Russ: Hi, I’m Russ Capper and this is BusinessMakers USA Live, brought to you by Insperity, inspiring business performance. And I’m very pleased to have as my guest today, distinguished engineer with IBM, and CTO of IBM Watson partnership, Sridhar Sudarsan. Sridhar, welcome to the show.

Sridhar: Thank you.

Russ: Ok, so IBM Watson sort of came on our radar screen in 2011, when those of us that watched TV saw IBM Watson crush Ken Jennings and Brad Rutter in Jeopardy. So, I want to go back to the beginning of IBM Watson. Certainly, there was already an initiative in order to start with cognitive computing, and after a couple years maybe, somebody looked up and said, hey we ought to, we ought to compete on Jeopardy. Is that how it went?

Sridhar: Uh, pretty much.  It actually started in a bar, like most of these stories do.

Russ: I’ll drink to that.

Sridhar: Yes, please. So, the technologies has been around for over three or four decades, so it’s about thirty, forty years of research, and papers, and science that we have all been working on; the research teams within IBM have been working on, had been working on. And, as I said, as a team from the research group in Poughkeepsie, when they went to the bar, Jeopardy and Ken Jennings was on a role at that time; his winningest streak, if you will. And so they looked at it and they said, ‘Hey, what if we built a computer? Can we beat that guy?’ And so that’s when it started, and it culminated, or that project culminated in the Jeopardy live showing, which hopefully many of you have seen. And then we looked back and said, well, looked forward and said, ‘Well, where can we take this technology?’ And so, that’s how we’ve evolved from that research project to commercializing the technology, to almost pioneering, if you will, the industry. So, IBM as a company, our Chairman has committed to transforming IBM as a cognitive company, and also transforming our clients into being cognitive. The industry, the clients company is becoming more cognitive.

Russ: Ok, but just for the record, it’s accurate to say that IBM Watson was conceived in a bar.

Sridhar: Sure.

Russ: Ok, so you (Sridhar: We’ll drink to that.) mentioning, ok, so you mentioned your Chairman in this transformation. In everything I read in the Wall Street Journal, IBM is kind of betting the farm on this;

Sridhar: Yeah, you are right.  So, I think as a company, IBM has, is undergoing a transformation, I would say. As I said, we don’t sell hardware anymore. We’ve moved to be more of a software as a service company, so we’re betting big on cognitive technologies enabled through the Cloud, which is offering these technologies as a service. And you’re right, cognitive is one of the big bets that we as a company are placing. And what does that really mean, is the Watson group, which is the group that I’m a part of, is really the foundation of these cognitive technologies, and if you look at, we have several business units now that are really based on offering cognitive solutions to the market. So, we have a Watson health care unit that focuses on specifically health care solutions. We have a Watson IoT unit, which offers Internet of Things solutions; the security space, financial services, and so on and so forth. So, that’s kind of how we’re transforming ourselves and offering solutions for our clients.

Russ: So one thing that comes to mind all the time when I hear about all the different kinds of data that IBM Watson can ingest and interpret and stuff, it’s huge, because I came from the data processing days when I was at IBM. And that’s when, you know, every piece of data had to be a certain number of characters and crammed into a certain format and form. And we had this term, I don’t know if there is anybody old enough to remember it, but it was ‘garbage in, garbage out,’ and that just doesn’t apply anymore right?

Sridhar: It actually does.

Russ: It does?

Sridhar: Yeah. So, you know, it does. So data is everywhere. I think the tolerance level of… and it is the number of systems and technologies, the advances that we’ve made in terms of how you filter signal from noise, has advanced significantly, but there’s still just so much of data being generated every day. Data has doubled in the last three years, and data is continuing to double every two years now, and by 2020, beyond 2020, 80% of the data is actually uncertain as to where it’s going to come from, what kinds of data. And when I talk about data, it’s really everything, right? So, these recordings that we’re doing today, videos, images, x-rays, things that we talk, all the interactions that you’re doing with your phone, you know, every movement that you have, so all the GPS data that’s being collected, all of this is really, are examples of data, and essentially, all of this does come to a point where you have systems that are specially built, and this is where you build adaptable systems within an industry for a specific purpose that learns to recognize information and signal from the noise, but there still are times when you find that there’s a lot of garbage in and you see garbage out.  And interestingly, these systems are now learning systems. So, unlike programs where, you know, if you set a set of rules, if you have a set of rules, which is how programmatic computing is, there’s a lot more opportunity for, possibility, not opportunity, possibility for putting garbage in and getting garbage out. With learning systems, it’s similar to humans. So, if people are talking, you know, things that are unrelated to a certain audience or in a context (Russ: Like fake news?), that could be. That could be an example. So, fake news is not so much a garbage in, garbage out, but it is an effect of that, right? Because somebody can game a system and spread news that is not necessarily true, so it’s not based on reality.  So, now if you think about how are we working around that, is we’re trying to build more ways to authenticate the information, and if you look at companies like Facebook, they’re putting more thought into how can they actually authenticate the kinds of news that travels fast, which is good, but it also means that there needs to be more security, there needs to be more authorization about how this news is being spread.

Russ: There should be, maybe somebody here should build a fact checker for debates, you know? We need it when (Sridhar: how novel.). Is it possible to do real time fact checking?

Sridhar: Absolutely.  And I don’t know if you’re referring to Presidential debates, or debates in general, or our discussion here and somebody might really be doing fact checking. But no, absolutely, it is very possible to do fact checking. And if you think about how we do fact checking today, even if you think about some of our presidential debates leading up to the election, there were these websites that were doing fact checking, but the way they typically happen today is, there are systems, there are people who are vigorously typing in messages or facts that are coming, that are being brought up at these debates, and they’re running searches, and then they’re sort of filtering, and sifting, and then they’re sort of providing the evidence. So, it’s really introducing a lot more transparency very quickly, but if you take it to the next level, systems like Watson can take an entire sort of set of conversations and run fact checking against data that it has access to; which could be news sites, which could be in articles, etc. So, yeah.

Russ: Ok, so it also seems to mean, and this might really be of interest to our audience, both people live and the people watching this, when you start looking at the clients, the customers today, the apps that they’re building, there’s a lot of real small companies that are using IBM Watson.  And it’s not like you’re just looking at the Fortune 500, or even the Fortune, I mean they’re startups, right? Do I see that right?

Sridhar: Absolutely.  So, one of the things that we, that we kind of learned over the last few years as we are commercializing and taking this technology to market is really focusing on the entire ecosystem of users, if you will, right? And again, if I might quote my Chairman, Ginni Rometty, she believes that there will be, or at least she said a target that we will have about a billion users touching, using, being impacted by Watson technologies, you know. So what that really means is, we can’t just focus on, we know we cannot do it all, but we want to build, we IBM with the Watson Technologies, want to build sort of the core foundational elements; the platform, the industry specific offerings, the methodologies for people who are working with and enabling the Fortune 1000 companies, or Fortune 2000 companies, if you will.  There’s actually over 33,000 startup companies that are working, sort of, in some way or shape in this space, So, we’re really opening the horizon, making it, and as I said, offering these services on the Cloud as a service makes it a lot easier. So, it’s very easy for, you know, two developers in a garage to actually get started and use technologies from IBM now.

Russ: Really cool.  So, I was kicking around on some of your websites, too, and that’s where I got this impression that, wow, these are startup small companies using this powerful technology. And I came across this one called Ampsy, A-m-p-s-y, and I was just kind of reading it, you know, as regular business stuff, but then I saw social media, and then I saw an event, and then organization. And so I said, ‘Wow, Watson and social media.’ And still, I was kind of getting interested and then I just flipped over to see who Ampsy’s customers were, and they were Guns N’ Roses, Brittney Spears, AT&T, and then I went back and looked at it, and it was kind of mind boggling. It was like, if you’re holding an event, but you know it was like they could draw a boundary around the event, monitor all social media, and build these pictures for their clients of their customers and fans. I mean, did I get that right, or was I even close?

Sridhar: No, absolutely you did. Actually, so it’s really about reaching an audience at scale, but reaching them in a very personalized manner. So, it’s about personalization at scale, and today with the number of devices that we all tend to carry on our body, with wearables, and mobile phones, and all kinds of connected devices, you as a consumer, I as a consumer want to be in control of what I’m looking at, when I’m looking at it, why I’m looking at it, and how I’m looking at it.  So, for the brands, for these companies, for these large event organizers, these

Russ: The customers of Ampsy

Sridhar: that’s right, the customers of Ampsy, they want to reach you, and so it’s almost a way to do so in a very personalized manner.  So, somebody, you know, AT&T trying to sell a package to you might be very different from how they sell it to me. So, it’s really about reaching an individual at the right time with at least a semblance of control of who they’re talking to. So, that’s where I think companies like Ampsy use technologies like Watson to look at a geographical area. So, you mention an event. So, if there’s a concert going on, let’s say in Seattle; being able to draw that boundary, know the kinds of people that are visiting, that are actually attending that concert, being able to then not just know who they are or what kind of Twitter IDs they have or Facebook profiles they have, but going deeper into the kinds of things that they’re talking about. So, what they’re talking about and how they’re talking about it, right? So, that kind of starts creating almost a personality. Understanding the tone of what they’re saying, how they’re saying it, and then using that in the interaction.

Russ: Really cool.  So, wow this is a fascinating, so let’s move on to medicine. You know, I’ve watched and interviewed a couple of these disruptor specialists that are around right now that are predicting doom and gloom because, you know, artificial intelligence, autonomous vehicles, robotics are going to replace many more jobs than they’re going to create. And specifically, I heard that about medicine, you know, and you guys are doing lots in medicine. I mean, 60 Minutes, two or three weeks ago you were on about that cancer report, which is very fascinating, but just talking about the general practitioner, you know? It’s almost like they might not need a, we might not need a human anymore because, you know, we’ve got Genome, we have the web, we have wearables, and we have this computer that we can ask questions, we can give them our symptoms and they can tell us what we ought to do, maybe better than a general practitioner. Is IBM Watson going to replace doctors?

Sridhar: No.

Russ: Ok, next question.

Sridhar: thanks. No, so I think we’re nowhere close to having systems, cognitive systems like Watson replacing doctors, for many reasons, right? A: the technology is still, I would say, in it’s very early stages. You know, if you think about it, the number of, the amount of knowledge that we have about our own human body after tens of thousands of years is very small. One of the reasons for that is that there is so much information that there are human limitations in terms of how we consume it. Our doctors have the same kind of challenge, and compounded by the fact that the more time they spend with people, like you and me, right, Their patients, the less time they have to consume that information, to absorb that, to digest and process that kind of information, that’s where technology steps in. That’s where a cognitive system like Watson steps in. So it reads, you know, 700,000 papers that get published every year, and it presents the information to the doctor, again, back to my earlier comment, at the time of need, in the context of when the patient is with the doctor, using all the information about that particular interaction that the doctor and the patient are going through, and being able to provide advice to the doctor or recommendations to the doctor, who then has to sort of pick on the best conditions, right? Pick on the best treatment advice, treatment recommendations. So it’s really about augmenting the doctors intelligence through the consumption of large amounts of information that is not humanly possible So, it’s timely, its scalable; that’s where systems come in and help the doctor to do that final mile, if you will.

Russ: Ok, so let’s move to IoT; Internet of Things. I mean, my god, it’s everywhere in stuff. Does Watson play, plan to play a role in IoT?

Sridhar: Absolutely.  you know, 2020, I think we’re going to have about fifty billion devices, and I that’s the current estimate. That might multiply. fifty billion devices, with a ‘B’, that are going to be connected with one another, and that’s a significant amount of, a significant number of devices, but what it really means is it’s generating a significant amount of data that is going between these devices, right? And the devices could be things like: your chair, to your lamps, to your thermostats, to cars, you know, to pretty much anything you can think of.  So, being able to recognize all these things based on your interactions and being able to respond to that, applying that in the industrial space, if you think about that in an aircraft, it’s all about turning things around quickly. So, recognizing that there’s a problem, so by the time you’re on the ground, being able to repair and have personnel, maintenance engineers ready to be able to respond to it, it’s really all about having these things talking to one another.

Russ: Ok. Alright, so we’re running out of time, but this is my favorite thing that I hope you’re going to tell us. Maybe it’s confidential, maybe there’s a back room somewhere that there are IBM Watson people inventing true, valid, workable, reliable, cyber security.

Sridhar: Yeah. I mean, so remember I was telling you we have several business units? So, one of our business units is the security, and we actually have some disclosable projects and others, not so much, as with everything else, but there is Watson now cognitive elements in the security models. So, think about systems where, you know there’s about 45,000 day zero hacks, about 45-50,000 day zero hacks that happen every single day.  We very likely, all of us have been hacked at some point or the other. The question is, either you know about it or you don’t. The second question is, whether you know about it or not, something has been done by the person who has hacked, or by the system that has hacked you. And that’s probably what will determine whether you know about it. Alright, so being able to stay ahead of that game again is really a race by systems to absorb the kind of information, the releases. There’s so much software coming from all over the place. There’s 60 million lines of code in a Boeing 787 aircraft, but there’s 100 million lines of code in a car, so there’s every single opportunity for cyber security is a real problem, and if you think about the utility industry, the impact can be bigger, with the advent of technology, like I said, being able to solve for that is something that we’re absolutely working on. We have a lot of research, we have offerings today as well. We also have partners. There are other companies who we’re working with who have solutions in this space as well.

Russ: Great.  Well, Sridhar, thank you very much. Let’s hear it for Sridhar Sudarsan.

Sridhar: Thank you for having me.

Russ: Thank you all.  And that wraps up this episode of BusinessMakers USA, brought to you by Insperity, inspiring business performance.