, I look for one critical factor, and it's part of our deep tech thesis: are we providing a level of human enablement? Are we giving superpowers and skills to people to do their job

Alan Cohen 10:00
better, to be more effective, to provide more information and data to that?

Alan Cohen 10:08
I've done this in robotics. I've done this in cloth manufacturing and wireless power. I think many of you have heard of Proprio. I was the lead investor in their Series A, which I believe is creating superpowers for surgeons, and I'll talk more about that. So

Alan Cohen 10:29
I kind of feel like if I can make my way to techmed, my background is a layup for you all because you actually understand this industry a lot better than I do. So let me take a minute and just define deep tech from our perspective. I think the other takeaway I got this week would be useful to explain how we invest and how we think about companies. I brought something for you. I've had somebody shipping. If you see on the table in the back, this is the DCVC deep tech report. It highlights several of our companies. It also highlights our techmed thesis, but it will give you a sense of what an investor in my space is looking for. It's very simple: we look for passionate entrepreneurs that are bringing a computational advantage to some scientific or tech sector and providing an enormous amount of leverage to what we think about as trillion-dollar problems. That's literally what we do. We swing for the fences. We fund rocket companies, robots, microbes that replace chemical fertilizer, and also passionate healthcare-oriented entrepreneurs that want to see a huge transition in this industry for many reasons, including democratizing the ability of lots of people to take advantage of this care, as opposed to folks that have good insurance and are fortunate enough to live somewhere where there are great doctors they can get access to. So that's our focus.

Alan Cohen 12:10
Let me talk about techmed. It's really not very complicated.

Alan Cohen 12:15
Our techmed thesis, and you can see this in the articles and in our report, is a speciation event. In biology, a speciation event is when a new species emerges out of an existing species. So techmed is emerging out of med tech. It's not something completely different, but it starts with a different focus. It's when an array of new technologies—not all of them are brand new—which include an array of sensors, computer vision, machine learning, and, most importantly, proprietary data sets, large individual, regional, population, and cross-anatomical data sets come together to be able to provide the medical establishment skills and capabilities to perform highly relevant clinical interactions and provide better outcomes for patients. Not really very complicated. But the thing that's different is that we start from the technology basis, and I'll share some examples of that, and then think about how they get applied to medicine, as opposed to somebody who is in an area, or they're in cardiology, they're in orthopedics, they're in an array of disciplines in medicine who go looking for technology to solve the issue. We come at it, and sometimes it's quite accidental. So that is what effectively techmed is.

Alan Cohen 13:51
When you look at AI technologies like ChatGPT, you're seeing what we think about as the superpower game of AI, where people are spending billions and billions of dollars to train large models, right? These models

Alan Cohen 14:08
take in audio, images, written words, text, knowledge, and they string it together, right? That's effectively back to the neural network. It's a very, very different game. In fact, if anything, Larry Ellison was quoted last week, by the way, if anybody saw that Larry Ellison, it's worth seeing. I want to look like Larry Ellison when I made it; he looks terrific. I don't know what he's doing. I don't think it's plastic surgery. I think he really has great doctors, and he's taking care of himself. He thinks to train the models that are going to win, the one or two winners, is going to take $100 billion.

Alan Cohen 14:45
I know there are a lot of companies here that are looking for funding. That's a different game. That's a very different game, but with that, you can use something like ChatGPT or MidJourney or Gemini or an array of tools.

Alan Cohen 15:00
You can ask it to create an image. It will do that. You could say, "I would like to see Gandalf, the wizard from The Lord of the Rings, fly fishing."

Alan Cohen 15:08
In fact, I might do a quiz. One of the images that are in my presentation was actually generated with ChatGPT; I used it myself for the presentation.

Alan Cohen 15:18
When I look at the opportunity, and I'll talk about this in somatic AI, it's actually the opposite. What the medical industry can do is take imaging data, the brainpower of practitioners, their knowledge, and actually create models in the opposite direction. You can take images and create knowledge, as opposed to taking knowledge and creating images. It's almost the inversion of what we're seeing in the ChatGPT movement. Now, if you're thinking about AI in your business, and we are AI-first investors, the DC in DCVC stands for Data Collective. We are big nerds. My partner, Zach Bob, 15 years ago, started the Big Data Drinking Club in San Francisco at Founders Den. We invested in lots of the infrastructure for big data, including Elastic, Confluent, and Databricks—companies that are really the on-ramps to do that. There is an enormous amount of money going into AI, but most of it, as you can see from this Goldman Sachs report, actually is going to the traditional industries that tech serves. So you see the information industry, financial services,

Alan Cohen 16:31
education—things that you would expect.

Alan Cohen 16:36
The growth of AI in healthcare, which, by the way, is a larger industry than many of the industries that are inhaling AI faster, is actually moving at a slower rate. I think that's part of the challenge, and I think it's the opportunity for the folks here at LSI and the people building in this industry to smartly apply AI into what they're doing. We should be seeing more of that money not just going into building models to write marketing copy, but into figuring out how to diagnose a pathology slide through computer vision, to work with patients, to build things like digital therapeutics and ways to keep them on them. To me, I think this is the fundamental challenge in front of the industry, which is that AI adoption is slow in an industry that, frankly, is dramatically more important.

Alan Cohen 17:34
Asking ChatGPT a question, or asking, "What do you think my tax return should be?" is fun, and you can do it, but using it as a practitioner or as a patient or somebody in administration to provide better health outcomes, to me, feels like a much, much more important use of this technology. This is something I think we as professionals have to really keep an eye on because I think this should not be a lagging indicator. It's going to start from a smaller base, but it should have a faster trajectory.

Alan Cohen 18:11
So here's an area, and if you're looking for a business plan to share with me and other tech investors in Silicon Valley, is something that I call somatic AI. The term somatic obviously relates to the body. I stole some images from my company Proprio. What you're looking at, many of you here know, is a view of the spine. One of the things that a lot of the companies that are building visualization systems can do is they can give you an enhanced view. This is a 3D view. It can also be brought down to the millimeter to examine for imperfections, to look for growths, and all of those things. It can also measure when something is in alignment, or if it is off. Somatic AI is what I think is going to be a branch of

Alan Cohen 19:06
technology that is going to dramatically enhance our ability to understand human anatomy, the interaction of anatomy with drugs, procedures, and medical devices. Now, for somatic AI to become successful, we need to fulfill the data promise of all of this, which means I serve a lot of really interesting companies with implants. Are those implants actually providing data about what is going on?

Alan Cohen 19:39
When you're looking at one of the challenges with a lot of imaging is it's a point in time, right? We all know if you have a check-up and you go to your doctor, and your doctor takes your blood pressure, it's terrific, except he or she took it once a year, and it's actually fairly irrelevant.

Alan Cohen 19:56
So the goal of somatic AI is to

Alan Cohen 20:00
master information, data about the human body, about people's health, about their interaction with all of the factors that they live their life—what you eat, how you move your body, how you sleep—and actually what your doctor and other professionals provide to you on how to administer your health. Somatic AI is a way to see that all, to visualize it, right? This is a very visual industry, right? It's either one or two things in the medical industry, so they're very physical. You can see what's happening, or it is microscopic, and it is about cells and molecules. So I'm talking about the physical side of this. This is an area that I think is going to really keep growing and take off. One of the ways to do that is to build better ways to continuously monitor information and build that data river. I stole this slide from our company Mic, which is a company that's built a monitoring system called SickBay. Its initial focus, it was built by machine learning scientists, is to take all of the measurements of a patient in a hospital—anything from infusion pumps

Alan Cohen 21:21
to glucose monitors—all these things and continuously monitor them. They did that. By aggregating this, they could actually make clinical staff a lot more effective, right? I mean, we know we have a nursing shortage, and if you are so the ability to map an entire hospital's population and know when something's happening. But the real Trojan horse about SickBay is that it is measuring. It is creating an enormous data set

Alan Cohen 21:50
and aggregating it. By the way, if there is a case of sepsis or an infection and it shows up in one patient, you would want to monitor the entire patient population. How does that start to get measured? For me, the next turn of the crank for this will be for the folks in the medical device industry to have their implants, their devices, talking to something like SickBay.

Alan Cohen 22:15
We are quantifying ourselves as a society, so there are lots of really interesting opportunities, and then ultimately to create foundation models and digital twins—the ability to create population-level, regional, sectional, and individual data,

Alan Cohen 22:36
data sets, and make them alive for that. I have a really close friend. We met 40 years ago at a running club, and she's a lead epidemiologist at the CDC. I called her last week, and she said, "Alan, I've been looking at data for 40 years. I'm glad you finally figured it out."

Alan Cohen 23:00
So I think we know this, but we now have the ability to make sense of this data, to find patterns in a way that we haven't, and that, to me, is an extremely powerful opportunity.

Alan Cohen 23:16
Not all of these capabilities have to be for the most specialized. I'm not the youngest person in the room, but I did not live when there were doctors making house calls. If you ask a couple of generations in front of you, and probably until about the 1950s, particularly in smaller towns, but sometimes in cities, doctors would actually come to your house. What did a doctor have in their bag? They had a blood pressure cuff. They obviously had a stethoscope, they had a tongue depressor, they probably had a thermometer, and then they had Advil, and they may have had syringes with some stuff to bring to you, and the doctor came to you. What was very wonderful about that era of medicine is you had a doctor, and they knew a lot about you, and if you didn't move around, they were with you for a large part of your life. I had a doctor for 20 years. He retired. It was heart-wrenching. He was an old-school doctor in my small town in California, and not only did he know me, he was the general contractor of everything in my health. When I messed up my meniscus, he was on the phone referring me to the orthopedic surgeon, right? He managed all of that. So on some level, he was the data lake, the data store for me because I didn't know anything about that. I think we're going to start to see this collection of data. I did see some folks that are, whether it's from wearables, whether it is from opening up the systems of these data to actually give the point-of-care practitioner an enormous ability to make decisions and work with you on your health based on comprehensive data.

Alan Cohen 25:00
What's the worst thing that can happen to you in your healthcare? You move cross-country, and none of your records, or frankly, the tribal knowledge moves with you. Enormous opportunity to improve health and health outcomes.

Alan Cohen 25:15
I want to share two companies with you from an investment point of view. I wanted to share, I think some may know a bit about Proprio. Proprio is a company that, interestingly, is a medical device company. They built a navigation and alignment system for surgery. But that's not why I invested in them. I'll give you a clue on how you can hook a tech investor. When I went to visit Proprio

Alan Cohen 25:45
in the fall of 2019, by the way, I made the investment during COVID and wired the money without visiting the company again. It was an interesting time as an investor, and we all know it was an interesting time for the industry. I saw something there that really struck me. I saw the most interesting and balanced team that I've ever invested in. There were real machine learning scientists; one of the founders of the Lytro camera and key architects of it was there. There were medical device professionals. There was a prominent PhD-level neurosurgeon, Sam Braude, as the co-founder. There was an energetic tech entrepreneur, Gabe Jones. What I saw were the fusions of disciplines that I thought could build a new platform in this industry. What's most interesting to me about Proprio is that it was not founded to be a medical device company. It was spun out of the sensor lab at the University of Washington. Jim Younghurst, one of the co-founders of the company, was a young PhD student working on light field technology. For those who are not familiar, light field is RGB. It is a video camera technology that is super intense in the amount of data it takes in. It does two things: it measures the intensity of light, and it measures the directionality, which would be pretty useful if you were a surgeon, and it allows you to create this 3D image. We discovered over the last couple of years, and what we're really working on is they're effectively building a digital twin for surgery, initially in spine.

Alan Cohen 27:35
They just had an announcement where they finished over 50 surgeries in the first half of last year. We collected 500 gigabits of data per surgery, which is all indexed, searchable, and referable. They are building a foundation model and a proprietary data set. Ultimately, I think that we will have to find that I'm an investor; I'm an optimist, right? I have to be.

Alan Cohen 28:02
But that's actually what we're betting on, and that's why DCVC invested in them. We invest in it because of the data opportunity. If you told me, in five years from now, they're actually not making equipment, but they're actually providing those models and data to other people, I'll be okay with it.

Alan Cohen 28:20
So a data-first mentality is really important, but no one in the sensor lab and Jim Younghurst were thinking about surgery when they put their time into this effort. So that's a key finding for that. The second one, this company is in stealth, but I'll share with you we invested in a company called Remedy Robotics. This is an authentic

Alan Cohen 28:43
autonomous catheterization robot. It does the high enough definition; you guys are smart; you can figure that out. David Bell, who is a cardiac surgeon who worked in a cath lab, took inspiration from this fairly outrageous entrepreneur called Elon Musk and became fascinated with autonomous vehicle driving. Now, I'm not a cardiac surgeon. Actually, my wife is a cardiac nurse practitioner. She worked in a cath lab. I don't know if you want your cardiac surgeon to think about running in full self-driving mode, but that's effectively what this is going to do. It does two things: one, it focuses on how to remove or reduce some of the challenges that go on in catheterization, right? It is not always a smooth procedure, and also the ability to deliver it autonomously over distance. It's important for stroke, but both what it is doing is taking expertise, and it is pushing it out to the periphery of the neural network. In this case, it's a catheterization neural network. So super interesting.

Alan Cohen 29:56
This clearly was a medical professional, but

Alan Cohen 30:00
he changed his entire life because what he saw in autonomous driving is really, really inspiring. These are just two quick examples of how we think about that. So if you're going to build a techmed company, there are a couple of things that are worth really understanding. We believe that techmed companies start—and by the way, many of your companies, I believe, already have these elements. They start from a proprietary data set. You know more about a certain medical challenge than anyone else. You have collected the data, and you've put it to work. That data, which in that model, whether you're using the transformers to build it, or you're just pulling it, can actually impact human technique—how a medical professional does their job while they are doing it. Obviously, it's easy to visualize that in surgery, but if I go back to my doctor's visit, he pulls out my chart, which I haven't seen him or her in a year. They ask me some questions, they touch me, they take my pulse, they do things, they make a decision, and then they prescribe care or give you advice.

Alan Cohen 31:13
What would change if more of this information was available to somebody who's really smart about making those decisions with you? So we see that happening, that post-procedure or post-interaction

Alan Cohen 31:32
is augmented in real time by that. What does it do? What comes out of the other side of that enhances the data set and the model on the other side. That's the virtuous circle. This is how you put data to work in techmed. I think it is fundamentally the most interesting technology opportunity for this entire industry.

Alan Cohen 31:53
But it means you actually have to change your perspective on where data is and how it's used.

Alan Cohen 31:59
Ultimately, like everything else in medicine, there are really four goals of doing this. One, to make the delivery of care more actionable, more precise, and more accurate. Obviously, in surgery, that's a pretty easy one to spell out. Number two is to push down the cost curve. If we know one thing from Moore's Law in the technology industry, it is that every great technology pushes down cost rapidly as it gets adopted. That has another aspect of the virtuous cycle: it makes it available to more people.

Alan Cohen 32:39
Ultimately, you can have one or two models; you can have something that's very expensive that few people can afford and will buy,

Alan Cohen 32:48
or you can have something that's very compelling, and you can make it affordable and let more people derive that. I believe that 90% of the opportunity in front of this industry are people that are pretty much unaddressed today,

Alan Cohen 33:05
and we have a giant growing middle class around the world that will expect the same great care that the people we all work with are delivering today, but they want it to happen in parts of the world where they don't have it. So they're not going to do that on the CMS reimbursement program. It's going to have to change.

Alan Cohen 33:28
You increase that availability. Then healthcare becomes integrated; it fuses through the human body and through our experience.

Alan Cohen 33:39
A couple of things that I'll wrap up: I coined this term, and I call it a data river. If you think about a lot of data science, people talk about data lakes, and they talk about

Alan Cohen 33:51
which are basically large pools of unstructured data, and they talk about data houses, which are structured data.

Alan Cohen 34:00
The power of the computing technology to actually draw influence from unstructured data is changing that game. Obviously, machine learning is a huge contributor to that. But the thing, the fallacy of both of those, particularly in healthcare, is that it is a static element. We are continuously—and I saw some great technology around glucose monitoring. There was a great company from San Diego I met with the other day. We have the ability to infuse our view of our health by constantly pulling data from aspects of that and correlating across that. I call that a data river. How do rivers start? A lot of them start with mountains. There are little streams, capillaries; they feed, and eventually you get something like the Mississippi or the Amazon or the Rhône River and other really important rivers around the world. So this

Alan Cohen 35:00
is the change in how to think about data—not a static set, but something that continuously moves all through your life.

Alan Cohen 35:08
For me, this is really simple. If you're going to build a techmed company, build with a data-first mentality.

Alan Cohen 35:20
There is a lot of sanity to meeting the customers where they are, meaning not—and a lot of logic on not changing how they do procedures. But I think over time, we're going to redesign how procedures, how operating rooms, how we deliver care based on what that data does and how it is utilized. I think that's coming; it's not coming tomorrow, but over the next five or ten years, I think we will start to think about that.

Alan Cohen 35:49
I see a lot of robotics, particularly in things like, you know, people's eyes. When robots are doing the work, how you do the prep room, how they go through the procedure, how they do the post; they're going to change. You're going to have leverage in different points and kind of intelligence and centralized.

Alan Cohen 36:08
You have to engage the regulators.

Alan Cohen 36:12
I think the most successful companies are going to be spending a lot of time educating the regulators about the change that is going on in the industry and working with them. I'm happy that we have an FDA. I don't want to be a guinea pig,

Alan Cohen 36:26
but they are running as hard as they can to keep up with the AI movement in healthcare. Finally, something that's a little bit different: you're going to have to

Alan Cohen 36:39
build with cybersecurity in mind. We are infusing computing systems and data into the medical industry. Well, that has—we've seen in the rest of our society—if you don't take cybersecurity seriously. So that's a bit of like the downer point here. But you actually have to think about that when you're building out what you do. I think that we're already in this movement. I'm going to steal from William Gibson, and people

Alan Cohen 37:10
know Neuromancer and his books: the techmed movement is already here. It's just not equally distributed. Thank you. I'll be happy to take any questions you

Alan Cohen 37:26
It's Friday morning, and I stunned you into silence on the third day.

Alan Cohen 37:31
Yes, sir, the way that you and your portfolio companies are thinking about that flywheel of data capture enhancement that powers the entire thing may be a little advanced for how some of the larger medical device companies, the strategics in this space, think about embedding AI into their existing businesses, adding AI features to legacy hardware products to sell more of that hardware product.

Alan Cohen 37:54
The question has kind of two parts: one, those large medical device companies are likely still one of the primary paths to liquidity and scale for these companies that you're investing in. How do you think about engagement with them and how they'll respond to this type of novel thesis? Where are they now, and where do they need to be? The second follow-up to that is, do you think that with a new movement or speciation event in med tech like this, we'll start to see more pathways for these companies to achieve scale outside of the traditional acquisition by the large company, that we could see more new larger techmed companies kind of come into the field of what we think of as strategics? Thank you. Thank you, Henry, for the most loaded question humanly possible.

Alan Cohen 38:39
Well, I think there are some people—the panel before me, I see Elizabeth here—actually, I think could provide a lot of insight on how strategics think about things. I think we're closer to the computing industry in techmed than we are to the med tech industry. When the PC came out,

Alan Cohen 39:01
interestingly, it was pioneered by IBM, but they built it across the country from their main headquarters. The people who built mainframes and mini-computers didn't really react.

Alan Cohen 39:13
Ultimately,

Alan Cohen 39:15
the big winners in that cycle were not the computing companies. They were really Intel and Microsoft, and they created kind of the forces there.

Alan Cohen 39:25
My company, Cisco Systems, which is the leader of network pharma company,

Alan Cohen 39:31
did it through rapid acquisition and a very agile mindset.

Alan Cohen 39:37
I don't think a lot of people, particularly publicly traded companies, are going to change their business model overnight because of an incumbent—sorry, as an insurgent with a different strategy. I think it's going to take a couple of the techmed companies going public, becoming compelling players, right? Effectively, alphas in

Alan Cohen 40:00
the ecosystem before we see fundamental change because it's really hard to explain to your shareholders why you want to slow down an existing product line and change it as you go. We would not have even the large language models. We would not have this entire change in generative AI if it wasn't for OpenAI

Alan Cohen 40:20
pushing Google, Microsoft, and Amazon to begin the change. So I'm not sanguine that the large companies—they will do things; they're very smart, and they're very well-run companies. I have huge respect for them; that is the issue. I mean, we saw, you know, there were a lot of surgery companies 25 years ago when Intuitive started, and people were like, "Well, let's see how that robot goes," right? I think the techmed industry is going to need an Intuitive Robotics—I'm hopefully investing in it. I'm about to find out in the next couple of years, but I think that is going to have to change,

Alan Cohen 41:00
or they're going to have to do what IBM did, which is empower an executive to create a division away from headquarters that will work as a skunk works to get there. Collaboration with startups—it's not that's not going to happen. But you know, nobody likes change. Nobody makes change unless they have to. You know, if somebody says to you, "You should exercise more, you should sleep better, you should do that," and yeah, you're right. Then if your doctor says you're going to have a heart attack or you had one, all of a sudden, come on, bring the GLP-1; I'm ready to go. I'm in the gym all day. So people react to change differently. I'm not giving you exactly the answer. I think this industry,

Alan Cohen 41:45
the business model has to change as well as the technology has to change. If you're selling,

Alan Cohen 41:52
if you're selling razors and razor blades, and somebody wants to go after your blade business, that's a really tough business to make that transition. Maybe you could make the blade smarter. Remember, razor blades had one blade, then two, and now there's five. But it's, you know, people haven't fundamentally changed how that is. I think the industry structure has to change as well. Time for one more question. Well, that was exactly where I was going to go with this next question. Thank you for giving us the architecture, the framework on how to think about this. When you look at it on the business model side, I've seen kind of two different approaches right now. One being, we're going to solve something we get paid for, and the data is on the back end, and the other being, the data is the value proposition that drives what the customer is paying for. As you think about driving down cost, and you think about greater accessibility globally, how do you think about that in terms of creativity in the business model as well?

Alan Cohen 42:51
I actually think you answered the question right, which is that you actually, if you have a data-first mentality, it actually opens up how you might develop products against it, right? It's those key insights.

Alan Cohen 43:07
You know, I had breakfast this morning with some folks, and they actually made me really think about this. There are three legs to the stool on how you innovate in this industry. There's this great technological element, right? There are the developers, there are the engineers, there is the data side of things. But I think the companies that are going to change the business model are also going to be very clinician-oriented. So you can't do this without clinicians. I think you have to kind of—if you're—I spent a lot of my career in marketing. I always think about who's going to use this and who's going to pay for it, and you have to hook them first in your design of a product and business model. If you can do that, they will pull you through on the business model. The great thing we've had with Proprio is that we have 20 amazing surgeons on the Medical Advisory Board. So I like to say it was built by surgeons for surgeons with surgeons, right? It's not like we're going to drop this in front of them and go, "What is this? I don't know how to use this." My company, Evolve,

Alan Cohen 44:15
is a company that sells a $120,000

Alan Cohen 44:18
weapons detection system. It's a subscription business. You pay by the month. It is 15 times more expensive than a metal detector, and price doesn't bother because it actually allows 10 to 20 times more customers to pass through without emptying their pockets. If you provide a condition for the users of these technologies that is so compelling, they will help pull you through on the business model. There are ways to finance, ways to do this. The biggest challenge, I think, with a lot of large-scale medical devices, it's like, "Hey, it's a million bucks." That's a very difficult opening conversation.

Alan Cohen 44:57
But if you can show—and I always

Alan Cohen 45:00
mine startups—visit the purchasing department early in your engagement with a customer and figure out how they spend money. Hopefully, that gives you a little bit of a flavor. Thank you very much. Thank you.

Alan Cohen 45:11
Thank you, Alan. I think we're coming up on time, and I know there are many more questions, so you may get hit on the way out here. Thank you very much. I appreciate everybody's time.