Episode Transcript
[00:00:00] Speaker A: Bill.
[00:00:00] Speaker B: Hari, what an absolute pleasure to have you on the show today. I want to start with each of you sharing why you started this entrepreneurial journey that you are on right now.
[00:00:15] Speaker C: I'll start, Hari. It's a very interesting question. I don't know that I have a very simple answer. From my background and how I grew up, it organically just kind of happened.
This company was 15 years ago, but I had done things, really, my whole life just finding ways to make a living, finding ways to help contribute to my family. And the term entrepreneur is, I believe, overused, but I was always just creative in finding ways to make a living. And the fellows that I grew up with and were mentors of mine used to say, you have to have income and you have to eat, so find a way to make a living. And that was really the genesis of this journey. And it's just evolved, I guess, to where I am today in over 57 years.
[00:00:58] Speaker A: In my case, I would say I'm an accidental entrepreneur.
I'm fundamentally a researcher and teacher and ended up working on problems and gravitating toward problems in research that are driven by application and driven by how to make people's lives better, and also very gravitating towards problems that are just really interesting. And intellectual curiosity is something that I feel I have and I really value in all my colleagues and people I work with. So this came about, this particular company came about because of some early research that I had led at MIT, together with my colleague Sam Madden, who is our co founder, and some of our students. And then I met Bill 15 years ago, and I, we hit it off. So that's really how this came about. But it wasn't a set goal for me in my life to become an entrepreneur. It's quite accidental.
[00:01:57] Speaker B: Tell us about your research. Tell us about the problem.
[00:02:00] Speaker A: Yeah, so I work in the area of network computer systems, and broadly speaking, it's had many different themes, but it's driven by this idea of creating societies that are safer and more resilient than before. And the Internet. My career started with the Internet. The web, the growth of the web in the mid nineties, early nineties to mid nineties. And early on, my research was in enabling wireless networking. And this was at a time when laptops were just coming around. There was no Wi Fi, there were no smartphones, and so there was no cellular network stuff. And the early research I did had a lot of impact on those technologies. And then I studied at Berkeley, and then I came to MIT as a professor in 1998. And at that time observed that the convergence of computing, communication, especially wireless and sensing, was starting to happen. You could create these devices in really small form factors, and today they're everywhere, right? It's smartphones and Internet of things devices and so on. But if you think back 25 years ago, this was not the case. But I created a set of projects that asked the question, if you had these, what could you do with it? And also looking at it in terms of various applications and things like transportation and mobility, to ask, what technologies do we need to create? So that's really where it came about. And today, and certainly over the last decade, it really is centered around communication, networking, sensing, and perception. When you take devices that are capable of those things, connect them to services running on edge devices or in the cloud over a variety of networking technologies, what can you do with sensor data?
And in the context of our company, what can you do in the context of changing how mobility works, both now and the future?
[00:03:54] Speaker B: Okay, let's dive into that. It's incredible. I mean, over the last 50 years, computers have gone through the size of an entire room to something that can be on your arm like a watch.
And the surprising thing is that watch is stronger computing power than what that room could do 2030 years ago. So it's been an incredible progress of computing. This concept of sensors that are now everywhere just opens up so many opportunities.
What is the problem that you guys have focused on?
[00:04:30] Speaker A: Yeah, you're right.
Ability to have sensors and manufacture those sensors and put them in consumer devices has just been game changing. And we focus on the problem of safety, particularly safe and sustainable mobility.
Our mission is to make driving safe and sustainable all around the world. And we do that by gathering data, with the user's consent, from the sensors on a variety of their devices, whether it be their phones or I IoT devices that they deploy, that they use, all with the intent of providing them feedback and incentives to make them better drivers and safer drivers. The technology is fascinating because the sensor data, like accelerometer data, gyroscope data, velocity position, et cetera, is extremely noisy and prone to error. So we take that data, and using a variety of algorithms rooted in signal processing and machine learning, take that mobile sensor data and convert them into actionable insights that help us understand and help our customers understand how well somebody's driving or how badly they're driving. Give them a score to assess risk, to provide incentives in the form of coaching and feedback, both real time and non real time, to improve their driving, to make them better drive and lower crash rates on our roads. And third, if they were so unfortunate as to be in a crash, to use the same sensing technologies to identify detect crashes in real time, at all levels of severity, from the most severe life threatening situations all the way to the fender benders, and determine the action, whether it be setting emergency or towing, or just a call, and then using that data to insurance claims to make them get them back onto the road quicker. So being able to complete that entire lifecycle in the context of mobility, working with insurance companies and ride share companies such as Uber, or the public sector and government agencies or carmakers to really make mobility safer. And this is really interesting, because a lot of the reason why mobility has become less safe, despite vehicles becoming safer and safer and more laden with technology is because of the smartphone or distraction. I call these weapons of mass distraction. I love that with these weapons of mass destruction. But we're using sensors on these devices and the engagement methods provided by these devices to make driving safer. So, for example, our partner discovery in South Africa, using our technologies, has shown we can work with them. We've shown 24% lower crash rates compared to the general population when you tie it all together. So we're providing a win for the individual, lowering their, improving their safety, a win for society, lowering crash rates, a win for our partners in insurance and broadly in mobility. And I think it's really creating a wonderful virtuous cycle where we can benefit everyone in a winning proposition.
[00:07:31] Speaker B: That number, 24% less car accidents, is an incredible number, and it's hugely surprising. Might I say, what is contributing to that? Why is an app on a phone where you're not even supposed to be on your phone? How is that reducing car accidents?
[00:07:48] Speaker A: Yeah, I mean, people should not use apps while driving, and these apps don't involve the user while they're driving. I mean, they're gathering data in the background, either through the phone or through IoT devices. We have a driver tag device that's a small two inch by two inch device attached to the windshield. We've shipped over 40 million of these devices, and in fact, over the last few years of our company, we've protected over 40 million users with a variety of technologies. So the reason why this works is because on the face of it, most people actually believe they're very good drivers, even when they're not. And what has happened here is bye. Providing the right type of feedback and combining it with the right type of incentives. For example, economic incentives like insurance discounts, which many insurers, partners of ours in the US, and abroad offer or providing rewards for safe driving or other types of incentives, or providing family safety modules so the entire family can be kept safe. Providing crash detection and roadside assistance services, as well as real time coaching. In many cases where people are in real time when they do something bad or something, I shouldn't say bad, but something that might be risky. Being informed of that. All of these combined together are able, in really good programs, to achieve significant improvements in crash rates. It's fundamentally because safe drivers are made, not born. I think almost anyone with the right incentives and feedback can be a safe driver.
[00:09:20] Speaker B: That is a lot to unpack. So let's take this, break this down slowly.
If I am an inherently safer driver, and I have a way to showcase that through the data that's being collected by the app, I can actually get a discount from the insurance company. Is that correct?
[00:09:38] Speaker A: That's exactly right. And these discounts today are quite substantial.
For example, progressive insurance, state Farm, all these companies that are partners of ours, they have significant discounts that could be, you know, north of 40% in some cases for the best drivers. In fact, even higher, I think. And they also provide a participation discount for just joining the program in the beginning. And then upon renewal, you get those significant discounts. And given the inflationary environment we're in, and because vehicles are laden with so much technology now, it's the point where even fender benders or relatively minor crashes involve actually quite a large sums, quite large sums of money to repair the. And so I think that as insurance premiums rise, the best way for consumers to lower their insurance premiums is to participate in telematics programs where their rate is determined fundamentally by how they drive. And these are very controllable factors. Put your phone away. Be more attentive to avoid heartbreaking. Don't excessively speed. Be attentive. Be calm while driving. And just being told about what you've been doing is quite revealing and can improve driving quality. I mean, my family uses it.
[00:10:52] Speaker B: This I love. So this is incredible. So we talked about the people who are already great drivers, right? And they should be basically rewarded.
But what you're saying now is incredibly interesting. You're saying, even if I've had, let's say, a history of bad driving, and maybe I've been dinged now, that's it. I'm carrying that burden for life. But what your technology is allowing is for people, actually create a track record, a history of showcasing that they are good drivers. And the platform is also giving feedback to create an incentive design to improve your driving.
So you're really changing driver behavior. That's incredible.
Tell us about how. I'm really curious, how do drivers, how do they experience this process? And what's the feedback that you give when they get this feedback? Because, I mean, my first reaction to somebody telling me that, oh, you did something wrong, maybe to flip them off. So how does that work out?
[00:11:56] Speaker C: So if you take a step back for a second and remove yourself from the technology, and I'm very intellectually curious as well, I believe you need to start with the human condition. So your question really focuses on the human condition. One of our first investors, Vitality health. From your home, your birthplace of Johannesburg, South Africa, Vitality Health focuses on what your healthcare, health insurance, life insurance costs should be by how you live. Do you take the stairs, do you walk to work? Do you drive? What do you eat? How often are you at the gym? And they offer rewards based on good choices and healthier living, which is really interesting. So the model now with what we're doing in the vehicle is similar. If you look at the human condition, people are motivated by two things. You know, one could be punitive, but the other could be rewards based. So one of the reasons that we've done so well is really focusing on the rewards base, giving positive affirmation, reinforcement of good behaviors. Once people realize there's very few really bad drivers, okay, there's some, but most people think they're good and are just not great. What we do is try to communicate to people micro and possibly macro changes in behavior that make a big difference. You know, we have, one of our safety advisors lost a son to distracted driving. And he does a wonderful presentation on behaviors that we used to do that no one thought anything of until we realized they were really bad. Smoking on an airplane could be considered one. Smoking or drinking while pregnant could be considered one. Drinking and driving certainly could be considered one. But what if you took that further and you looked at being fatigued behind the wheel, slow reaction time, what time of night, what time of day is bad for you to drive? You look at these aura rings and the feedback loop that we're getting there. All of these things are interconnected and are going to be helping to have a better, more sustainable society. So our technology is brilliant, but without educating the human condition on how to improve, it's just really good technology.
We believe we can change the world and save lives while we're offering people rewards. Safer driving discounts are beneficial to the consumer, beneficial to the society, beneficial to the insurance company, beneficial to all involved. Similar to not smoking on an airplane. Hari, anything else you'd like to add on that?
[00:14:30] Speaker A: That was great. Thank you.
[00:14:34] Speaker B: The car industry is developing, even though I would argue to a certain degree, it's been dormant for the last 20 years.
Over the last, let's say, three, four, five years, the technology in cars has been accelerating, maybe due to Tesla and other manufacturers. Are we going to see a time that these technologies are just going to be part of the car?
What does that mean for you guys?
[00:15:03] Speaker C: Yes. So at the risk of telling you some statistics, that changes the question. Hari, why don't you take this and I'll fill in after?
[00:15:11] Speaker A: Yeah, it's a good question. It depends on the timescale. So let me just lay it out. So, since 2017 in particular, there has been a tremendous amount of new technology with semi assisted driving and all sorts of warnings and interventions and even takeovers of the driving. But what's been very, very interesting is that over that same timeframe, crash rates have actually gone up, not only in aggregate, but also per mile. So I call this the vehicle paradox. We put in more and more and more technology, and I think a lot of it is very good, but it's become the data has not kept up. In terms of the reduction in crashes, there's a few different reasons for it. One of the reasons is, in fact, that these are not complete replacements for humans. They sometimes take over from the human. The human is still heavily required to be involved. So that's why I said it's the nature of timescale. Now, 20 years from now, we're going to certainly have in large some parts of the world, a significant, real, true autonomous driving. And in that world, crashes won't be eliminated, unlike what some people believe they will take on a different form. However, that world will still be hybrid, involving a combination of vehicles and human drivers.
The other part of the equation that's pretty important to understand is that worldwide, about half of the people seriously injured and killed in crashes are outside of the car. They're not in the car. So these cars might have made people in the car, may have made some of them safer, and probably have, but I don't know that they've done the same for people outside. And we need to take a more holistic view at the question. What it means for us is that I actually think that as there's more technology in these vehicles, the ability to determine the safety of a vehicle will fundamentally be based on telematics, based on understanding this quality of sensors, the quality of the AI algorithms, and the quality of the software itself. So today we're using this to measure human drivers and improve human drivers. In 1520 years, we'll be using it for that, but also to measure computer drivers and to use the same algorithms to improve those drivers.
[00:17:29] Speaker C: Ari, if I could touch on that a little bit. So think for a second, the average lifecycle of a vehicle today. If a new car came off the lot today, what would the average life cycle of that car be? Do you have any idea?
[00:17:43] Speaker B: I would guess less than 20 years.
[00:17:46] Speaker C: It's about twelve years.
[00:17:47] Speaker A: Twelve years on average. But there's a tail.
[00:17:49] Speaker C: But then those cars are typically sent to places like eastern Europe, Southeast Asia, parts of the Middle East, Latin and South America. So these vehicles are not quote unquote intelligent every day it's not. They're either autonomous or something else. There's varying degrees of that.
What we're doing at CMT is connecting the unconnected, and we're agnostic to the data source. For instance, whether it's a phone, a purpose built IoT device that connects to the car, which connects the unconnected, whether it's information from the technology that's in the vehicle. We've created a platform called the Fusion platform. So we're agnostic to the data source of where we bring that in and what is ingested. But then the output of that needs to be harmonized for the end customer. Now, I don't mean the consumer, but perhaps there's a government that is building infrastructure, roadways, highways, bridges, to be ready for autonomous driving. Perhaps there's an insurance company that wants to know what's happening in a particular city based on weather patterns or time of day. Perhaps it's a car maker that wants to fuse the data between their, their brand new intelligent vehicle ten years from now with other vehicles in their fleet. So we are and believe strongly whether it's autonomous, semi autonomous, intelligent, traditional. We are purpose built to help solve all of those data problems, because it's not just collecting the amount of data. We've got petabytes upon petabytes upon petabytes. It's really harmonizing that data for the use case on the other side.
[00:19:32] Speaker B: Bill, this is such, and Harry, this is such a wonderful mission. I don't think we get to work in companies that actually save lives. So having the opportunity to bring a product like that to the community is absolutely wonderful. So I appreciate the work that you're both doing. I want to ask you both a difficult question. What would you tell 20 year old Bill or 20 year old Harry, Bill, go ahead first.
[00:20:05] Speaker C: So Hari and I, this is a deeply personal question for me, Ari, so I might take you out in an area you haven't contemplated when asking that question. So 20 year old Bill grew up in Boston. 20 year old Bill was homeless at 1720. Year old Bill knew that MIT existed from where he grew up, right around the corner. But because they had really nice fraternities and they had parties on the river, 20 year old Bill would never consider that he would become a brother with a much younger man from India who grew up with parents who worked for the equivalent of NASA and were academics and brilliant. And so it would have been hard. But looking back, I would hope that 20 year old Bill would be like, hey, that guy is really smart and really cool and a really good human being, and we should try to do something together.
On paper, we have very different backgrounds, but our core intellectual curiosity, compassion, wanting to leave the world better than we found it, is actually almost from birth. I feel like we're the same.
The irony is what's happening in our world today, and I'm not going to go political, I'm going to be very gentle here, but what's happening in our world today is so divisive. And we both believe that doing good matters and you can do a really good job and still have a highly profitable business. I'll pause there.
[00:21:33] Speaker B: I love that. I'll share. Before inviting Harry to share his story managing the Siemens consulting business, I got to travel all over the world and I met a gentleman basically living in the middle of the Amazonas forest. It's a place called Manaus in Brazil.
And I felt such a brotherly connection to this gentleman more than anybody else. That was my neighbor or my school friend from school. And that really shocked me because it showed me that really, people are people wherever you go. And it really doesn't matter where you were born or the color of your skin or your religion, there are things that fundamentally bring us together as human beings. So I tremendously appreciate your comment, Bill. Thank you.
[00:22:23] Speaker C: I think it's a part of our superpower, Ari, that, you know, I'm big on. Hari and I are big on diversity. Diversity of background, diversity of thought, diversity of life, diversity of experience, diversity of education, diversity of country of origin. I mean, I'm talking true diversity of all of these things. And if you humble yourself to listen to it, you can actually learn a lot. Hari, go ahead, man. We've been chatting.
[00:22:56] Speaker A: This is very nice to hear. I know your story. We've been together now 15 years and Ari, that's such a good question. I think I would say that 20 years ago, when I was 20, I was mostly focused on my work and just very driven to, you know, whatever the research or the academic work I was doing. And I was also very focused on a few sports like cricket and soccer that I would be part of and compete in. But I think I wasn't really paying attention to interpersonal relationships and people and sort of understanding that. I mean, I've learned that over the years from many people, including, especially Bill, just observing how you interact with people and learn what makes teams work, but also what makes each individual, you know, what matters to them and how you can create something. So I would say that, yeah, it took me a long time to learn about people, and I'm glad I've done that now and still learning.
[00:23:57] Speaker B: Bill Hari, thank you so much for joining the show today. I truly appreciate you both.
[00:24:01] Speaker A: You're welcome. Thank you for having us. Hari.
[00:24:03] Speaker C: Thanks for having us. Harihe.
