 Live from New York, it's theCUBE. Covering theCUBE, New York City, 2018. Brought to you by SiliconANGLE Media and its ecosystem partners. Everyone, welcome back to theCUBE Live in New York City for CUBE NYC. This is our live broadcast. Two days of coverage around the big data world, AI, the future of cloud and analytics. I'm John Furrier, my coach Peter Burris. Our next guest is Paul Appleby, CEO of Connecticut. Thanks for coming back to theCUBE. Good to see you. Great to be back again. Great to visit in New York City. It's incredible to be here on this really important week. Last time we chatted was in our big data at Silicon Valley event, which is going to be renamed CUBE SV because it's not just data anymore. There's a lot of cloud involved, a lot of new kind of infrastructure, but analytics has certainly changed. What's your perspective now in New York as you're in here hearing all the stories that are on the show and as you talk to customers? What's the update from your perspective? Because certainly we're hearing a lot of cloud this year. Cloud, multi-cloud, analytics and infrastructure, proof in the pudding, that kind of thing. Yeah, I'm going to come back to the cloud thing because I think that's really important. We have shifted to this sort of hybrid, multi-cloud world and that's our future. There is no doubt about it. And that's right across all spectra of computing, not just as it relates to data. But I think this evolution of data has continued, this journey that we've all been on from, whatever you want to call it, systems of record to the world, a big data where we're trying to gain insights out of these massive oceans of data. But we're in a world today where we're leveraging the power of analytics and intelligence, AI, machine learning to make fundamental decisions that drive some action. Now that action may be to a human to make a decision to interact more effectively with a customer or it could be to a machine to automate some process. And we're seeing this fundamental shift towards a focus on that problem. And associated with that, we're leveraging the power of the cloud, AI, ML and all the rest of it. And the human role in all this has been talked about. We hear, obviously in the U.S. and the political landscape that data for good, we've seen Facebook up there being basically litigated publicly in front of the Senate around the role of data and the elections. And people are talking in the industry about the role of humans with machines is super important. And this is now coming back as a front and center issue of, hey, machines do great intelligence, but what about the human piece? What's your view on the human interaction component, whether it's the curation piece, the role of the citizen analyst or whatever we're calling it these days and what machines do to supplement that? Yeah, you're a really good question. I've spent a lot of time thinking about this. I've had the incredible privilege of being able to attend the World Economic Forum for the last five years. And this particular topic of how robotics, automation, artificial intelligence, machine learning is impacting economies, societies, and ultimately the nature of work has been a really big thread there for a number of years. And I've formed a fundamental view. First of all, any technology can be used for good purposes and bad purposes. And it's- It always is. And it always is. And it's incumbent upon society and government to apply the appropriate levels of regulation and for corporations to obviously behave the right way. But setting aside those topics, because we could spend hours talking about those alone, there is a fundamental issue. And this is this kind of conversation about what a lot of people like to describe as a fourth industrial revolution. And I've spent a lot of time because you hear people banding that around, what do they really mean and what are we really talking about? And I've looked at every point in time where there's been an industrial revolution, there's been a fundamental shift of work that was done by humans that's now done by machines. There's been a societal uproar and there've been new forms of work created and societies evolved. So what I look at today is, yes, there's a responsibility and a regulatory side to this, but there's also a responsibility of business and society to prepare our workers and our kids for new forms of work. Because that's what I really think we should be thinking about, what are the new forms of work that are actually unlocked by these technologies rather than what are the roles that are displaced by the steam-powered engine? Well, Paul, we totally agree with you. I want to, there's one other step in this process that kind of anticipates each of these revolutions. And that is there was a process of new classes of asset formation. So you could go back to when we put new power trains inside row houses to facilitate the industrial revolution in the early 1800s. You could say the same thing about transportation and what the trains did and whatnot. So there was always this process of new asset formation that presaged some of these changes, today it's data. Data is an asset because businesses ultimately institutionalize or re-institutionalize their work around what they regard as valuable. Now, when we start talking about machines telling other machines what to do or providing options or, you know, pairing off options for humans so they have clearer sets of things that they can take on, speed becomes a crucial issue, right? And at the end of the day, all of this is going to come back to how fast can you process data? Talk to us a little bit about how that dynamic and what you guys are doing to make it possible is impacting business choices. Yeah, two really important things to unpack there and one I think I'd love to touch on later, which is data as an asset class and how corporations should treat data. But you talk about speed and I want to talk about speed in the context of perishability because the truth is if you're going to drive these incredible insights, whether it's related to a cyber threat or a terrorist threat or an opportunity to expand your relationship with a customer or to make a critical decision in a motor vehicle in an autonomous operating mode, these things are about taking massive volumes of streaming data, running analytics in real time and making decisions in real time. So these are not about gleaning insights from historic pools or oceans of data. This is about making decisions that are fundamental to the environment that you're in right now. You think about the autonomous car. Great example of the industrial internet's one we all love to talk about. The mechanical problems associated with autonomous you've been solved, fundamentally sensors in cars and the automated processes related to that. But the decisioning engines that need to be applied at scale in millions of vehicles in real time, that's an extreme data problem. I mean, the biggest problem to solve there is data. And then over time, of course, societal and regulatory change that means that this is going to take some time but we're just... I think it was a hundred, Tesla's generating a hundred terabytes of data a day based on streams from its fleet of cars that its customers have. Yeah, we firmly believe that longer term when you get to true autonomy, each car will probably generate around 10 terabytes of data a day. And that is an extremely complex problem to solve because at the end of the day, this thinking that you're going to be able to drive that data back to some centralized brain to be making those decisions for and on behalf of the cars is just fundamentally flawed. It has to happen in the car itself. I totally agree. So this is putting supercomputers inside cars. Which it kind of is kind of happening. In fact, that hundred terabytes a day is in fact the data that does get back at Tesla. There's, as you said, there's probably 90% of the data that's staying inside the car. It's just unbelievable scale. So the question I want to ask you, you mentioned, you know, industrial revolution. So every time there's a new revolution, there's an uproar you mentioned, but there's also a step up of new capabilities. So if there's new work being developed, usually entrepreneurial activity, you know, weird entrepreneurs figured out that everyone says they're not weird anymore, it's great. But there's a step up of new capability that's built. Someone else says, hey, you know, the way we used to do databases and networks was great for moving, you know, one gig ethernet at the top of the rack. Now you got 10 terabytes coming up a car over wireless spectrum. We got to rethink spectrum or we got to rethink database. Let's use some of these GPUs. So a new step up of suppliers have to come in to support the new work. What's your vision on some of those things that are happening now that you think people aren't yet seeing? What are some of those new step up functions? Is it on the database side? Is it on the network? Is it on the 5G? Where's the action? Because who's going to support the Teslas? Who's going to support the new mobile revolution, the new iPhones like, you know, like the size of my two hands can put together? What's your thoughts on that? The answer is all of the above. And let me talk about that and what I mean by that. I think they're there because you're looking at it from the technology perspective. I'd love to come back and talk about the human perspective as well. But from the technology perspective, of course, leveraging PowerGPU is going to be fundamental to this because if you think about the types of use cases where you're going to have to be giga-threading queries against massive volumes of data, both static and streaming, you can't do that with historic technologies. So that's going to be a critical part of it. The other part of it that we haven't mentioned a lot here but I think we should bring into it is if you think about these types of industrial internet use cases or IoT, even consumer internet IoT related use cases, a lot of the decisioning has to occur out at the edge. It cannot occur in a central facility. So it means actually putting the AI or ML engine inside the vehicle or inside the cell phone tower or inside the oil rig. And that is going to be a really big part of shifting back to this very distributed model of machine learning and AI which brings very complex questions in of how you drive governance and orchestration around deploying AI and ML models at massive scale out to edge devices. Inferencing at the edge, certainly. It's going to be interesting to see what happens with training. We know that some of the original training will happen at the central site but some of that maintenance training, it's going to be interesting to see where that actually is probably going to be a split function. But you're going to need really high performing databases across the board. And I think that's one of the big answers, John, is that everybody says, oh, it's all going to be in software. It's going to be a lot of hardware advances. Well, the whole idea is provocative to think about and also intoxicating if you also want to go down that rabbit hole is that if you think about that car, okay, if you're going to be doing essentially machine learning at the edge, okay, where's the, what data are you working off of? So there's got to be some storage. And then what about real time data coming from other, either horizontally scalable data sets? So the question, what do they have access to? Are they optimized for the decision making at that time? So this is a real, again, talk about the future of work. This is a big piece, but it's the human piece as well. I mean, are our kids going to be in a multi-massive, multi-player online game called Life? They are, they are now. They're on Fortnite, they're on Call of Duty, and they're all this gaming culture. But I think it's one of the interesting things because there's a very strong correlation between information theory and thermodynamics. They're the same exact, in physics, they are the identical algorithms and the identical equations. There's not a lot of difference. And you go back to the original revolution, you build, you have a series of row houses, you put a power supply all the way down, you can run a bunch of looms. The big issue was entropy. How much heat are you generating? How do you get greater efficiency out of that single power supply? Same thing today. We're worried about the amount of cost, the amount of energy, the amount of administrative overhead associated with using data as an asset. And the faster the database, the more natural it is, the more easy it is to administer, the more easy it is to apply to a lot of different cases, the better. And it's going to be very, very interesting over the next few years to see how does database come in memory? Does database stay out over there? A lot of questions are going to be answered in the next couple of years. We try to think about where these information transducers actually reside and how they do their job. Yeah, and that's going to be driven, yes, partially by the technology, but more importantly by the problems that we're solving. You know, here we are in New York City. You look at financial services. There are two massive vectors in financial services going on. What does the digital bank of the future look like and how the banks interact with their customers? And how you get that one true one-to-one engagement, which historically has been virtually impossible for companies that have millions or tens of millions of customers. So fundamental transformation of customer engagement driven by these advanced or accelerated analytics engines and the power of AI and ML. But then on the other side, if you start looking at really incredibly important things for the banks like risk and spread, historically because of the volumes of data, it's been virtually impossible for them to present their employees with a true picture of those things. Now with these accelerated technologies, you can take all of the historic trading data and all of the real-time trading data, smash that together and run real-time analytics to make the right decisions in the moment of interaction with a customer. And that is incredibly powerful for both the customer but also for the bank in mitigating risk. And there the sorts of things we're doing with banks up and down. The city here in New York and of course right around the world. So here's a question for you. So with that in mind, this is kind of a more of a thought exercise. Will banks be even be around in 20 years? Oh, wow. I mean, we've got blockchain saying we're going to have new crypto models here. If you take this Tesla with 10 terabytes going out every second or whatever that number is, if that's a complex problem, banking should be really easy to solve. I think it's incumbent on boards in every industry, not just banking, to think about what existential threats exist because there are incredibly powerful, successful companies that have gone out of existence because of fundamental shifts in buying behaviors or technologies. I think banks need to be concerned. Every industry needs to be concerned. But every industry needs to be concerned. At the end of the day, every board needs to better understand how they can reduce their assets specificities, right? How they can have their assets be more fungible and more applicable or appropriable to multiple different activities. Think about a future where data and digital assets are a dominant feature of business. Assets, specificities go down. Today, the very definition of vertical industry is defined by the assets associated with bottling, the assets associated with flying, the assets associated with any number of other things. As assets, specificities go down because of data, it changes even the definition of industry, little on banking. And auto industry is a great example. Will we own cars in the future? Or will we consume them as a service? Auto manufacturers need to come to terms with that. The banks need to come to terms with the fact that the fundamental infrastructure for payments, whether it's domestic or global, will change. I mean, it is going to change. It has to change. It's in the process of changing. And I'm not talking about crypto. What form of digital currency exists in the future, we can argue about forever. But a fundamental underlying platform for real-time exchange, that's just the future. Now, what does that mean for banks that rely heavily on payments as part of their core driver of profitability? Now, that's a really important thing to come to terms with. Or going back to the point you made earlier, we may not have banks, but we will have bankers. There's still going to be people who are providing advice and counsel, helping folks understand what businesses to buy, what businesses to sell. So whatever industry they're in, we will still have the people that bring the enterprise to the data. We got to break it there, run out of time. Paul, love to chat further about the future of banking, all this other stuff. And also, as we live in a connected world, what does that mean? So we're obviously connected in the data. We certainly know there's going to be a ton of data. We're bringing that to you here in New York City with Cuban NYC. Stay with us for more coverage after this short break.