 at Big Data SV 2014 is brought to you by headline sponsors WAN Disco. We make Hadoop invincible and Actian, accelerating Big Data 2.0. Okay, welcome back everyone. We are here live in Silicon Valley. This is the Cube, our flagship program. We go out to the events, extract the signal from the noise, talk to the thought leaders, the tech athletes, the people making it happen, entrepreneurs, VC, CEOs, anyone who has knowledge, we want to share that with you. I'm John Furrier, the founder of Silicon Angle. I'm joined by co-host Jeff Kelly from wikibon.org, just published a Big Data Survey. Again, on the path of $50 billion market, total available market. And our guest here is Cube alumni. And been there from the beginning is Abhi Mada with Truseta. Abhi, welcome back again. And another Hadoop world, we'll call this Strata. It's in California, Silicon Valley. Big Data SV is the event here. All the Big Data happening in Silicon Valley is about innovation and no other, the best person to talk to than you about that. Being an entrepreneur, building a business in this market, challenging technically, a lot of business growth, what to shoot at, there's so many opportunities. Welcome back. Well, thank you so much. And I'm especially proud of coming back today because this is our first sponsorship of the Cube. And hopefully we'll keep this on and continue it. We've lived the Big Data journey together. I'm incredibly proud to be here again. Thank you. And you believe in the Cube. You saw, you were one of those magical people who saw what we were doing with the Cube, saw our mission of opening up the data and the underwriting support. I know anything is always appreciative. And our goal is to have that underwriting support from Truseta and folks like you guys for our mission to get the data and share it. So thank you very much. Absolutely, anytime. So let's talk about what's happening, right? Obviously Hadoop is here to stay. That's the theme again. Hadoop is real, it's in all the customers. Big Data just isn't Hadoop. It's a lot of other things hitting the stage. You got legacy vendors, you got data science. It really is an innovation party here at Big Data SV. What's your take? And what are some of the new technologies? Like we see Spark, which I know you're really interested in. What's happening? What is your take on the current market? I think the take remains the same, John, which is the thing I'm the most proud about, having built the business the way we have. You know, and there's this interesting metric where people equate success to how much money you've raised. I equate success to how much money we have made. Our customers are funding our CDSA. We had an awesome year. The report that Jeff released talks about the revenues we were able to hit. And I think the theme hasn't changed, which is the market is seeing the light of what you, Dave, and I spoke about four years ago, which is data factories, the ability to actually monetize data in ways never possible before. But the ability to do that depends squarely on domain knowledge and science and technology coming together seamlessly, right? What we have since called advanced analytics applications. That hasn't changed. I think the era, you made an interesting point. Big data is a bigger ecosystem than just Hadoop. Completely agree. The reality though is the era of building and commercializing tools is fast behind us. You can't, if you're building a database in the big data market, don't worry. Shut your company down. BI is commoditized. Democratization that everybody talks about and commoditization are two sides of the same coin. And as big data technologies democratize, they will commoditize. The only way to extract value is going as high up the stack as possible and finding problems to solve with big data. That hasn't changed. And as much as you push people, I love that you push people, to talk to you about that, I am disappointed that the market has moved slower in there. But I'm excited because it's moved slower, I have success. So it's this interesting, interesting part of it. I'll ask you, because again, this is an interesting point. The commoditization angle is great. So we were at the Open Compute Summit just recently here and there's a revolution happening at a hardware level. It's a whole nother conversation. It's like the homebrew club of the future. The modern version of what's going on with the cloud hardware. But the comment we were talking about there is someone said on Twitter to us during your crowd chat, if you're going to commoditize infrastructure and you're going to commoditize the middleware and you're going to commoditize the application, what's left, where's the value? So the conversation was interesting. At the end of the day, you mentioned value. So value can never be commoditized because at some point people will pay for it. So if the stack becomes commoditized, where does the value shift in your opinion or can apps never be commoditized? No, I think it's a good question. I think, again, if you have to break it down into its elements, we have to realize as an industry that companies that are hundreds of billions of dollars in revenue on a stack that includes databases, storage, BI, analytical software are under attack, under massive attack, both with their own legacy stacks, as well as this revolution called open source. The track record isn't hard to find. Lilix did it to one of the biggest technology companies for one of the smartest guys on the coastline in the world. So the track record is there to see. The question that you ask is, does commoditization keep going up the stack? The answer is absolutely yes. However, that's why I don't think the big data market is a technology market. I think you've heard me say multiple times. This is a business revolution. Where you add value is in technology, is not data, is not science, is the combination of the three. So when you go to a customer, when you go to a bank, when you go to a retailer and you say, look, with this massive explosion of data, if you want to understand customer behavior, if you can understand customer, can you make money on it? Everybody will say absolutely yes. But understanding customer behavior is an incredibly hard thing to do. It's a combination of bringing in open data, open tech, and open science together. When you bring those three components that are fundamentally open together, you build something that's unique. That's where the value is. That can never be commoditized because the expertise that Tristata has in financial data is very hard to replicate. And no one will replicate exactly what Splunk has done. So yes, commodization will rapidly expand up the stack. We have to remember that the value at stake is $2.4 trillion, right? IT, IT is a $3 trillion market. 80% of it is enterprise software, right? Something that you guys cover and we are in. We can't forget that. We can't forget the value at stake. Isn't the trillions, not in the billions? And we've spoken about that quite often. Let's talk about the business revolution. I want to get your take on this. You mentioned that if you're doing a database tooling, you should fold your business down and shut down. But I want to ask you about it. The current crop of startups, are you impressed with them? Are you seeing some goodness there? What are you seeing in the market around some of these startups? Are they solving hard problems? Because the enterprise market has always been a hard market, right? But now what does that mean? What does enterprise mean? It's all business. It's a business revolution. That's the enterprise. You got a consumerization angle going on here. We got the consumerization of IT and a business with mobile, data, tsunami everywhere. So it's that enterprise problem is still there. No one can just say, hey, I want to be an enterprise entrepreneur. What are you seeing in the crop of startups here in the big data world, the strata conference? It's interesting. I love the fact that there's so much confusion and lack of quality startups. Because it, in a way, helps to set out in an obtuse kind of way. We saw the market so early, it is actually shocking to me that there are three clear trends, John, that we all have seen. And it surprises me when I ask people like yourself, I ask everybody, so where are the interesting startups solving business problems in big data? Tell me, because I want to learn from them. The answer, the room goes silent. The three trends are clear. Trend number one, the days of selling databases for hundreds of millions of dollars or analytical tools is over. The enterprise buyer is smart enough to realize that algos are free. BI is free. Database functionality is free. But he's struggling with it, trying to figure out, how do I take those open source components and find value, right? That's number one. So enterprise, we've got to give credit to enterprise buyers that they're smart enough to realize that open source works at scale in the enterprise, right? Not just in the operating system level, all the way to analytics. Number two, the buyer itself is changing. You're not selling to the CIO anymore. You're selling to the CMO, you're selling to the CRO, you're selling to the CFO, you're selling to the CEO. I was with the CEO of a large bank in North America talking about big data. And he sits there and he goes, you know what? This is important enough as an initiative that for I need to run it. Not my CIO. I personally am going to guide this conversation. How many startups do you know who have talent and people who can sit and have a CEO conversation on how big data will revolutionize their business? Nobody. That's what fundamentally differentiates, the domain knowledge differentiates, we say that. The last and most important part is applications I hate to say. There is a suite of software we're calling Next Generation Predictive Analytics Software. They are delivering actionable intelligence at scale at a segment of one for a variety of problems that you can monetize tomorrow. You know the funniest part was when we tried, we have trademarked the term monetize big data. I was shocked it was available. Because why is someone not talking about monetization of big data? It's all around big data tech and big data investments. So why aren't they, why aren't it starts working on it? Is it just lack of awareness? Hard. It's incredibly hard. You have lived the Tristella journey with me. Tristella was announced at theCUBE, right? Three years ago. And here we are with $10 million in revenues funded by our customers. It's incredibly hard. It's incredibly hard to combine domain knowledge, technology expertise, and the leading edge science to solve problems. Not easy to do. And when you've tried to focus and build a company on tech, and just domain, and just science, individually, I'll kick butt for all three of those guys, right? So I think it's just not easy to do. It's not easy to pull together business leadership, thought leadership, with deep technical expertise, and the best algorithms in the world, and say, I'll apply this particular solution to solve this problem. A great example is, now that we've had the initial success, and I always say, we're still changing the odds. I like my chances. With initial success, we are picking up a audacious goal to go after every year. So this year we picked fraud to go after. Financial fraud in the world is a $100 billion problem. Even if we together made a 10% dent in it, that's $10 billion. Why isn't anybody applying big data to fraud? Because I think the challenge becomes, do we understand the elements of fraud? Do we understand the network? Fraud is a network problem. Can we do it real time, right? A big announcement we made this week was announcing the first real-time graph traverse a network discovery engine we call NET on Spark. So can you actually take those capabilities, deliver them real time for massive problems by applying frameworks like energy-based models, which are brand new artificial intelligence frameworks in big data. That is just hard to do. Well, let me push back on that a little bit. So it is definitely difficult to do, but why can't a company like IBM do that? They've got all this technology. They've got the services arm. They've got the research arm. Why haven't they been able to do it? But I think, you know, I think IBM gets a bad rep for the wrong reasons. Smarter Planet is, in my mind, by far the smartest marketing genius campaign ever. And I think IBM is trying very, very hard to do exactly that, right? Their big push into healthcare, their big push into looking at disease diagnoses and prevention is a step in absolutely the right direction. I think the whole picture on Watson and Margini came out and said the big investment they're making is set in the right direction. So I think I actually don't believe that no one else is doing it. I think the reality is, what technology stack do you build it on to deliver advantages of scale, speed and accuracy at a cost that's feasible? So if I want to build a platform to eliminate financial fraud in the ecosystem, it needs to work across four billion different nodes on the planet. So I think that's where IBM's technology will, I think, evolve. Because what they're lacking is the scale and the speed, not the intent. I think they actually, absolutely understand the smarter planet is the future. So they've got that legacy business and that legacy technology, which, you know, it's the innovator's dilemma. Do they, they don't want to give up too much of that revenue, of course, at the expense of some of the new things that are happening, but they're going to have to make this transition if they want to survive and thrive. Well, that's our bet. I think the reality is that this is a Darwinian moment in our lifetime. I think I've said multiple, this is a 50 year boom cycle where data as the fuel, right? We spoke about this four years ago. Data as the fuel will fundamentally transform how we measure, evaluate, and act on consumer behavior across a host of problems, from education all the way to fraud or disease prevention and disease management. And I think that journey will take time, but the ecosystem is rapidly, rapidly evolving to solve exactly those things. So I think that, I agree with you. I don't think that they will have a choice but to jump on the band line. So another thing you mentioned that struck me was talking, you've got to be able to talk the language of business and not necessarily the language of tech. And that's a challenge, I think, one of the, to John's point about why aren't some of these startups having more success? And part of it is because they speak the language of tech. They don't necessarily speak the language of business. They think, well, we're going to build these technologies and we'll put them in the hands of customers and they'll figure out how to use them. And that's a hard thing to do because frankly, we've seen a lot of the enterprises don't know yet how to do it. They're begging for leadership and thought leadership and some direction about how to use this technology. So talk a little bit about how you go about building a company who can, you know, not just the sales but the whole company who can speak the language of business while at the same time bringing in the science and the tech. I mean that, as you said, it's an incredibly hard problem. How do you go about finding these people? Because I'm sure others would love to find similar people. Absolutely, not giving any secret sources away but I think, you know, building a company is hard and you guys are entrepreneurs yourself and we have learned a lot of lessons along the way around and we've made mistakes. We made this new transition last year where we made a decision that a bottom third of our talent, we will rotate out, right? To the bottom third of our talent, we will proactively rotate out because the only way we can achieve our ambition to be the next billion dollar company is if we have the best people on the table having the right conversations. So what, the way we solve the problem, we put data at the center of every conversation. So when you come to me, if you come to Truseta and say, well, let's think about understanding how you can underwrite risk or think around how can you deliver omni-channel offers at a segment or one. We always start with the data and get such a deep understanding of transactional data or social data or interaction data, understand what elements, what triggers, what observations can inform the right answer before we do anything else. So they bring data to the party? They bring data to the party. But I think we will have to, I'm also surprised that we're the only big data startup with a chief data officer. So when me and my co-founder, you know, starts putting more structure on the company, the most important function in a big data company should be the chief data officer. You know, so why is it that while every single financial services company, every single retailer, every single healthcare company announcing chief data officers, the startups in the space who are trying to monetize data and don't have a chief data officer is the most important function. So I think that's a key, key tenet. Bring data to the party because the tools exist, the technology exists, the real value, the secret sauce is this beautiful, magical combination of data coming together and making love for technology and the algorithm. If you can pull that together, you know, tomorrow's Valentine's Day, so we'll keep going with your love and allergy, the answers are beautiful and wonderful, right? You will propagate little minions of insight, like there's no tomorrow. So I think that's the secret sauce to say that we always start with the data. If we want to be the only company or the first company to truly understand customer behavior, we have a read on customer data like no one else, like no one else in the market today. So let's talk a little bit about the announcement you alluded to a moment ago. So you guys have built a new application on Spark, which is getting a lot of attention at this show, kind of, you know, we've heard from Cladera's made an announcement around Spark. Tell all our viewing audience a little bit about Spark, what it is and kind of what it enables and use your application kind of illustrate some of the power of the technology. This is great. You know, that's why I love coming here. I mean, go from business to deep tech. Yeah, we're all over the place, isn't it? This is great. We covered the Spark Summit this year. Oh, awesome. With the crowd chat and really intimate group. Amazing technology. Absolutely. Charmele Mulligan from ClearStory. Absolutely. Supporters to go there. Amazing. Amazing ecosystem. Yeah, I want to drill down on this a good. Absolutely. So I think, again, we should take credit for this. So when we spoke in New York at Big Data NY, you guys asked me what I'm at the most British about, right? And we did the news as to, I said Spark, by far. And the reason I was the most British about this was, two years ago when we had the conversation, I used to get asked the question, well, Hadoop is great. It's awesome, but it's batch. And I said, don't underestimate the open source community, right? Don't underestimate the power of the common man as they say. And Spark was the genesis of that. So it's been a project that we actually have been following and contributing to for 18 months. And it brings to Hadoop an in-memory real-time computational capability. So Spark is, in a simplistic way, an in-memory database-like framework that allows algorithms like ours, advanced algorithms to run in memory at scale. The cool part about Spark is, it works natively with HDFS. So unlike other in-memory systems and engines, where you have to go and repopulate data, they're trying to find insight on different populations, Spark will do this dynamically. There's something that they're proud of us about. So we have this in-memory graph traversal engine at scale. So say you're looking at a particular segment, a particular cluster. Say it's the telecom cluster. And now you want to say, well, I don't want to look at the telecom cluster. I want to look at the energy cluster. In traditional systems, you have to go back to the database, resample, recut, wait, hours, days sometimes, and get the information back. Spark does it natively from memory to disk real-time. So it completely changes. So we were asked the question often, well, why is it that you guys are building this beautiful ecosystem and beautiful software? Why don't you build visualization? And we said, we would only build visualization when scale is the insight. In graph network application, which is the killer application in big data, graph today is scale is the insight. So that is a easy to use, beautiful visualization tool, already deployed at a massive bank to understand money flows across the world at scale. And Spark is the framework behind that. So I got to ask you on the analysts, you guys are just a great job. Where are people missing the boat on analytics? I mean, I'll say some great examples of greatness, but in your opinion, where are people, where's the blind spot for the folks doing analytics? You know, it's this question, it's a great question. I was debating this yesterday. The term that I've grown to not like in big data is this way it was being defined in the past. There's two terms I'll critique today. One is allowing people to ask questions they did not know how to ask before. Well, that's kind of yes. The questions are very well known, right? It could be anything from if a particular, if a particular supplier of yours goes bankrupt, right? What will the impact do you be? You know the question, you got to find the impact, right? So it's not like you're searching in this mountain of data. Maybe it's a question they couldn't ask before, is the right one, because they all know their question. Correct. When am I going to go out of business, when am I going to lose a customer? Exactly, right? The language of business. Absolutely, the language of business is quite simple, right? Every business has three core goals, right? Get customers, grow customers, keep customers. You have that, I have that, everybody else has that, right? ExxonMobil, Walmart, they all have the same goals. The questions are already very well known. How do you apply new, bigger, better data? So newer data, bigger data, better data. To those existing existential questions for every corporation is what people should be worried about. But people don't talk like that. I come over here and I see these beautiful visions and rainbows, but the pot of gold is missing. And I go, well, the pot of gold exists if you ask the right question, if you know where to look. It's a medical marijuana cards while you're out here too. They're smoking a lot of, you know, Z-er sense. Silicon Valley always has always been dreamy in the sense of inventing the future. And that's good. And you gotta dream the future before you can invent it. Absolutely. However, at some point. You gotta apply the dream. You gotta make money, right? Because if you keep dreaming, you won't get married, you won't have a house, you know? You'll be sitting here doing something fun, smoking pot, I guess, you know? But I think the second thing is this misnomer. Seattle and Denver, they won the Super Bowl. All right, I'll go. I'm getting big thumbs up for a smoking pot. Getting back to subject. The second big issue that the industry needs to realize is this wrong focus on lack of talent. I don't think we have a lack of talent issue in data science. In fact, data scientists have become the most confusing term ever invented by man for any profession, you know? Because what does it mean? What does a data scientist do? Unfortunately, when you call somebody a data scientist, they chase the sexy part of data science and not the real value of data science. We believe the challenge that we have as an industry is not to create the talent. It's to actually deliver intelligence to existing talent to make better decisions. So I wanna ask you on the talent point. So this comes up a lot when we talk about hyperscale, right? Facebook, Google, Amazon, huge hyperscale, talk about scale, they're at scale and they're kicking butt doing it. So the issue was why didn't they just buy Oracle licenses? Why don't they just buy off the shelf software? Well, because they had smart talent that didn't wanna pay the money. So they used the talent to build their own, hence the history of what's going on with HBase and Hadoop and Facebook and other environments. So the question is, everyone's saying, I wanna be like Facebook, I wanna be like Google, I wanna be like Amazon. Build my own computers, write my own software. But the problem is not everyone can do that. So the talent gap is huge between the mid-range hyperscale, the enterprise. So how do you talk about people saying, I wanna be like Facebook, but the reality is they just don't have the talent and might not get the talent. But here's the good news though. I think what people forget is this is the beauty of open source. The world is developing for you. You can be like Facebook. Guess what? Facebook's entire backend is this thing called Hadoop and it's open source. You can absolutely leverage it. What is lacking? What, the way you drive the hyperscale growth, the way you transform businesses, is there need to be more companies like Trisada that are making existing talent more intelligent by delivering them the knowledge. That's what's missing. If we keep talking about data scientists and their 1.5 million jobs that are waiting to be filled, we'll always chase tail because five years from now, the talent gap will look different. What the tools are enabling us to do, what Siri is enabling us to do, what Google now is enabling us to do is take people who are not the Einstein's and give them actionable intelligence and make them smarter. When we can build tools to make the population make Earth a smarter planet, no pun intended for IBM, that's the way to win it. So the only way you'll get hyper growth is you can automate it. You can absolutely, as I've always said, machine learning can automate 99% of the work done around data analytics. The last 1% is human, giving better information to you and me to make the right decision at the right time. On the machine learning, obviously machines have to learn, so you have to teach it. Correct. So if you don't have the knowledge, you can't teach the machines. So how do you fill that gap? And you guys are doing some work in this area. Correct. Obviously in memory speeds up and accelerates some compute power, but assume compute might go to zero cost and be infinite. Right. If you believe that, I mean it's a thesis, but let's just stay there for a minute. Right. Compute is infinite and the cost is zero. Where's software go? So here's how do you break it, right? There are four key components to building analytics application, right? Data, so you gotta store it. You gotta make sure you can catalog it, right? Which is what databases do. Then you gotta make sure there are algorithms you have available to find the insight, right? And the last piece is delivery of the insight. You gotta make sure you can actually act on it, right? Decision tools. Let's look at those components. Storage, free, open source, commoditized. Data processing, free, open source, commoditized. Algorithms, free, open source, commoditized. In fact, we're talking to a large customer. He says, you're the only company, Avi. Trisita is the only company that I have seen and who has been able to maximize the algorithmic power in Hadoop. Can you give me the underlying algorithm to your application? We said, sure. We have an open source product called Gneth. We're open sourcing them. Go get them. They were into pay us for it. We said, no, algorithms are free. Why should you pay me for it, right? The last part is decision tools. Decision tools have two components, technology and human. That's where we play. We are building the next generation decision tools to enable intelligence by combining all those elements. So that's where it comes in, John, is the open source community in the next 50 years is going to dramatically commoditize and democratize access to intelligence tools for the world, not just the enterprise. We were talking yesterday around everything being free. How do you make money? So there's the customer support and the red hat and Horton was just doing that. And one of our guests from InfoObjects, Rishi said, the get effect, the network of trust. So you have a trust equation now developing. What's your vision on trust? Because what you're talking about here is ultimately you're offering algorithms for free, which I would argue is good move on one end, but it might be risky, some would say. However, if you do it for free and you aren't, how do you make that a differentiation for you? Is it the trust? Is it the halo effect? Is there an element there? What's your vision on trust? It's quite simple. It is the combination of those elements that are fundamentally free, coming together to offer a business solution, right? We had a customer sit down with us and said, I said, how would you characterize Trasada? In your mind, our average deal size is seven figures. Trasada's average deal size is seven figures, right? If you do the math on how we hit the numbers, we don't have hundreds of customers. We're not setting BI software for $10,000 a pop, right? We're doing seven-figure deals. And that's based on this very simple premise, which is if I can take... There's 10 customers. If I can take, your math is brilliant, you know? No wonder I love you even more now. But if you take, John, if you take the elements around data, a business problem, combine what even we will push on democratizing. The algo isn't the secret sauce. It's the combination of the algo, the data with the domain knowledge, solving up specific problems. If I can come to you, you're a buyer, and say, Abhi, I have a problem in fraud. It cost me $2 billion a year. And I said, John, I can reduce it by 10%. And you only pay me. Of the 10% I save you, you pay me 10% off that. Do the math. You say, where do I sign? You won't ask me technology I use, because you don't care if it's a do or not. You won't ask me if the algo underlying it is open source. You will say, if you can prove to me that you can take my burden. And typically, so it has a metric field. Typically in a large enterprise to manage financial fraud, it takes a combination of an act of God, thousands of people, and some software with 98% false positives. If I can just walk in and say, I'll drop the false positives to 90%, you and I will sell that software, sitting in our drawing rooms in our pajamas, with no sales, with zero sales, because who will not come and say, that is value for me? You have to figure out what are the pain points? Where is the value delivered? And as long as we can do that, you can extract a piece of it. That gets back to what you were saying earlier. You've got to speak to the business. You've got to speak about the business problem. They don't care what the technology is. Business people will open up the checkbook if you can solve a business problem and show them savings or money that they're going to make. You know, I have a statement which made, drove me to be an entrepreneur. Life is short. You got to remember, so are enterprises. You know, enterprises are no different than our own human interactions. They all have limited lifespans. They realize it, right? Competition drives them. They don't have enough resources. They don't have enough intelligence. They don't have enough software. They don't have enough time to build everything themselves. That is the nugget. If anybody wants to enter the open source ecosystem of big data, right? And drive value from it. Drive value from it. And derive value from it for themself. Don't focus on databases, BI tools, you know, storage. You've got to be able to add value at a differentiated tier in that stack. Okay, we got a break here but I want a great conversation. Always great to have you on. It's intoxicating and informative for all the folks involved, including us. Great to share and, you know, kind of riff back like kind of musicians. It's fun. It's really great. Thank you. The final question I want to ask you is, I want you to summarize for the folks watching about this moment in history here in Silicon Valley, big data SVR event that we're having here with the Stratoconference. At this point in time, what is the most important thing that's happening that's the big story in this moment right now? I think the VR in the middle of the second industrial revolution it started five years ago. I would hate to say the time was when you and I spoke at that summit. And the most important driving factor is a massive change in consumer behavior that is forcing every single enterprise to rethink their business model. It is across every single industry. Every industry vertical is struggling to figure out how do they go from no growth where GDP is growing at 2%, how do you go from no growth to fundamentally redefining how products and services are understood, designed and built to meet customer needs and wants, not to exploit customer needs and want to meet customer needs and wants. So we call it becoming customer advocates. Enterprises who understand that can leverage data to become a customer advocate are the winners of the future. It is a Darwinian moment in time. That when we look back, when you and I look back 40 years from now and you will be probably on ESPN and doing 24-hour something fun. Curling. Yeah, curling. Check curling. I think we'll look back and we'll say, wow, I never thought a bank would look like this. I never thought a retailer is delivering products and services when they ask me before they design it. I think that is the most exciting part for me. We'll go back to the footage. Look at what Abbie met. It was just a tiny little startup. Now he's got his mansion and the next Microsoft. Well, Abbie, great to have you on theCUBE, as always. Thank you so much. Thanks for hanging on. This is theCUBE. We'll write back here. We're live in Silicon Valley here in Santa Clara, the Hilton Live, broadcasting Big Data SV and the Stratoconference. We'll write back.