 From San Jose, California, it's theCUBE, covering Big Data Silicon Valley 2017. Okay, welcome back everyone, live in Silicon Valley for Big Data SV, Big Data Silicon Valley. This is SiliconANGLES, the CUBE's event in Silicon Valley, with our companion event, Big Data NYC, in conjunction with O'Reilly Strata Hadoop, Hadoop World, our eighth year. I'm John Furrier, my co-host, Jeff Frick, bringing down all the action, and our super guest, Abhimeta, the CEO, Trasada. We've been on every year since 2010, and you see a very successful Trasada building out the vertical approach in financial now health. Welcome back, good to see you. Thank you, John, always good to see you. It's an annual pilgrimage to have you on the CUBE. It is literally a pilgrimage. I was exchanging messages with your co-host here, and he was picking me saying, you got to come here, you got to get this to this thing, and I made it, the pilgrimage is successful. Yeah, and so a lot's happened, right? So, I mean, data's the new oil, we've heard it over again. You had the seminal, first interview in 2010 calling the oil refineries, the data refineries. Turns out that was true. You know, we always love to talk about that prediction every time you're on, but it's so much going on now. You can't believe that the shift. Certainly, Hadoop has got a nice little niche position. Of course, as batch, but real-time processing is huge. You're seeing the convergence of batch and streaming and all that good stuff in real-time with the advances of cloud, certainly more compute, Intel processors get more powerful, 5G over the top, you have connected cars, smart cities, on and on, IOT, Internet of Things, all powering this new deep learning and AI trend. Man, it is game-changing, so I see this as a step-up function. Agreed. What's your thoughts? Because this is going to create more data, more action. I agree with you. I always remind myself, John, especially when I talk to you guys that, and this was, we were chatting about this right before we went on air, which is as smart as we as humans are, trends repeat themselves. And now we were talking about AI. We all went to school and did things in AI. The whole neural networks thing has not been new. It's almost like fashion. Bell bottoms come in fashion every 20 years. I will never be seen in them again, hopefully neither will you. And AI seems to be like that. I think the thing that hasn't changed, yes, absolutely agree with you, that as S-curves shift, as we have said, almost at this point, eight years ago, a decade ago, there's a new technology S-curve, fundamentally new technology S-curve, S-curve shift underway. And S-curve shifts take time. We will look back at this 10 years saying, it was literally the first, second inning of this new S-curve shift. And I think we are entering the second innings where the conversation around batch, real-time storage, databases, the stacks is becoming less important. And conversations and AI and deep learning are examples of it. Conversations on how can you leverage cheaper, better, faster technology to solve and answer unanswered problems is becoming interesting. I think the basics haven't changed though. What we have spoken with you for almost eight years remain the same. The three basics around every technology trend remain the same. Do we, I think you guys will agree with me. So let me just play it by you and you can either contest it or agree with me. Data is the new comparative weapon. It is unequivocally clear that the new asset, the new enterprise, the most valuable enterprise asset has become data. And we've seen it in data companies, Facebook, Google, Uber, Airbnb, they're all fundamentally data companies but data is the new comparative weapon. So the more you have of it, the better off you are. So I always love people who say, oh, big data is not a bad term, it isn't. Because big data fundamentally, in those two words, defines the very thesis of what we built Trasera on, which is the more data you have and if you can process and extract intelligence from it, boring your term, extract signal from the noise, you can make a lot of money on it. So I think that fundamental basic hasn't changed. Well, I mean, one of the things that big data was to me was always about, okay, big storage kind of a view. And then we coined the term fast data on theCUBE. So that now speaks to the real time. But it's interesting, I mean, I just see that the four main new areas that are being talked about in outside of the big data world are autonomous vehicles, smart cities, smart home, and media and entertainment. And each one of those, I would say that the data is the new weaponization. So you, there's an article that was great this month called weaponizing AI, and it had to do with Breitbart and the election, and that's media entertainment, you got Netflix, all these new companies. Data is content, content is data as a digital asset. So yes, so this AI component fits into autonomous vehicles, it fits into media entertainment, it fits into smart cities and smart home. But you also raise a very interesting point, and I think that we can take comfort in the fact that we have seen this happen. This is not an idea anymore, or it's not just a wild idea anymore, which is we've seen massive disruption happen in consumer industries. Google has built, it created a brand new industry on how to market stuff, could be any stuff. Facebook created a brand new way of not just being in touch with your friends globally because people have thousands of friends, not true, but also how do you monetize deep preferences, right? So a twist on deep learning, but deep, deep preferences. If I know what Jeff likes, I can market to him better. I think we're about to see, the industry's just mentioned, is where will success come from in enterprise software? I always ask myself that question when I come to any of these conferences, Strata, others, there's now an AI conference. What will the disruption that we've seen happen in consumer industries, in broader industries that we just mentioned, automobiles, media entertainment, et cetera, what is gonna happen to enterprise software? And I think the time is ripe in the next five years to see the emergence of massive scale creation. I actually don't think it'll get disruptive. I think we will see, just like with Facebook, Google, Uber, Airbnb, the creation of brand new industries in enterprise software. I think that's going to be interesting. Mark Andreessen, Mark Andreessen. Mark Cuban said it South by Southwest this week where the Cube was with the AI lounge with Intel. He was on stage saying, the first tech trillionaire will come out of deep learning. And deep learning is kind of the underpins for AI if you look at all the geek stuff. So to your point that a new shift of opportunity, whether it's comes in from the enterprise side or consumer or algorithmic side, is that there's never been a trillionaire. No, there hasn't. But I want to push back a little bit because I don't think it always was that way with data because we used to have sampling. It was all about sophistication on sampling and data was expensive to store, expensive to collect and expensive to manage. Absolutely. I think that's where the significant change is. Now the economics of collecting and storing and analyzing are such that sampling is no longer the preferred method to your point. It's the bigness. Absolutely. Well, you know exactly where I stand on that. Now it's an asset, right? You know exactly where I stand on that. I said on the Cube at this point almost a decade ago sampling is dead. And it's for that particular reason. I think the reality is that it has become a very tricky area to be in. Buzzwords aside, right? Whether it's deep learning, AI, streaming, batch, it doesn't matter, flash, all Buzzwords aside, the very interesting thing is, is there are we seeing as a community the emergence of new enterprise software business models? And I think ours as an example, we are now six years old. We announced to say that on the Cube, we have celebrated our significant milestones on the Cube. I'm going to announce today that, you know we're now a valuable member of society. We pay taxes as a company, you know another big milestone as a company. We have never raised venture money. We had a broad view when we started that every single thing we have learned as our industry enterprise software, the stack, databases, storage, BI, algorithms are free, you know Dave was talking about this earlier this week. Algorithms, analytical tools will all become free. So what is this new class of enterprise software that creates value that can then be sold as value? Because buyers, corporations are becoming smart to realize and say maybe I can't hire people as smart as some of the web industries on this side of the coast, but I can still hire a good talent. The tool set is free, should I build versus buy? It fundamentally changes the conversation. Database is a $2 trillion industry. Where does that value shift to if databases are free? I think that's what is going to be interesting to see is what model creates the new enterprise software industry and what is that going to be? And I do agree with your Mark Cuban's statement that the answer is going to lie in if the building blocks are free and commoditized, which you guys know exactly where I stand on that one, if the building blocks are commoditized, how do you add value in the building block? And it comes from what the point you made. Industry knowledge, data, owning data and domain knowledge. If you can combine deep domain expertise to bring advanced applications that solve business problems, people don't want to know if the data is stored in a free Edge DFS system or in some other system, quantum computing. People don't care. All right, so I got to get you to take on the data layer because this is where we had a lot of guests on stand. With the cloud, you can rent things, algorithms are free, so essentially commoditization has happened, which is a good thing. More compute, everything else is all great, all the goodness around that. You still own your data, so the data layer seems to be the land grab metadata. So how do you cross-connect the data layer to be consistent fabric? Here's how we think of it. And this is something we haven't shared publicly yet, but I believe you see us talk a lot more about this. We believe there are three new layers in the technology fabric. That is what we call the hardware operating system. The battle's been won by a company that we all like a lot, Red Hat. I think mostly won. Then there is what we call the data operating system or you call the data layer. But I think there's a new layer emerging where people like us sit. We call it the analytics operating system. The data layer will commoditize as much as the hardware operating system, what I call the layer commoditized. So data operating system fight is moot. Metadata should not be charged for. Master data management, draining the swamp, whatever you want to call it. Every single thing in the data operating system is a commodity, where you need volumes as another view on our businessmen. You need volumes in the P times V game. You need volumes to sustain a profitable business model. The interesting action, in my opinion, is going to come in the analytics operating system, where you are now automating hard core, what I call finding intelligence questions. Whether it's using deep learning, AI, or whatever other buzzword the industry dreams up in the next five years. Whatever the buzzwords may be in material, the layer that automates the extraction of intelligence from massive amounts of data sitting in the data layer, no matter who owns it. Our opinion is, Treseda as an enterprise software player is not interested to be a data owner. That game I can't play anymore, right? You guys are a content company, though. You guys are data owners, and you have incredible value in the data you're building. For us, it is, I want to be the tool builder for this next Gold Rush. If you need the tools to extract intelligence from your data, who's going to give you those tools? And I think all that value sits in what we call the analytics operating system. And the world hasn't seen enough players in it yet. So this is interesting. Mine, Bender, if you think about it, when you said analytics operating system, that kind of rings a few bells and gets my hair standing on the back of my head up because we're in a systems world now. We kind of talk about this in theCUBE where operating systems concepts are very much in place. You look at this ecosystem and kind of who's winning, who's losing, who's struggling, who's kind of falling away. Is the winners are nailing the integration game and they're nailing the functional game. Like I think like a core functional component of an operating environment, AKA the cloud, AKA data. So having those functional systems, is it systems management, systems operating system game? So what is your view of what an analytics operating system is? What are some of those components? I mean, most operating systems have a linker, loader, pile or all those things going on. So what is your thoughts on this analytical operating system? What is it made of? Three core components that we have now invested six years in. The first one is exactly what you said. We don't use the word integration. We use it, we now call it the same word we've been saying it for six years. We call it the factory, but it's very similar, which is the ability to go to a company with unique or enterprises with unique data assets and enrich, integrate, I will borrow your term, integrate, enrich. So we call it the data factory. The automation of 90% of the workload to make data sitting in a swamp usable data. Part one, we call that creation of a data asset. A nice twist or separation from the word data warehousing you all grew up on. That's number one. The ability to make raw data usable. It's actually quite hard. If you haven't built a company squarely on data, you have to be able to buy it because building is very hard. Number one. Number two is what I call the infusion of domain-centric knowledge. So can industries and industry players take expert systems and convert them into machine systems? Because the moment we convert expert systems into machine systems, we can do automation at very large scale. That is, as you can imagine, the ability to add value is exponentially higher for each of those tiers, from data asset to now infusion of domain knowledge to take an express into machine system. But the value generated is incredibly large as well. So you can afford, if you actually had the system built out, you can afford to sell it for a lot of value. That's number two. The ability to take expert system, go to machine systems. Number three is the most interesting. And we are very early in it. And I use the term on the cube, and I'll offer a slight, I'm going to do more forward thinking over here, which is automation. Today, the best we can do with leveraging incredibly smart machines, algorithms at scale on massive amounts of data is augmenting humans. I do fundamentally believe, just like self-driving cars, that the era where software will automate a tremendous amount of business processes in all industries is upon us. How long it takes, I think we will see it in our lifetimes too. You know, when you and I have both a little more gray hair, we'll be saying, remember what we said about that? I think automation is going to come. I do believe automation will happen. Currently, it's all about augmentation, but I do believe that business- We're going to have some cube bots. Automated cube broadcasting. John will give them your magnificent hair, and then they'll do it. But I do believe automation of complex human processes, the era of enlightenment is upon us, where we will be able to take incredibly manual activities, like hailing a car today, to complex activities, looking at transactional information, trading information in split second time, even quicker than real time, and making the right trading decision to make sure that Jeff's kids go to college in a robo-advisor-like mode. It's all early, but the augmentation will transform to automation, and that will take some time. So those are my three tiers in the AOS. So then, if we are successful at converting the expert to machine system, will the value of that expert system quickly be driven to zero due to the same factors that automation has added to many other things that have been cited to the machine? You guys always blow my mind. You always push my thinking when I talk to you. No, I just love the concept. But then, will the same economics that have driven asymptotically approaching zero cost then now go to these expert systems? You know the answer. The answer is absolutely yes. The question then becomes, how long of an era is it? What we've learned in technology is S-curve shifts take time. This era of enlightenment, what I'm calling the era of enlightenment that enterprise software is about to enable and leaving aside all of the buzzwords. Whether it's deep learning, AI, machines, chatbots doesn't matter. The era of enlightenment will absolute, I think there'll be two things. It will first of all, it'll take time to mature. So yes, whether it's 50 years, 40 years or 30 years, does it at some point become its own commodity? Absolutely. The value, the marginal value we can deliver with a machine at some point does go to zero because it commoditizes it at scale. It commoditizes it. Absolutely. But does that mean that the next 30 years will not be a renaissance in enterprise software? Absolutely not. So I think we will see, let's take the enterprise IT market, about $200 to $300 a year. All of it is up for grabs. And we will see in the next 20, 30, 40, 50 years that as it is up for grabs, tremendous amount of value will be retraded and recreated in completely new industry models. I think that's the exciting part. It's great. I won't live for 50 years, so that's okay. I know we got a couple of minutes or so left, but I want to get your thoughts on something that we're seeing here at theCUBE this year pointed out. And we've kind of teased around it, but again, batch and real-time process streaming, all that's coming together, the center that's IoT data and AI, is causing product gaps. So there's some gaps that are developing. So either a pure-play batch player or your real-time, some people have been one or the other, some are integrating in. So when you try to blend it together, there's product gaps, organizational gaps, and then process gaps. Can you talk about how companies are solving that? Because one supplier might have a great batch solution, data lake, some might have streaming and whatnot. Now there seems to be more of an integrated approach of bringing those worlds together, but it's causing some gaps. How do companies figure that out? I believe there's only one way in the near-term and then potentially even more so in the long-term to bridge that divide that you talk about, because it absolutely is a divide. It's been very interesting for us especially. I'll use our example to answer your question. We have a very advanced health analytics application to go after diabetes. The challenge is, in order to run it, not only do you need lots and lots of data, IoT, streamed real-time from sensors you wear on your body, so you need that. Not only do you need the ability and processing power to crunch all that data, not only do you need the specific algorithms to find insights that were not findable before, you know, the unanswered questions, but you also need the last point, you need to be able to then deliver it across all channels so you can monetize it. That is a end-to-end, what I call, business process around data monetization. Our customers don't care about it. They come to Tracera and they say, I love your predictive diabetes outcomes application. I have rented the system from the cloud, Amazon, Azure, I think it's at this point only two players, we don't see Google much in it, I'm sure they're doing something in it. We've rented you the wheels and the steering and the body, so if you want to put it together to run your car on the track, you could. Everything else is containerized by us. I call them advanced analytics applications. They're fully managed, they're on any environment that is given to them because they are resource-ready for whatever environment they play in, and they are completely backwards and forwards-integrated. So I think you will see the emergence of a class of enterprise software we call Advanced Analytics Applications that actually take away the pain from enterprises to worry about those gaps, because in our case, in that example I just gave you, yes, there are gaps, but we have done enough of an automation cycle on the business part of itself that we can tie it over the gaps. We've got to go, glad we could squeeze you in. Quick, 30 seconds. The show this year, what are you seeing? What's the buzz coming out of it? What's the meet? What's the buzz from the show here? What's the story? I continue to believe that we are in an era that will redefine what we have seen humans do. The people at the show continue to surprise me because the questions they've been asking over the last eight years have got a slightly chin. I'm just done with buzzwords. I don't pay attention to buzzwords anymore, but I see a maturation. I think I said it before. I see more bald heads and big pates. When I see that in shows like these, it gives me hope that when people who grew up in a different S-curve have borrowed a new S-curve, the pace will strengthen. So as always, phenomenal show, great community. The community's changing and looking at less, are you different or in a good way? Well, we feel your pain in the buzzword as we proceed down this epic digital transformation over the top, 5G, autonomous vehicles, big data analytics, moving the needle, all this headroom, future-proofing, AI, machine learning, thanks for sharing. Thank you so much. All right, more buzzwords, more signal from the noise here at theCUBE. I'm John Furrier with Jeff Frick and George Gilbert, we'll be back right after this short break.