Nikki Vincent:
It's my great pleasure to introduce my friend Craig Siiro. He's an ACG member, board member, chair of our programming committee, and president and CEO of Integrated Consulting Services. So Craig, please take it away.
Craig Siiro:
It's my pleasure to introduce our speaker today. It's always interesting, these events, learning new things and I'm really looking forward to this presentation. So, I'll introduce Bill Studebaker, he's the president and CIO of ROBO Global. is an index advisory and research company wholly focused on helping investors capture the unique opportunities of fast-growing robotics, artificial intelligence, and healthcare technology companies around the world. So I am looking forward to the conversation. Take it away, Bill.
Bill Studebaker:
Thank you Craig, and Nikki for inviting me back. For those that may remember I had an opportunity to speak to you all roughly about two years ago. We talked about the innovations in robotics and AI and how those applications were going to begin to change the world as we live and work, fast forward a couple years later, we couldn't be more convicted by the changes that we're seeing and importantly we'd like to talk about the advances and the innovation that we're seeing in healthcare that's being driven by AI and it's a truly exciting time to be an investor, to be a patient, a consumer, because we're going to move into a world of prediction and prevention and individualizing medicine. I'm fortunate to have my esteemed colleague Dr. Manish Kothari on the call with me. Manish is a strategic advisor at ROBO Global and a shareholder and importantly Manish is the head of ventures at the Stanford Research Institute Ventures Group, where he has a front row seat to what's happening in innovation. For those that may not know and Manish can talk about it but SRI has had some amazing incubations, they created the first satellite TV dish, Siri, the brains behind Intuitive Surgical’s, DaVinci, the internet just to name a few things, the mouse ... So it's pretty amazing to have Manish's perspective. He's one of 1,800 PhDs at SRI that again are really on the frontlines of seeing where the innovation is occurring.
Again, what's so exciting about healthcare is that we're really on the front ends of one of the greatest technological shifts and it's being driven by the performance capabilities of computing which are essentially doubling every 18 months and our costs of computing are plummeting. This is now creating an array of use cases that a few years ago were only Elon Musk science fiction and healthcare is certainly ripe for disruption. Healthcare as we know it isn't healthcare, it's sick care. We're moving to a world of prediction, prevention and essentially what's going to happen is that we're going to be lowering costs, improving patient care, and essentially extending human longevity. So kids that are born now are not going to live until they're 70 or 80, try 100, try 120+ as we have the ability not just to arrest the signs of aging but to begin to reverse it.
Just a little bit about who we are and why we created ROBO, Craig touched on it briefly for those that don't know ROBO Global. Again we are a research and advisory company focused on robotics and AI in healthcare technologies. Feel free to look at our website which is ROBO Global.com to get more details on our team and our focus, but we created the firm over eight years ago with the belief that we were entering a period of ubiquitous automation and at the time there was no way to play this theme, let alone to actually think that this might be a good idea. At the time we were one of only seven thematic strategies or exposures in the world and now we've got close to a couple hundred. But from the early days, we always articulated the importance of kind of having the longer term lens when looking at disruptive innovation because the green light ... It doesn't always go on or you don't really know when the coast is clear to go and buy.
In fact, innovation falls what's called the J-curve effect, where the performance often falls at the beginning and then rises gradually to a point higher than the starting point forming the letter J. This phenomenon applies to a variety of areas captured by innovation. It requires management discipline and commitment to stay the course because it reflects a period typically of unfavorable returns that's followed by a period of gradual recovery that rises to a higher point than the starting point.
We're seeing that everywhere from ... You've seen it from Tesla to Intuitive Surgical, there's Surgical Robotics to iRobot with the cost and penetration and in home cleaning to sequencing the first human genome. So to get it right requires I think the discipline to understand where the world is going and to stay the course and investing in disruptive technology is obviously very difficult because everyone wants to find that works. But it's difficult to be there because investors typically sell what works, they sell what doesn't work and they're stuck with what's in between which typically is not that glamorous.
So we've put together a team of industry, experts and financial experts to help investors capitalize on this. Briefly, our strategy, not to tell you too much because that's not the focus here, we want to talk about disruptions in healthcare, but our strategy for those that they may care is research-driven where we try to identify the companies that are best in class in robotics AI, in healthcare technologies. We tend to focus on companies that are more small or mid-cap because we think that's where the innovation is happening, not to mention they tend to be in the catbird's seat of where the M&A is following. We have been fortunate since we launched our robotics index and ETF ... Eight years ago we've had upwards of about a third of the index has been acquired and that just speaks to the innovation that's occurring. These strategies have very ... Excuse me, low overlap with traditional indices and so we think this is interesting for investors to consider.
The next slide just briefly identifies our three indices that we've created. The first was our robotics and AI index which we did eight years ago. We essentially created the first classification system, think of it as the NASDAQ if you will for robotics and AI. We had to create our own proprietary sectors to identify these companies as either a technology or an application so you can see on the top we have the technologies which are things like autonomous systems and sensing and actuation and down below we have the use cases, where they're being applied. So in this case into manufacturing, into healthcare, into food and ag as an example.
We then also launched two other indices, our AI index which is Think which capitalizes also on the technologies and applications within that area and obviously we have our healthcare innovation index that looks to capture the future of healthcare and we'll talk more about this.
Our team is ... Nothing better than sensational. Manish is one of our esteemed PhDs on the team. He is surrounded by a team of industry experts. Everyone from in the bottom left hand corner Raff D'Andrea who co-founded Kiva Systems. That's Amazon Robotics. He and his two partners were working on warehouse algorithms back in 2005 and in 2011 and sold that to Amazon for $800 million and that's started what we see as the robotics arms race in eCommerce. We have Daniela Rus who's the head of AI at MIT. We have Henrik Christensen who is kind of the godfather of robotics who has been in the industry for the better part of four decades doing research in the area. You can look at our website for more details but certainly this is what I think makes us quite differentiated.
Moving on to just the discussion of healthcare and investing and innovation, the world of healthcare again is ... I talked about going to the world of prediction, prevention, and individualizing medicine, and HTEC is designed to capture and invest in the future of healthcare. This a technology-driven revolution in healthcare and it all starts with going on ... Moving to early detection. All solid tumors fall a pretty predictable path from local and treatable to metastatic to lethal, so this is providing the rationale for early detection and as prices drop below $1,000.00 nearly all age groups above 40 will be screened for cancer cost-effectively, potentially saving up to 1.4 million lives in the U.S. alone and the convergence between among innovative technologies has pushed the cost of multi-cancer screening down twenty-fold from $30,000.00 in 2015 to $1,500.00 today and is expected to fall precipitously in the years ahead.
So again, the early screening, the world of prediction prevention, is a direction we're going. Manish will talk a little more about this in his remarks.
Moving onto the next slide, you can see that the healthcare innovation has seen a lot of milestones. The seeds were actually planted years ago. Costs were way too high and they still are high and much of the early growth was very linear but we're moving to a world now of exponential growth.
The first DNA was sequenced as you can see in 2003, it was sequenced for $2.7 billion which required 13 years of computing power. It now costs $500.00 and we're going to $100.00 to then $10.00 and once you understand mutations, you can understand disease and then cure it. So that's where we're going and that's why we really are so excited about this.
As the next slide points out, what's super exciting about healthcare is it is one of the least digitized areas of all sectors of the economy. There isn't an area of healthcare that is not ripe for disruption. The costs are too high, having a bad or misdiagnosis, the list goes on and on, and so the opportunity for healthcare is going to go on for years if not decades ahead of us, and you can see here that innovation is going to be needed to solve the problems that we have, whether it's aging population or whether its rising cost, again healthcare is 20% of our GDP. That's simply unsustainable and it must change. We have a shortage of skilled workers. If you look at places like ... In the U.S. I think there are roughly 2.4 doctors per 10,000 people. In China it's more of the order of 1.4 doctors per 10,000 people. As you move into more acute areas, it's even more intense, I believe there's roughly 20,000 oncologists in all of China to treat a population of 1.2 billion. The only way to solve this is with automation, AI. A medical error or misdiagnosis is actually the third leading cause of death so again, a lot of areas are very, very ripe for disruption.
As you can see, the future of healthcare is going to be profoundly changed as we're going to a world that has become much more digitized, as the previous slide had outlined ... We're at low single digit penetration rates for digitization but that's going to change. We're going to a world that's become much more decentralized, that's going to result in better patient care and lower cost. We're going to a world of prediction and prevention and that's the direction we're going, we're super excited about this and I think the opportunities are endless.
As the next slide highlights, healthcare has seen a tremendous amount of innovation. It's in the fast lane which is why it is ripe for disruption and money continues to pour in. As you can see the number of deals and the funding has continued to increase. Here you're seeing companies actively develop everything from AI solutions for clinical trials to AI for remote monitoring to machine learning for electronic health record processing to cybersecurity or data protection.
CB Insights who also put the data together for this announced their fifth annual list of the 100 most promising AI companies in the world and among them were actually eight healthcare companies with solutions for everything from drug R&D to clinical trials to surgical intelligence among others. So this continues to be again kind of fast lane in terms of where the innovation is going and as the next slide points out, the opportunities for investors to invest here is certainly evident. You can see the returns of our healthcare innovation index have far outpaced that of global indices. While nothing does go straight up we have seen a little bit of a pause in the performance this year. The index is up about 3% lagging the market ever so slightly after being up roughly around 68% last year, but again we think we're on the frontlines of a tremendous amount of disruption ahead of us.
So with that being said, I'd like to pass it over to Manish and he can give you some insights on the disruptions and the excitement that he's seeing at SRI.
Manish Kothari:
Thanks a lot Bill. So maybe I'll just start by a quick introduction to SRI. We've been around for 75 years. As Bill said, things like the computer mouse, the first internet message, ultrasound for medical diagnostics, intuitive surgical, Siri Nuance which was recently in the news was created by us early back 20 years ago or so, so we've been around for a long time. We've done some great things and really excited to be here today.
Before I go to the next slide, I just want to talk about two personal stories that sort of reflect what I'll be talking about in the next few slides.
The first one is about my mom. So my mom passed away a few years ago with uterine cancer. When she first got diagnosed over 10 years before that, she had the usual cocktail of therapies, chemotherapies, massive side effects, but very successful. Eight years go by, it stopped working, and my dad in his desperation tries treatment after treatment after treatment, most of them with horrible side effects and with no outcome that worked. That was 2015, '16, '17. So just a few years ago, three or four years ago.
Today, the situation would be very, very different. For exactly the same person, my mom would have a biopsy of one of her tumors. We would do a full sequencing of the tumor. We would actually in vitro test whether any new drug therapy would work against that tumor and only give that one to her. What a different world from even four years ago. All the pain she had to go through trying 10 different therapies would not occur today. Roche provides that for example, and there are others too. Genomic Health and others. So this is just a reflection of the profound changes that have happened in the last five years from sequencing.
The second story is on a totally different note and I will have a video about this later on too. So my grandfather, who also did just pass away recently at the age of 99, he needed a walker and when I would go to see my grandfather the first thing he would do, the very first thing he would do, was try and chuck his walker behind the couch. He didn't want me to see him with a walker. He wanted me to see him as he was. This is an important story and lesson. It's not just about being healthier, it's about mobility, it's about dignity, it's about many more things than health. But the underpinnings of what we're talking about here today are going to be what lead to those changes, both in terms of longevity or agelessness as Bill had described it, as well as quality of life when you are living those lives. Both are equally important, and we're in a great time.
Last comment before the next slide or we can get to the next slide, my background is ... I started my career way back when 30 years ago building hips, knees, and spines. Then I built bigger things like MRI scanners, PET scanners and then I fell in love with the software world because I saw how the software world was changing the world and we did ... The first telemedicine company that was probably sold was one that I founded, and I've been around the healthcare area ever since, focusing a lot on software amongst other things.
So, let's just say ... There are really a few things we want and let's talk about it. We want increased effectiveness of the drugs and techniques, therapies we are using. We want increased convenience and I think Bill described it beautifully earlier. We want it with fewer side effects and to get that we need two things. We need faster innovation because we don't want to wait another 10 years before there is a drug for Alzheimer's. None of us do. And we want true personalization. In the ideal world, we would like a drug to be created that is uniquely applicable to our biochemistry and our lifestyle.
So, these are the extremes and what I'd like to say is they are not that far away. We are talking about true personalization starting in a few years from now and I don't mean 15, I mean a few years, with broad true personalization happening within a decade. I'm going to give some examples of faster innovation which enables that. I think let's step back. First there's really complicated ... Healthcare is not easy. Here is an example of how it's not easy. You have very complex biochemistry defense. In the blood, in the brain for example, you have a thing called the blood-brain barrier. It's a barrier that's designed to allow nutrients to pass through to the brain but no dangerous things such as viruses, bacteria, which are small, and other elements. This same barrier causes a tremendous difficulty when you're trying to treat diseases such as Alzheimer's or even metastatic brain tumors. The same barrier doesn't allow drugs to be transmitted through. So, the very things that our body has been designed to protect us is causing some challenges in how we do it, and I'm going to talk about how these are being resolved now, if you go to the next slide.
Another thing that's a challenge, so I'm trying to highlight that there are some challenges ... If this were easy it would be done already. In the world of windmills and physical objects, over the last four or five years, ROBO has really captured this trend beautifully with its ROBO stock indicator. We have started making digital twins. We have actually started making very complex digital twins. We can actually even model things like friction and lubrication which are not easy to model at all to digital twin. Once you've made a digital twin, you can do a lot of things. You can run a lot of simulations on software, reducing your cost. You don't need to make 50 windmills of different sizes to see which works. You run it on software.
In the past a digital twin was not good enough, but with the combination of better sensing and of AI, we can do that. In the human body, making digital twins is still hard. We don't have a full sense of what the digital body is like. We don't have sensing capabilities to know how much potassium is being used by our muscles as we exercise or our nerves. So these are all elements that make it challenging. If you go to the next slide.
There's other challenges too. So I don't mean this to be a damper, I know it's lunch but I'm going to start by just talking about the challenges and then move over to the solutions here. We all have heard this, right? Highly siloed data and not just that it's highly siloed, it's really poor quality. I mean you've heard again and again physicians are tired of what the electronic health records do so they fill it full of gibberish and they actually keep the notes that are relevant on a piece of paper on the side of what they do. So you've got a lot of data but it's siloed, it's poor quality, and it's not ... It's incomplete. How do you create, take full advantage of the AI revolution when you're in this circumstance. Next slide.
So coming back, let's ... I just sort of highlighted some of the real challenges that we're facing right now and I'm going to talk about some of the changes that are happening that we'll go through this. If you go to the next slide. So first thing, I mean you've heard the term data is the new oil. It's a partially correct answer. Data, another better way to think of it is data is the new sand. Sand is actually used for a lot of things. Sand requires a lot of processing but when you process sand you can make things like glass. You can make things like ceramics. You can make things that form the underpinning of your phone. Data needs to be processed to do it. I just hear a comment that data is the new bacon, is that what I see down there? Sure. Everybody loves bacon, so that's an easy, easy one.
So the first big change that's happening is there is a lot of synthetic data being created. Let's talk about COVID for example. You had the situation where every hospital had its own data sets on COVID, the response, what therapies, how it was working, but because of privacy issues, nobody was willing to share that data. This is a real challenge. On one hand you have maximum privacy but zero utility. On the other side you have maximum utility but zero privacy. That is not an acceptable answer. What's happened in the last year was a combination of new AI algorithms allow us to create perfect anonymized synthetic data. This is a very hard problem because what you do when you create synthetic data is you may lose sometimes the small signals that are in there that you're really trying to discern. This is possible to do now.
Think about Biobank, which is the U.K. genomic database and All-in-One which is the one in the U.S. Those data sets contains tens of thousands of deep genomic information, but they've never been able to be exchanged with each other. This is possible now. COVID was one of the drivers that transformed the privacy issues and the creation of synthetic data and data aggregation. Next slide.
Data virtualization, now this sounds like a fairly boring subject but I think we all know ... Look what happened in the IT world as virtual machines were created. I mean AWS, Google, Microsoft, a bunch of others, from Akamai and others, have built massive systems based on virtual machines or machine virtualization. VMware's hole existent is on machine virtualization which is creating ... Instead of each person having one server, one server hosts many virtual machines that are distinct from each other, allowing you to scale efficiently. We have had challenges on the data side. I think you can listen to big companies even Fortune 100 companies that will say that they've spent the last five years of their AI journey creating data lakes, creating data marts, loading things into systems. This makes no sense. What you're doing is you're reducing, not increasing the flexibility of companies to be nimble to changing circumstances. By creating this complex systems with flow-throughs and cleaning, that doesn't work. This next generation of startups that are coming out over the last two years have really focused on data virtualization or creating virtual machines but for data. So in other words, you don't have to spend billions of dollars creating data lakes that all these companies will have to do. You instead say if there's a new sensor, it's spitting out new data, I'm just going to virtualize it and incorporate it the same way I deal with virtual machines.
This is a fairly complex technical problem, but I would like to point out that this is actually one of the underpinnings, this and the previous slide, are two of the underpinnings that are actually going to transform healthcare in the next 12 to 24 months. All the issues you've been hearing about siloed data, poor quality data, these are being resolved with these two fundamental technology underpinnings. Next slide.
We talked about digital twins, so I talked about the data side of things. What we also have now is exquisite sensing. I'm going to show a product that's come out into the marketplace just in the last couple of months. The thing here, the story of my grandfather is one where we go, "I don't want to wear an exoskeleton." If I'm not willing to put a walker on, I'm not going to be willing to put an exoskeleton on, but I'd like to create something that enables me to wear something that is clothing, soft, wearable. I could wear it under my clothes or over my clothes or just by itself. That allows me to increase mobility. To do that I have to create a personalized digital twin first to understand how I walk. Do I have arthritis, does it change things. All of this can be done very, very easily with off the shelf sensors today, and if you play the next movie, you'll see a product that's coming out ... It's actually just come out.
You get a sense here, right, that this is the change, so this is ... This would not be possible without advanced sensing, advanced and without AI because it is ... How do you process all of those signals ultra-fast, come up with a response that is there. It wouldn’t be possible, well wearables like the watch, which allows a person to modulate how they want to feel when they use this. Do they want more stiffness today or less stiffness today? Do they need help sitting or standing or just standing while they're cooking? These are all possible now and these are going to change the way our future are.
Let's talk about another area. So I think we've all heard and we all know that the time for clinical studies is 15 to 20 years from the creation of a ... The identification of a possible target you can impact to change something to the creation of lead compounds to the creation of drugs and then the testing of drugs. It's a 15 to 20-year journey. Now this is something, and the costs have gone up from hundreds of millions to billions and billions of dollars, so it's both very expensive and takes too long. I don't think that's a satisfying solution and that will not enable us to get to that extreme personalization that I discussed.
First of all, AI itself is transforming this. Today AI can ... One of the big challenges in these processes is finding out medicinal chemistry. Trying 15 to 20 drugs, finding which ones have potential issues with heart toxicity, with other types of toxicity. All of this now can be done via AI. You can actually take existing compounds, train them, create a model, decide what you want your outputs to be, and create new chemical entities. So already there are over dozens and dozens of companies and three or four very strong ones that are starting to create more lead compounds to go after targets.
Now you're running into the next challenge. Okay, we've created these designs. How do you make them? If you go to the next slide and just play the video, this is around a minute and a bit long.
So as this says this was funded by DARPA so what is DARPA trying to do here? DARPA is trying to say in the ideal world once I have a target chemical process in the body that I want to influence, can I create leads via AI so the human in the loop but via AI, create potential target compounds, go out and create them, test them, see the delta from what we want versus what it does. Input that back into the AI system, closed loop create new compounds, and in the course of two months which would normally take two years find the lead compound to go after.
DARPA has invested around $40 million into this at this point of research. We already have systems up and running. We recently won another $100 million grant to say can we make kilograms instead of milligrams. The answer of this is beautiful, it's just like a parallel processing on computers, we just create many of them because they're cheap and they just go one after the other and we can create ... Now when you talk about extreme personalization, this is what you need. If you want to think back about how Novartis or Roche, today Novartis or Roche makes hundreds of kilograms of their medicine at any given time. This has a couple of real problems. One, if you're making the same thing hundreds of kilograms, you're going to sell hundreds of kilograms of that. It's very hard to do extreme personalization. Two, most of these reactions are exothermic meaning they release energy. You can't create these things by doing continuous flow, you can make very small quantities that don't release energy, exothermic reactions or so much energy so therefore you don't need complicated cooling, environmentally degrading systems. You can do things much better for the environment as well.
These new methods are what we're looking at. Ely Lilly has set up a massive center in San Diego that does exactly this. So this is the future of drug innovation on the chemistry side and actually drug manufacturing on the other side which is a minimum requirement to then sell at a more personalized, create and sell at a more personalized level. Go to the next slide.
Then the last thing, you can just play the video here and then I'll talk about it. Volume would probably need to be up.
So the point I'm making here is that when we talk about speech recognition, we think about Siri, we think about Google, we think about Amazon. We realize that most of the portion is thinking about words. They're trying to catch our words, but when we think about how we interact in real life, most of our reaction is actually through the tones. We use our voices as gesturing agents or explaining what we have or not. This is a concept of think of it as emotional intelligence from voice. There are systems now that can do this extremely well and in the beginning pre-AI days, people were trying to do sentiment analysis, so they were trying to classify are you happy, sad, or otherwise. There's lots of data now that shows that it's actually very hard to classify if somebody's happy or sad or otherwise. These are very difficult things to do, everybody's different. However, using neural nets, you can start classifying certain responses and behavior.
On the right what I have is this concept of GPT3. Some of you may have heard about this. GPT3 is really a breakthrough in natural language creation. So GPT3 is a very large dataset that was created by OpenAI, a group founded by Elon Musk and a few others, that now enables you to create synthetic language that is almost indiscernible from natural language.
Why is this again relevant? The biggest challenges ... In healthcare, why is it relevant in healthcare, it's relevant in many areas. The biggest challenges in healthcare is a mismatch between the patient communication ability and the physician communication ability. The patient is inherently lack of knowledge, nervous, worried. The physician on the other hand knows too much. They're trying not to say all the possible outcomes, many of which have low probabilities. The ability to use these techniques, GPT3 and detection of tone actually is changing telemedicine and in the course of the next two to three years, you are going to see digital front doors to almost all physician capabilities that are very smart. The simplest way to think of the difference is you're going to have a virtual medical resident rather than a virtual scribe or a person receptionist. A medical resident is able to really help you and then explain what the results are. You could go to the next slide.
I'm going to end with this slide. This is a paper that came out I think last week and I think it's a really, really important paper. All of you have heart about CRISPR9 and Cas9, I'm sure you've had conversations about it. This is the first study that allows people to program epigenetic memories as opposed to breaking DNA strands and inserting something new, this can temporarily switch on and off genes that create epigenetic reactions, so reactions to your current environment. These changes can be heritable, so you can actually ... They can be passed on from gene to gene but they can be turned on and off so you can turn it off again when you need to turn if off. So think about the example where you can turn off the gene that creates tau protein tangles that are considered to be a part of Alzheimer's. You can turn them back on if you want to. So CRISPR on and CRISPR off is a really revolutionary upgrade to CRISPR and what it can do and this combined with some of the other big revolutions happening in CRISPR. CRISPR is being used a lot in lab settings. There's a lot of work going on and some great start-ups that are talking about how we can actually provide CRISPR in vivo in humans.
That combination of those changes and this publication arguably is going to be one of the most transformative things that happens in the next 10 years in terms of therapy. Happy to take questions. I'll pause here. There's a lot more we can talk about. Thank you very much.
Nikki Vincent:
Thank you Manish and thank you Bill. This was definitely exciting and a lot of information to cover in a short time. Craig, I'm going to start to you if you have any questions that have come your way.
Craig Siiro:
One question I have is ... You talked about the speed of innovation. Is there a risk that basically with the speed of innovation, innovations that you are invested in today are going to become obsolete before they actually can become fruitful for the investors. Does that make sense?
Manish Kothari:
So I'll start and then if Bill wants to say ... That's a great question. So yes, that is absolutely true depending on how you choose where the innovations are. So let's give a specific example. Around three years ago, AI platform companies were the rage in the startup world, and I remember looking at dozens of dozens ... I did not start even one and I start 10 to 15 startups a year. Why didn't we start one and it was simply because I was looking at Google, Amazon, and the other established players and saying, "Some of this is going to come out really fast from these established players. It's much better to invest in a chip which is going to be harder to reproduce, Mythic for example, or an edge AI which is going to be much tougher for somebody like Amazon or Google to come up with easy protocols for investment. So those are areas we invest in and so there are some areas where the bigger entities, the bigger large caps, which are leading a lot of innovation are open-sourcing elements and taking a financial loss leader approach. But there are so many white spaces that are there, and in healthcare in particular, because IP is such an important portion, as long as you can corner a very strong IP position, there is ways to actually tell if you've got a solid shot of succeeding.
Bill Studebaker:
I guess the only thing that I would add there is areas that are ripe for disruption are going to bring in a lot of other entrants and this is certainly the case if you look at surgical Robotics back in 2000 or 1998 Intuitive Surgical began selling the DaVinci and they were obviously the first company to market, they're arguably the gold standard in surgical robotics but there are now more than 40 other companies that are working on various other surgical robotics and as Manish talked about, what's kind of different which is a positive and negative is the advent of what's called ROS which stands for robot operating systems. So essentially it is much easier for incumbents ... I'm sorry, entrants to come into a market and start out with some sort of innovation because if you could imagine a surgical robotic activity that maybe was being funded at Stanford 10, 20 years ago, you would typically start, do your research and what would normally happen is you would run out of money and the innovation would essentially collapse, there would be no scale to it, and there was no shared learning, and now with ROS, there's shared learning. So people aren't starting out at the ground level, they're starting out much higher.
Bill Studebaker:
So there is risk of disruption. We kind of spend a lot of time focusing on companies, understanding their technology, understanding their market share and understanding their path and routes to commercialization because there are obviously a lot of great technologies that never make it to market because of poor management, leadership and execution so that's always a risk that you have to be mindful of.
Craig Siiro:
All right, thank you. A question from the audience, how would you view the privacy within the evolution?
Manish Kothari:
Right. So maybe I'll take that one and I'll take that one and then the next one which is under there about the transparency. I think those are ... So I think privacy is really, really critical and privacy is something that we are seeing everywhere and in health care of course it's hyper critical. This is why I've been talking about ... There is two technologies that are really underpinning and reducing the risk on the privacy side. One is this concept of synthetic data that I described, the concept that you can create identical data sets that are not traceable back in any way, shape and form and in COVID these played a very major role in helping come up with decisions quicker.
The other is this very unique technology called partial homomorphic encryption and it's a technical term but think of it this way. If you can ... So one of my starts of joining SRI was I came actually from a non-profit motive where I was seeing a real problem in pediatric care. The trouble in pediatric care is a very rare occurrences, so say pediatric cancer, so even a group like Stanford or Mayo or Harvard may only have 15 to 20 patients of a particular type in their data sets. This makes it really challenging for a doctor to figure out what to do. So what does a doctor do today? They say, "Oh, I had a friend who went to medical school who is now in Harvard and I am in Stanford so let me call that person and see what the case." It's a highly inefficient way to do these things.
Another way is to create a central bank which everybody's data comes partially encrypted in a homomorphic encryption rate so it's an encryption method where you cannot discern the original data but you can do computations on the original data. That's one way that people are doing to get around the privacy issues. I think privacy and healthcare will always require some effort but I think it is being solved faster than you would think, and I'm the sort of person who doesn't get positive unless I really feel that there is some positivity here but I actually think the privacy issue in healthcare is going to go through a year or two of problems where there will be hacks and other things but the technology to solve it is well underway.
The second one, which is a lack of transparency and the ability to explain, is a really big deal. There's two issues here. One is are AI neural nets and deep learning techniques accurate enough for the problems and then can they explain why they came up with their decision, what they did. It's a great question. Over the last 10 years, DARPA has probably funded half a billion dollars worth of research in explainable AI, and these are somewhat oxymoronic. The whole aim of a neural net is to do things without having to explain things, having it do that. Yet at the same time you need a degree of explanation. There are already solutions where what you can ... Think of it the simple way. If you could take all the clinical guidelines or any rules if you're a finance company, take the rules of your finance company, take the document, ingest it as a set of rules, and then from the rules create synthetic data that becomes the backbone of your neural net. You can actually go backward and forward. The neural net first will have constraints in it now. Two you can work backward and compare it to the rules and say, "This met those rules or it didn't meet those rules." So there's a form of explainability that can be achieved as a result of that.
So I think you're going to see explainable AI coming out. It's never going to be perfect. It's never going to do things like rules but it's going to at least be able to explain the rationale behind why it did it to some extent.
Craig Siiro:
Thank you. That's a great explanation. I have ... Maybe to build on the original comments, you talked about living till like ... 120, 100 whatever, I wonder what the economic effects are going to be to that and I just shudder to think my kids are going to have me living with them for 40 extra years after I get to 90, so any comment on like the long-term economic effects of the longevity of life?
Manish Kothari:
Bill, you want to take this one?
Bill Studebaker:
Yeah. Well I think that what we're going through now is again because of the world of prediction and prevention, we're going to cut down the costs of healthcare dramatically. There isn't an area of healthcare where you're not going to see cost reductions. Again, you're seeing it across the board. I mean it takes upwards of 29 days to get into see a medical professional, if you need to, if you can go and solve this with a virtual care appointment it may cost $40.00 to $100.00 for an appointment. You're solving a big constraint which is saving a lot of money. Not to mention as advances in 5G occur, we're going to be able to provide ubiquitous healthcare coverage globally in places that haven't had healthcare. So the cost will continue to climb, you can be a doctor situated on a desert island in the Caribbean and be able to provide healthcare so the opportunities I think are kind of endless. Manish, any thoughts?
Manish Kothari:
Yeah. I think just a couple more comments which is reducing side effects and reducing medical errors actually is going to be a fairly dramatic reduction in the cost as well. So I think you're going to see the cost curve bending. I think you're also going to see the productivity of people increasing over their lifetime because of the improvement in healthcare so you are going to see somewhat of a reduction in the cost and I don't think it's going to happen immediately but it will happen over the next three to five years and you're going to see an increase in the productivity or output as a result of improved health. So those two things.
Now whether you want to hang out with your kids till you're 90, 95, I guess it depends on you and your kids, right?
Craig Siiro:
Right, no, exactly. All right. That's all I have, Nikki. I think we're almost at the end of time so thank you.
Nikki Vincent:
We are and this is perfect. Manish and Bill, thank you both. Thank you, Craig, this was fascinating, and I know there's some more questions so for those of you who can stick around please do so. Manish and Bill will be able to stick around for the next 15 to 20 minutes and answer additional questions and/or just frankly chat and network as you choose. But thank you again Bill, it was lovely having you back.
For those of you who are new to ACG, we are the largest networking group in the world for private equity and we drive growth in the middle market primarily through mergers and acquisitions. So please reach out to me if you're interested in learning more and let's give a big shout-out to our platinum annual sponsors who support this organization along with all of our sponsors to make these virtual and our live events happen. When you're doing business, please think of the following firms. Ballad Spahr, Boulet, Bremer Bank, Frederikson & Byron, Datasite, Growth Operators, Highland Bank, Lathrop GPM, Lurie, MidCountry Bank, Oasis, a Paychex Company, Oxbow Industries, Quazar Capital Corporation, and Allied Executives and Redpath and Company. When you're thinking of doing business, please check these folks out and reach out if you need more information. And by the way, it's not too late to get in on sponsorships so if you want to have your firm name shouted out on virtual events or live from the podium, come and raise your profile with ACG Minnesota, the leaders in middle-market deal-making.