 Live from San Jose, California, in the heart of Silicon Valley, it's theCUBE. Covering Hadoop Summit 2016, brought to you by Hortonworks. Now, here are your hosts, John Furrier and George Gilbert. Okay, welcome back, everyone. We are here live here in Silicon Valley in San Jose. This is theCUBE's Silicon Angles flagship program. We go out to the events and extract the signal from the noise. We're wrapping up day one of three days of wall-to-wall live broadcast coverage. We are live in San Jose for Hadoop Summit 2016. I'm John Furrier with my co-host this week, George Gilbert and Peter Burris, who couldn't make it in the meeting. Guest analyst is Abhimeta, CEO, founder of Triseta. Welcome to theCUBE, filling in as our guest analyst on our club. It's great to see you again. You guys gave me a promotion to guest analysts. I appreciate it. Well, usually it's for unemployed entrepreneurs that come on or sold their company. It's kind of like they're not on the sidelines, coaching or playing, so they come on. No, but seriously, you're the CEO of Triseta. For the folks that don't know, Abhi was on theCUBE seven years ago, the official second one was the first time we were there at theCUBE. Hadoop Summit, the real world didn't start. Hadoop Summit seemed around. That's right. You were for Bank of America at the time. Now you went outside your own company, Triseta. Now we've lived through this. You've worked in the big company, financial services, where at that time you had clear sign of blind sight on analytics for fraud detection, a lot of things. Went out on your own, started a company. What's the status? Give us an update. Well, first of all, thanks for aging me for your viewers seven years. It's been a long time. And it's so good, always good to see the success you've had as an entrepreneur and the success we are seeing as Triseta. So a quick update on Triseta. We have nothing but good news to share. I think a big thing that is happening in the market today, especially with enterprise software, more so than B2C, but B2B software, is this ability for entrepreneurs. You call them tier one entrepreneurs. We finally have an ability and a arrow of an innovation economy where you can build a company that is capital light. What I mean by that is you can actually build a company without raising oodles and oodles of venture capital or private equity money built on revenues, which is the secret of Triseta's success. So we at this point have funded our CD's A and CD's B on organic revenues. We had become profitable two years ago. Last year was the first year we hit our striil as operating company. We hit our operating numbers, gross profit, net profit, how to pay taxes, which my lawyers and my accountants tell me is a fifth avenue problem. I shouldn't worry about it, much as much as it hurts. We are a contributing member of society now. We're not a company losing money. But more importantly, I think we have proven the vision that we have shared with you, that this era of data powered economic models is going to unleash a wave of business transformations, not just technology transformations, that is going to empower billions, right? I think you were the first ones to say that the economic value at stake is not billions, it's in the trillions. And we are seeing the early successes and the early signs of that unbelievable revolution. And you know, we'd love to share the data that's certainly live, and now Facebook does doing a lot of live, so it's good to see Facebook kind of copying the cube model, which is great, you know, they're worth billions. Now, all kidding aside, we've been successful, you've been successful. Congratulations. I want to get your take on it because you said something in 2010 that it still holds true today, data capital, data factories are going to drive the innovation and competitive economic advantages. I don't think you used the word valuations per se, but you were intimating, certainly talking about the value of the data that's going on. LinkedIn sold to Microsoft for $26 billion. You know, some would look at LinkedIn and say, hey, that's a Rolodex, that's a job board, but not really a social network, but they had unlimited, high quality data. They're not unlimited, high quality data. Absolutely. That's a showcase for this ecosystem we're living in. Question for you. Did Hadoop fail on its promise to build an app economy? Not at all. I think, and I love when you asked me to make controversial statements on theCUBE. Absolutely not. It did not fail. It succeeded. It is unleashing. Let's take a step back. Let's take a step back for a second. What we are seeing is a dramatic shift in S-curves. And when you have dramatic shift in S-curves that are driven by some tectonic underlying change, which in this revolution happens to be technology around data analytics, it takes time for the entire ecosystem to evolve because we are fundamentally rewriting business models. I know you guys believe that big vision, right? We are fundamentally- You can get more step up and value- Absolutely. With data. With ever low investments, right? With markedly lower investments. Well, the question is, do you shrink costs or increase value? So you can actually increase your value faster than you can increase, lower your, I mean lower your cost. But what if we came over here and we proved to you that we can do both? Yes. I think what this ecosystem is proving to us is you can do both. That you're enabling an innovation economy on a new S-curve. Give an example. So let's do a quick example. The one with this ecosystem doing that. Let's actually use the example that you use and we'll use two examples. We'll use an enterprise software example, which of course I'll talk about Trasada. And we'll use a data company example, which we use LinkedIn. LinkedIn and the value of LinkedIn is way beyond the social network that LinkedIn built. But we have to think about LinkedIn as creating a new data asset that never existed before, which is one single global location of not just professional data, but professional hierarchy and timelines. There is no company in the world that if they wanted could predict looking at global talent pools, global education certifications and global professional lineage, what people could do in the future with their careers when they start. You marry that information, that plethora of knowledge of a pathway of careers that is now owned by a behemoth in professional software, aka Office, and you can do not just sell more office software, which all the analysts looked at, but I think you start creating new products and services geared to make professionals successful in the pathway. That is the poster child of, in my opinion, because it's so sexy news-wise. And the number's big, of the value of data. George, I want to get your take as the analyst on the Wikibon team. He's really talking about, I was talking about the trends in the meta Uber business economic side and the entrepreneur side. Under the hood, okay? Are you happy with this ecosystem? Do you think that the engine that's in the car, so to speak, is working? I mean, obviously Hadoop was supposed to be the holy grail, it morphed. What's going on under the hood? What are the key innovations that you're seeing? Is there a new kind of component? Is there a turbo-charge engine? I mean, what's going on? I'd say 2016 is big data's moment of truth. And by that, I mean, we've assumed for several years, for more than five years in the commercial market that Hadoop was the big data platform. And by using the word platform, we were implicitly saying, and actually often explicitly saying, applications were coming and that they would reside on that platform. But the problem is the governance model, Apache Software Foundation, and then the companies that made distributions of Hadoop off of what came out of that foundation created these platforms that innovated in so many different directions so fast that there is no one platform. We can't say it's like Linux and there's just a little bit of difference in each distribution. You cannot build applications on Hadoop. You can build one on Hortonworks, you can build one on MapR, you can build one on Cloudera. And I think what's going to happen is and we were hearing this from our guests all day, that the very services that define the core of Hadoop are being swapped out, like instead of HDFS on-prem in the cloud, it's S3 or some other file system or object store. And so what we're seeing is, it's almost like a disarticulation of the Hadoop components. So, okay, I'd buy that. But to the question of what's working, what do you see working on the moment of truth on the public side? What's working is the principles that Hadoop stood for, which is huge amounts of storage where you don't have to decide upfront exactly how you're going to use it, the ability to send the compute or the analytics to the storage, and then sort of the ability to, well, expand in the elastic capacity and pricing, a metered pricing model. Those things. So- The two other things that are working John, I'm sorry to interrupt, but the two other things that are working that are dramatically important for us to talk about. I think the oil tanker has been shifted to, Hadoop did win the battle to become the de facto data storage platform for the industry. And I call it the HDFS ecosystem, number one. Let's talk about- Not MapReduce. Not MapReduce. HDFS, right? A distributed file system. Great place to store the data. Cheap, scalable, redundant, and works at massive, massive scale, irrespective of structure of the data assets. Hold on, hold on, hold on. Before you go to the second thing. Sure. Just interject. Do the cloud guys care about Hadoop? I think- Amazon, others? I think the reality in my opinion, having seen no one cares about infrastructure anymore because of what happens with infrastructure, which is at some point, when Cloud Foundry, Docker, Amazon have abstracted away the complications running applications at scale, no one's going to care about them. Okay. They'll make a lot of money about it with it, but no one's going to care about it. The second thing that's very important on the app question, because we've debated this for five years or six years now. Yeah, let's go again. Is there's a reason why we haven't seen a plethora of applications? And I know why you were turning back and looking at it. And here's why. What has happened is, we see the emergence of three tiers of new platform services emerge in the future of the enterprise software ecosystem. I would love to get your views on it. Sure. The bottom tier is the server operating systems, right? We call it the SOS. The historical operating systems have been Linux and Windows and the next generation are Cloud Foundry, Docker, et cetera, right? Then is the data operating system, which HDFS is the next gen of what has replaced what we had grown up with. And HDFS is absolutely one that battle on the data operating system. There's a critical missing component of writing applications because the data operating system is not building, the building of the data factory to enable application writing. We have added a third tier to it. We call it the analytics operating system. We have our own word at Trasera Optimus, and it's building the core functions necessary to arcaded apps really quickly. Without the analytics operating system that works on that tier, you can't write the apps. And that's why you haven't seen anybody but Palantir on a 15-year cycle building what they call Gotham, which was an analytics operating system, to set up with Optimus and IBM with Watson. And that regeneration cannot happen without that missing tier. And no one in the data operating system tier has focused on it, because to your point, when you look back, they're building functions that are essential to store and for them to be able to build a marketer on. All right, so here's my take on this. So, I mean, pretty dogmatic, obviously it's your company doing the analytics operating system, but here's a trend that would support that thesis. I've been kind of teasing it out in the cube and it hasn't really kind of come out. So I've been stitching it together in public in the open. But here's my take. What DevOps was was infrastructure as code. And that was like, wow, are you crazy? So that created LinkedIn, Facebook, all these web-scale guys, Yahoo, Cloudera, Hordemarks, the DNA of this community, they had to build their own stuff. That was called DevOps. Eatin' glass spittin' nails is Dave Vellante and I always talk about that hardcore dudes. Well, DevOps just went mainstream, in my opinion. So with Docker containers, the application market, you're seeing DevOps is now a paradigm that's gone mainstream, which means the ITOps group has to accept the new narrative, the new reality, adjust operations to be agile. That's its own little challenge. That's a done deal. That's what we're covering at DockerCon, Red Hat Summit, and some extent, not so much here. But all the IT shows. But if you believe what DevOps did, maybe infrastructure as code, you could even take the same leads, move up the stack and say, data as code. Yes. So you'd argue that the benefits of a developer to say, whoa, I don't want to provision any hardware, I want pools of resource, be elastic. Of course. That is what developers wanted and that's what DevOps did. The next level is I don't want to deal with all the data, I just need data. So programmable data, data as code. Machine data. Machine, it could be machine to machine, augmented by AI, but my point is that's a new data ops layer. Yeah, it is, absolutely. So, data ops, DevOps is data ops. You see that? I do. Can you share your thoughts on that vision? Absolutely, and I think, I like, that's why I like being challenged by you, because you're not using different terms, but the data ops layer, what you call program data, what I call machined data, machine data is the automation of data ops into its own operating system. There are fundamental building blocks needed to architect analytical applications that solve business problems. And the goal is to make the developer go faster. And in some cases, abstract the way the developer itself or herself. Because the reality may be, let's take a step back. You look at the three big economic engines of an economy, agriculture, manufacturing, and services. And when each tier got machined, AKA automated, people moved to the next tier. We have lived in a boom of services with humans doing all activities around every single service, whether it's a data service or a product service. And we are now witnessing the era, the start of an era, where we will machine services and remove humans from it. With your terminology of yours being data ops. Data ops will be automated, but the core functions around it are... Data ops with bots. With bots. And the bots, in this case, will be not like the fuzzy bot running around and jumping rocks. It will be software. Yeah, software help make my life easier. Absolutely. So this data ops is going to enable a new class of developer, the machine developer. Absolutely. It's Terminator. It's Terminator. It's Skynet. It's Arnie. I mean, it'll be great to get Arnie to talk about. But you're absolutely right. I think we are witnessing the starts of artificial intelligence, not the scary part of artificial intelligence, but the ability for humans to program machines to think like humans. And you asked me for an example. I'll give you a very good example. Money laundering. Massive problem in the industry. Two trillion dollars. After every single automation, every single machining, every single technology innovation after 15 years remains a two trillion dollar problem. Which today... They fixed the X trillion dollars to only a two trillion. No, it's been growing. At 15% compounded every single year. So it hasn't even gone down. So the question you are... So they stop on the bleeding with their current solutions. And if they want to, because they keep screwing up part of my French and the regularity coming after them. So the question becomes, if we are so intelligent, if machining is truly coming, if Terminator is truly coming to data ops, why haven't we fixed the problem? And we took a crack at fixing the problem and we realized that by simple machine learning, as step one, and complex unsupervised learning, AKA AI as step two, we could eliminate 90% of the false positives and completely automate the human process of running down a suspicious transaction. It wasn't possible even when we spoke seven years ago. So to answer, now to bring it all down, it wouldn't have been possible without the HDF ecosystem that you have so passionately supported yourself, including companies like ours, because without that, the economics of doing so didn't exist. So absolutely, not only can we build Terminator, every single human being, every single family, can build their own Terminator to make human life better. Well, a final wrap up on day one. We've got two days of more coverage. Tomorrow will be day two. Got a great lineup, more great guests. Abhimedas filling in for Peter Burris and Dave Vellante. He's so powerful. We're filling in for two answers. Of course, George Gilbert and I'm John Furrier. You're watching theCUBE. We'll see you tomorrow here, live in Silicon Valley in San Jose for Hadoop Summit 2016. Thanks for watching and we'll see you tomorrow.