 Live from the Mandalay Convention Center in Las Vegas, Nevada, it's theCUBE at IBM Insight 2014. Here are your hosts, John Furrier and Dave Vellante. Okay, welcome back everyone. We are here inside theCUBE live in Las Vegas for IBM Insight. This is IBM's premiere show around big data analytics, formerly called Information On Demand Now, rebranded as IBM Insight. I'm John Furrier. I'm with Dave Vellante and our next guest is Steve Mills, Senior Vice President, Group Executive, Software Assistant with IBM, Industry, Legend, Leader of the Helm, Captain of the Ship, welcome back to theCUBE, good to see you again. Always great to be in theCUBE. We love chatting with you. It's like, yeah, tech athletes roll on us, please. You're on the day to day, IBM is still on course. I mean, the shift is not happening. You guys are staying on vectoring into this market, cloud mobile, software descent and the value proposition. This show's focused on the big data analytics. We talked to you recently at Pulse. What's different? And just quickly break down the core thing between this and the other event, Pulse, and now Comical Interconnect. Well, obviously this event's very much focused on analytics, information management, so whether it's data warehousing, database, big data, little data, structured, unstructured, SQL, no SQL, SQL on no SQL, every conceivable combination of technology associated with all things analytics, right? And so that'll pick up all kinds of areas, including things like internet of things, which is very much a big data challenge. Security, these days, cybersecurity, big data challenge. Fraud, any money laundering, obviously into sentiment analytics and customer relationship. And so it's a very broad range of topics, but they're all tied back to this whole area of analytics and data. So one of the things we're seeing in the theme here is speed, large-scale analysis software, right? So this is you're in the software systems, you're the head honcho there. So talk about what's going on with the cloud dynamic, okay? You've got large-scale computation available, softwares at the center of the value proposition. The new models are emerging. Ray Wang calls this area of insight around the future where IBM is going to really shine with their software. So you're leading the software team at large scale. Cloud is large-scale. What are the things that are in place and where do you guys see the leverage for IBM? Well, you know, specific to cloud, cloud is obviously a dimension of what the market is doing today. And it relates to where do things run? Where can I run things? And can I, through leveraging cloud-related deployment, can I get better leverage on economics? Can I move faster? Can I be more agile? I mean, there are a lot of different attributes that customers are seeking as they talk about the cloud. I think this morning in the main tent, we talked about the new announcements, new things. And there was a lot of focus on analytics in the cloud. You know, we talked about the data warehousing, our dash DB initiative, Watson analytics, things we're doing around data refinery and data cleansing, big issue for customers. And to what extent can shared cloud services be a way in which customers can more rapidly get at these capabilities, solve some of the underlying problems and get to analytics? In a consumption model that they want, which is cloud-like as a service. Yeah, it's certainly many want that. But by the way, they want a lot of things. Yeah, our customers want it. You know, that's the definition of your customer. We want quality. Thankfully, they're coming to you asking for a lot of things. So they want in the cloud, they want on-premise, and they want hybrid, something in between. Where, well, I'm not sure I want to move all my data to the cloud. A regulated company, my data's important, maybe I'm not quite ready to put the data in the cloud. Can you give me analytics capability? Can I leverage cloud? Are cloud services in some way? Can I keep my data on-premise? Can you give me a hybrid? Can you overflow? Maybe the data is not as critical from a regulatory perspective. Which case, how much analytics could I push out to the cloud? We have customers that are very interested in leveraging our grid technologies, platform software, doing Hadoop and big data analysis, using those technologies, ideally suited to a large cloud-based grid type of thing. So lots of scenarios and this ability to mix and match and pick the way you want to go is one of the unique aspects of the IBM portfolio. I wonder if I could ask, you had talked about earlier the information management data warehouse, integration, Hadoop, NoSQL, SQL all coming together at this event, and it seems to be coming together. Why is that? Is that because of the cloud? Can be this transport data is a transport? Normally these things are very, or historically, very fragmented for the horizontal layers. Is that changing? Well, first of all, our clients are telling us they want to be able to look at it holistically. They don't view what is available to them necessarily from the perspective of any one mechanism being the perfect answer to their business problem, but rather it's the combination of technologies. And look, we're uniquely positioned to do that. We've been in the Hadoop distribution business for years. We've obviously been in relational database, but we've done non-relational, document stores, image, video, all kinds of things in all these different domains. IBM is unique in that regard. So when we run an event like this, we're putting all the pieces together. That's what the event's all about. Show our clients all the different things that come together. Show them, by the way, how our business partners fit into those offerings, the kinds of things that they're doing, and take them through the full, you know, if you will, structure from end to end of all the possibilities, and then work with them on how do we grab at that portfolio and match it up to our business needs. I think you guys are doing a good job by having these mega events, because there's a lot of crossover between customers. Some might want a little bit of big data, some DevOps in here and there, and there's no clear general purpose product anymore. So I got to ask that question. You mentioned that they want this integrated model. You're seeing integrated suites or stacks where there's a lot of mixing and matching. So as a grandmaster on the chessboard, you've been in the industry for a while. Can you correlate this to other trends and inflection points? Because now you're seeing customers saying, okay, I'm going to build a stack for this application and workload and be good with it. And that's a great solution. It might not be up to shelf. It might be something that can cobble together some really cool technology, a little Watson here, a little data fabric there, and some computing. What does this all mean? Help us tease this out. Look, I think when you step back from all of this, you look at it both historically and in terms of what the technology permits, you begin to see an evolution that's taken place around information technology. And when it's expensive and complex, you tend to do it in smaller chunks, more in isolation. I want to solve a series of problems. But time passes. Decade after decade, you're buying more technology, adding technology, and you hit certain inflection points where the cost of computing comes down, network bandwidth goes up, the possibilities of being able to do things more flexibly now come into focus. And frankly, the client says, aha, wait a minute, I always wanted to do these things in combination. It just wasn't affordable before. I had this pattern, but nobody could respond to me. The vendor community didn't have a solution. It was always too expensive to bring it together. Well, guess what? This confluence, big data, analytics, mobile cloud, all these things allow us to begin to put things together in ways that frankly, we would have struggled 10 years ago to put together in the same fashion. So what you're saying there is is that, and then we heard some comments on theCUBE earlier where the old software would throw away skew data that was off the median because it just didn't have any way to explore that. What you're saying is now you have the ability to go in and do things that you could never do before because of the low barrier of entry for the high performance stuff. The affordability's there. The cost of computing's coming down. So from a value standpoint, what does that mean for the customer? That means a variety of new choices, right? That's where analytics is. Well, it means solutions that they could have only have dreamed of before. Now they can afford. They can find patterns and relationships that they can monetize they couldn't find before. They can hunt out the hidden jewels, the weak signals, the things that sometimes eluded them. They're now in a position to actually apply technology in an affordable way. They're able to do something that perhaps could have only been done a decade ago in a research lab, or by a heavily funded government entity that did it some special project. Now this stuff is available, literally to be deployed in businesses of all sizes. The cloud opens that aperture to almost any company of any size, and the speed and performance of the hardware makes it possible now to do things that previously you just would have passed on. But some of those things you could do in the past with a services led engagement and IBM could bring in services on top of all these technologies. How is that changing? I mean, you still have obviously big services component. Are customers getting more self-service? Are you partnering in a different way with customers? Talk about that. Well, I think that the requirements for skills remain every bit as high, if not higher than ever before. Right. You know, the technology moves, the affordability improves, and the bar gets raised on the kind of challenge that many of the customers here are very adept at standing up basic data warehouses, reporting systems. They've had those for years. It's familiar technology. You know, now they're going to move to the next level, right? Increase the size of the data, the velocity of the data, and suddenly the degree of difficulty now goes up. I mean, no one here at this conference views a terabyte as being a hard thing to do. In fact, they're all talking about hundreds of terabytes, and many are talking about pet-a-scale, pet-a-bite scale. Now it used to be, we used to run contests in the tech industry back in the 90s who had the first terabyte, full terabyte database. Yeah, how big was the room? Yeah. Who was a terabyte database, you know? And there were very few terabyte databases. That's nothing today. That's trials play. You can buy a thumb drive. It's a terabyte. You know, we can stand up a pet-a-bite of storage, pet-a-bite of solid state storage inside of a single frame today. You know, just amazing density is possible. So that puts pressure on the software. So let's get back to the software piece, because now you have unlimited storage essentially. I mean, not unlimited, but you know, it's not a lot. A lot, you know, it's not available. That means the streaming of the ingestion of data. So call it data exhaust. Call it, okay, I want everything that's connected to my network, Internet of Things, where people are things. I suck it all in. Now I got to do something with it. It creates more noise. The noise barrier is really hot. Well, you have to, you know, come up with technologies and approaches to how you filter. You know, do I want to look at all of it? Does it matter? I mean, if I'm looking at financial markets, I probably care about every piece of data. If it's an Internet of Things project with a power company and you're measuring the grid, a lot of it is just status information. You don't care about most of it. But the perturbations and the changes you care about a lot, you know? So you may ingest more data per day than almost anybody else is outputting, but yet you only care about a modest percentage of it. You know, so you get this range of scenarios. And the jury's still out on all of this. It's still early. So we were, Dave and I were talking about some of the new companies like Splunk, ServiceNow. A lot of these cloud, born in the cloud like companies, well, they weren't necessarily born in the cloud, but they're now in the cloud where you have the land and expand business model that and the customer's consumption is kind of, they, you know, buy before I buy by the drink, take a POC, grow it. Now they can, they're not as big as IBM. So they're reinvesting all their profits into growth. So I got to ask you the business model question, you know, as the general looking over the battlefield, you see the skirmishes of these young companies coming up, they're public, they're losing money, they're reinvesting it all in. That business model viable for you guys? You guys going to come in? How do you guys roll out to that landscape with a cloud like business model with all those assets? What's your, how do you view that as a business? Well, I mean, it's always been the case that the things that have been around for a while are the ones that make money and the things that are new don't make money. And that's been the case forever. It's not like this is all of a sudden a new phenomenon. Bubbles, Silicon Valley's been there before. You know, I'm involved with startups and you know, obviously in IBM we've invested in many technologies that took years, you know, five, seven years or more to finally get to the point we're now making a buck on this. You know, it's finally making money. So we're quite used to that. You know, we invest over six billion a year in R&D. We do a lot of new things that we know have long-term payback, but we have a hundred billion dollar company. And we do a lot of things for customers that they pay us for every day. And obviously there's a margin there associated with the value we deliver. And then we plow that back into R&D. So we're our own banker in that sense. And we do a lot of things that we know have a long-term proposition around making money, but we're not going to see that money in their term time. But on the business mix side though, have you seen some changes with the cloud? Kind of not forcing your hand, but just adapting to the marketplace. You guys have been very adaptive. You have great sales force. You've got a great field service force. Has the customer consumption of the technology? That's a lot of very service-oriented components. You mentioned a few of them. Has that changed anything? There's a lot of reconfiguration of the... Well, the good news in the market is that the customers are doing all the above. They're doing everything. You know, I think sometimes there's this illusion that you read about the hot new things and well, everyone's doing that. Well, it's true. A lot of people are doing it. But they're also doing a lot of things that we're no longer talking very much about that are an integral part of the way they run their business every day. This is a very, very big market and businesses are investing across a spectrum of activities. Clearly, we need to continue to invest as we have been now for many years in the new things they're doing. Cloud, analytics, mobile, social, security, hot growing areas. We need to bring those capabilities onto traditional platforms and systems because that's what keeps those platforms and systems vital and growing. So, you know, we have over 100 SaaS offerings in the IBM portfolio today. Cloud represents a big chunk of revenue for us. Certainly, nothing to sneeze at big in terms of the industry. But frankly, modest when you measure it against a $100 billion aggregate company. And that's one of the things that we grapple with is we're always going through these transitions and the idea you move the investment to where the market's going, you build mass and we've put tens of billions of dollars of revenue on the boards in these new CAMs related areas. But we're on a journey to $100 billion. And they're still not big enough to offset the decline in the other stuff, right? So, slow growing things, fast growing things and that's the world we live in. You got to make the bets. I mean, you got to make the right bets. But the rich keep getting richer in the software business. And it's true. And I had to miss your comments last September in Greenwich. I had to leave early, but I read them. You talked about the percentage of contribution in software versus hardware and how you guys are changing your portfolio. But as they say, the rich keep getting richer. There aren't a lot of new vorice. I mean, maybe Salesforce, I guess Salesforce is one. Why is that? And does that continue? Well, I mean, we all understand that in these shiny new things, the valuations relate to the way the market forward values, the stock price, right? It's not that they're getting rich off of the net profits of the company. Especially the private companies. Notable private companies. The net profits are not obviously where the money is. The money is in the perceived value associated with working on the next new thing. And so you're given that forward value benefit, right? And certainly more than a few million. And someone already, that's irrational on the private side. But more than a few millionaires that have been minted in the tech industry as a result of those kinds of things. But it's one of the things that brings energy. Obviously, the venture firms bring liquidity and that funds the innovation. And they're going to be winners. Well, you're a buyer of those guys. And you're a buyer of those guys. And we're an acquirer of these companies. You can argue it's overfunded, but isn't overfunding in a way a good thing? You get good innovation. You get to identify value. I think it's a good thing from innovation, from an economic perspective. It's that liquidity. If you can, the multiplier effect on liquidity, the creation of liquidity is incredibly important to power the economy. And if that's backed up with good successes and things that really contribute positively to employment, to the GDP, then it's a great use of cash. On the other hand, if you build a lot of companies that go into chapter 11, that's not so good. Yeah, but you're a very selective buyer of companies. I mean, every now and then you'll pop a few billion on a company, but generally speaking, you guys are very disciplined. You're a very finicky buyer. You talk about 100 SaaS acquisitions, basically, or SaaS products, many of those acquisitions. Some we've created, some we bought. And so, I'm presuming that hasn't changed, that investment discipline, right? We're talking about that a little bit. Well, we're held to a different standard, frankly. We're not in a position to not deliver on earnings and cash. We're a different kind of company to invest in. And so we think very carefully about how we use the shareholders' money. And are we buying things that we can get a good return on? We look for synergy with the rest of the portfolio. We look for a lift. We look for- It's also integration costs too. If you misfire on an acquisition, if you overpay and it's not the right call, there's also the cost internal, right? Overpay is always bad. Yeah, it always is. Overpay is always bad. But if you look back 20 years, what's it been a better use of cash? Acquisition or R&D? I know it's kind of a loaded question. Can you give us an honest opinion on that? We really can't do, no. Clearly, given the IBM franchise in the market, the R&D investments that we make are absolutely critical. We would not be where we are today had we pulled back on R&D. And we could have invested money in buying companies that would not have gotten us to where we are. And there are massive franchises out there. The 28 million mips of mainframe capacity installed in the world today. The world's businesses run on these systems. That's a level of investment that we have to make uniquely. No one else has that knowledge and expertise. So there's a lot of... R&D is critical. I think you guys really did the right thing. And also you're staying on the right vector in the market. You guys are eyeing the prize. We can see it with the big data and the systems. There's a lot of stuff that's kind of, you're kind of like getting modernized on with the customer. But I think you guys have a good track. I've been very impressed with that. But, and I got to ask you though, there's some, besides the tuck-unders and product kind of buyouts you might do that. It might be someone that pops up. Hell yeah, pluck them in, put them into the portfolio. What is the game-changing area that you're watching the most right now? In terms of turbulence, where there's opportunity that's, the need's de-risking. That you're like, okay, is it, where's the straight and narrow on the business that you see and where's the area like, okay, and what's in that sector over there? Well, I mean, in these critical areas that go under the acronym CAMHS, we're obviously looking to... CAMHS stands for? Cloud, analytics, mobile, social, and security. These are some of the hottest growing areas in the industry today. So we're invested in all of them. And obviously you want to invest wisely so you can make poor decisions. But the fact of the matter is that those are the fast growing opportunities. You want to ensure your development dollars are going in that direction. You go to market structures are moving in that direction. That is something that is, you absolutely have to keep pace with. And you can really never run fast enough. And that's a mix of organic. We report on these things every quarter. We talk about how's IBM doing in these new areas. These have been very solid significant double digit growth spaces. Cloud, 40, 50%, strong analytics growth, big mobile growth. But look, that's where the market's going. So we like to grow that fast, but we also know those are fast growing spaces. So we're certainly grabbing, if you will, our fair share. We'd like an even bigger share who wouldn't. We got to keep moving the investment, moving the skills, moving the resources. In that direction, it's what builds the IBM of the future. And software is the key component of all this now. You're seeing that clearly out there. But it's the defining technology it what enables that capability. By the way, software has to run on hardware. At least the last time I was in the lab, that was true. Gotta run on something. And so there are investments there that are critically important. And the customers often can't get at the value of the technology without assistance. So that could be IBM providing assistance or it could be the thousands of business partners that are here at this event providing assistance. But you're sharpening your hardware focus and narrowing it. There are a lot of skeptics out there. Why are they wrong? Well, because these are good hardware franchises that do make money. And we've obviously honed in on where the profits are. IBM used to be a much bigger hardware business. We shared this with the Wall Street analysts earlier this year. At one time, IBM, 1980, IBM was more than 80% hardware. In 2013, IBM was 17% hardware. Dramatic shift, services and software. So we're endlessly rebuilding this company. And that rebuild of the company is always focused on where are customers going? What does the technology allow that opens the floodgates on the next new thing that they want to do? Because their aspirations for using technology to change their business are what we have to tap into. It's interesting, on those years you mentioned one of the things that was really in place was kind of an end-to-end stack. You had SNA architecture. Remember the day you had networking protocols? So a lot of stuff was wired in with the hardware. So again, that was proprietary in some cases, IBM. But now you've got cloud, you've got a conversion infrastructure. Kind of an interesting wiring going on, but it's open now. So are you guys looking at that? Because we see Oracle doing the same thing. They have very specific solutions that really work well with workloads. So is that kind of the thing, stay standard and open? And is that the stack model? Is there a certain criteria? Certainly what our customers mean. Look, those that vote with their dollars, the customers, they want their cake and they want to eat it too. That's what I'm saying. Faster, cheaper, smaller. But by the way, they also want it integrated, but they want it open. They want it modular. They want to be able to mix and match. They want to be able to make changes. So they want all the above. But by the way, make sure it works well. So those demands are there. Do people really care if it's a proprietary piece of hardware, if it runs really well and fast and no one really interops with everything? I mean, I think openness is important. At the same time, they know that once they make an investment in any technology platform, there's an aspect of them therefore being on that platform and the cost of changing. So they want to know the vendor is investing in the platform. It's hardened. And by the way, they want to know that you're pursuing an open strategy such that more possibilities emerge for them, not just from you, the provider, but also other companies are coming in and supporting that environment. I mean, openness is about building ecosystem. So it's okay to have a hardened solution encapsulated together as long as there's choice. Kind of what you're saying, right? Open means choice, right? Not locked in. Open means choice. Well, and the proof is in the ecosystem, as you're saying. And the ecosystem tends to be the test of that, right? The truth is the larger the ecosystem, the more sense of openness that exists. So it's a subtle definitional issue. And then there's a technical view of openness which is very absolute. Keep you honest. The ecosystem keeps you honest, right? And then there's another view of it which tends to be more driven by choice, right? Which can be ecosystem related. So IBM's always had a pretty heavy planning culture. Given the push toward agility and now the contribution of the ecosystem as a sort of an adjudicator of openness, has that, and I know you've changed in terms of speeding up your planning culture, is it changing again? Are you moving faster of the things that you're doing internally to speed up decision-making? Well... Be more agile, all the buzzwords associated with that. We work with a lot of companies and we acquire a lot of companies. We don't have a problem being as fast as others are, all right? So there are always these notions, somehow or other, if you're big, you can't be fast. And that's just simply... So it's a misperception. And that's just simply not true, right? Because we're constantly moving people, we're moving technology. It's a relative term. We're making changes. We had nine months go to marketing on Watson. It's a herculean. We're making changes all the time. And we have teams that are unbelievably skilled and do, you know, all of our teams are on agile, you know, small team, very rapid development processes and techniques. So we've been very good at getting things out the door and out the door quickly. As always, our customers expect us to not just deliver quickly but also deliver reliability, scalability. You know, we're held a fairly high standard. We have to make sure those things are, you know, evident in what we do. I think that very often there's a real need for creative discovery. In other words, you don't find out what the market really wants until you go to market. So you can build in the lab, you know, forever and never get it right. Or you can go to market quickly with something that's less than perfect, you know, and then refine it, refine it, refine it. It'll run through, that's agile, right? You know, and I'm sure if you talk to many customers that are here, they can tell you stories of things that IBM delivered to them and they say, oh, I wish it ran better the first time I got it. But the good news is I got it and then they worked on it, worked on it, fixed it, upgraded it, improved it, you know, and I'm on version 10 today. And they like it and they're happy. Yeah, yeah. And IBM, you did exactly what you promised to do. And when I was an early adopter, I tended to suffer a little bit with some of the rough edges on the code, you know, but I saw you innovate rapidly and you moved to market leadership with your technology because you were dedicated to rapid innovation. And the alternative is to have a product that has way too many features that they don't want, that's sub-optimized. Or where you're trying to come up with the pluperfect integration model with all kinds of things. You missed the market. You know, nicely fitted together and you missed the market, right? You missed the market. So, you know, those are the balancing acts that you have to go through. I'd rather be fast and I'd rather have the market direct me to where I need to go long-term than attempt to figure it out in the lab. Steve, we really appreciate you coming on and we love talking with you. It's like talking to sports. It's so much fun for Dave and I. We love talking to industry trends and whatnot. But I'm going to ask you, what's next for you? What are you working on now? And what are some of the highlights of what's in the moment for your next couple of months? And also share with the folks something about IBM that they may not know and that would shed some positive insight into what you're working on. Well, I'm just relaxing here in Vegas. I'm the cube hanging out. I'm going to talk more if you want. Who's going to win the World Series? Giants or a KC? No, we don't want to get into that. Well, you know, our passion, our obsession obviously is technology and the next things that can be done technology. There's probably nothing that we're doing in IBM that excites us more than what's happening around Watson, around cognitive computing, around inferencing technologies and all the different things we bring together to make the Watson system. We're working on a lot of fascinating things in hardware design and systems design that begin to use the structures in a way that start to do a better job of mimicking some of the decision-making characteristics and profiles and patterns that we're seeing customers want to get into. There's a fascinating, it's not that computers are in any way human or becoming human, nor will they take the place of humans, but they're taking on attributes that have more and more human-like characteristics. I think one of the most incredible things we're working on is we're working on ways in which Watson, as a technology, can see, you know, of the five senses that all of us have, the one that delivers just an extraordinary amount of information at an incredible rate is site. Site is an amazing learning mechanism and a product that can see, and by seeing, I mean, it's just not a collection of pixels, but there is an interpretive engine behind it that you're now taking in sensory input and new things, new entities, new relationships, new aspects of a particular problem now, the richness that comes from being able to add another sense, if you will, into a tool like Watson, is really just extraordinary. So, I mean, we're clearly just scratching the surface, but of all the things in the technology world we're working on, that's one that obviously has this tremendous attention. It's super popular in the mainstream, but it's also got this really interesting data intelligence layer fabric that's developing out of it, and I was commenting to Ray Wang earlier, so the demo last night, I cut the line, hey, come on, I want an accident. Wait, yeah, it's just small. They didn't say that, but that's my interpretation. And I said, and I started kind of complaining to Ray Wang, he's like, dude, nine months go to market, that's Herculean. So, pretty big accomplishments with Watson. I mean, is that how you guys see it? Are you guys hurrying up, peddling too fast, or taking your time? Well, no, I mean, I'm constantly encouraging the team to go faster, right? You know, everything built. But not at the expense of quality, right? Because you have some big customers. No, I mean, they obviously have to be able to meet the customer needs and requirements, but the knowledge that you develop is something that you build upon, right? And when you talk about something, you talk about moving faster, you're moving faster off the back of the things that you've learned that work. How do you capture that? There are patterns in relationships, that becomes the base in which you then move forward, and it widens the aperture on the kinds of things you can do. So, what began as a relatively small number of customers, early adopters, getting on Watson now was quickly moving to be not just dozens, but hundreds. And we have many hundreds now that are queued up going into 2015 that want to take advantage of this technology, advisory services, and not just some of the leading edge to the limitless application space, but all kinds of things where advice is required, right? And wouldn't we all love to have an advisor, right, that had all the facts, all the data, all the information? Well, Apple series train the mainstream on this notion of, you know, what's the score of the Giants game? They could parse that pretty quickly. There are facts. There's a difference between facts and advice. Yeah, yeah. So, facts is just, you know, hand me back exactly what that is, and there's not a lot of choice. But advice means I have choice. It's ambiguous. It's unclear to me. What are the best choices? And what's the evidence that supports the choice that the system is telling me to make? That's what Watson does. That's what Watson does. It's extremely hard to do. No, it not only gives you an answer, it gives you multiple answers. It shows you why it chose the one that it chose, and then it gives you the evidence back that allows you to go, aha, now I understand why that is the right answer. I'm going to act upon that advice. Or, that's triggered a new thought. I'm going to put in some more information, see whether or not the advice changes. I'm going to explore this topic even further. You know, and so you think about it in all kinds of domains, whether your job is advisory, you can think about academia, you can think about some very tough problem solving. You know, imagine if everything that you engaged in, your auto mechanic had an advisor, you know, my bill would be a lot lower. Or you as a consumer had an advisor. Yeah, you could have an advisor, you know, all folks out there, how hard this is to do. I mean, we always say this is super hard. It's why we, all the geeks get loved, Watson, because they know how hard it is. But in your own words, how hard is it to really do the difference between, what's the score of the Giants game, how tall is the Empire State Building, versus a medical scenario with predictive, prescriptive? Well, it's profoundly different and difficult, in the sense that we've had search technologies for many decades. So long before anybody, you know, understood anything that we see on the web today, obviously there were search technologies out there. Finding facts, you know. You put in a simple request, it brings you back facts, add another one, it brings you back facts. Whether it gives you one fact or dozens of facts, you then, how to apply your brain to then understand, okay, so it's telling me this, does that dovetail with what it is I believe to be? It's a huge challenge. You know, true and correct, right? Then you take inferencing. Inferencing is a set of concepts that have been around for forever. Within computer science, these are 60s and 70s concepts that were pioneered, you know, going back literally 40 plus years ago. Early AI stuff, right? The problem was is that, you know, could I ever build anything that was flexible? I could build inferencing in a set domain, but I couldn't have a product that matched many domains. You know, so flexible ontology. That's a linguistic ontology kind of thing. Flexible ontology, right? And then if I could create a product, would it scale to a large number of users? Because it used to work well with a couple of people using it. But could it work well with thousands of people using it? You know, that flexible ontology, that ability to... And the user interface, how they're accessing it. Understand language. You know, that ability to continue to feed the system and continue to increase the amount of information it has because it's ability to get at the truth increases with the quantity of data. That way back on our earlier point here, with what's different today, you know, and big data's possible, the petabyte club now, you know, and you can build petabyte-based systems. And once you get into that domain, the number of things you're not going to know about become relatively small, you know? And that's the amazing thing the computer does. Because none of us can remember all this stuff. And even the things that remember, sometimes we question whether or not we remember it correctly or we're right. So imagine you had an advisor that never forgot, right? That always made the same associations, right? That didn't make a mistake along the way. And that gave you statistical accuracy to seven decimal places as to whether or not the conclusions that it is reaching really matches true statistical probability versus what your brain does, which is, yeah, that's pretty close. Or I got too much information in overloading. And we all know this, you know, we all make assertions and assumptions and we've all in our lives declared something to be true and we know it's true. And then what happens? We find out, well, it wasn't exactly what we thought. It wasn't because we were 100% wrong. It's because we just couldn't assemble all of the facts in our head, you know, at the moment to be able to reason through the right answer and we ended up in the wrong spot. You know, that's exactly why it's so exciting. You just basically outlined three major functional areas in the computer industry in one sentence, like all integrated in. Ingestion, computation, software, user experience. Man, it's super exciting. So once it could only improve my golf game, I'd be a happy man. It's called a virtual golf, fantasy golf. You played yesterday, but you don't know it yet. It's predicting your score. You shot. Don't show up. Yeah, sadly, my score is predictable. Dave Mills, great to have you on theCUBE. We love chatting with you. We could go another half hour, but the planes are backing up, as Dave said. Appreciate the time. Thank you. And supporting theCUBE. We love what you're doing. Dave, thanks. Watson and Analytics is a cognitive, predictive reasoning all coming to help us out. Powerful software here at IBM Insights. It's all about the insights. This is theCUBE, sharing our insights with you. We'll be right back after this short break.