 Hi, welcome once again to Action Item. I'm Peter Barris and this is Wikibon's weekly research meeting where we bring together some of the best minds in Silicon Valley to talk about some of the trends that are most important. We're broadcasting from here in the Cube studios in beautiful Palo Alto, California, in the studio I'm being joined by George Gilbert and David Foyer, and on the phone we have Neil Raden, Jim Cabilis, Dave Vellante. Team, thanks very much for being part of this conversation today. What we're going to do today is we're going to bring forward some of Wikibon's predictions for 2018. In a previous show we discussed what we learned in 2017. So some of the trends or some of the expectations that didn't play out as expected. This year we're going to dig a little bit deep into what we think is going to happen in 2018, and it all starts with the proposition that even as we go through significant industry change, we're not necessarily going to see the economics of the industry change as fast, which leads to prediction number one. David Foyer, what is it? So my prediction is that volume is going to take a key role in the evolution of disruptive technologies. So for example, in AI and IoT and in true private cloud, volume is going to be the key determination of when it starts to take off, when it starts to hockey stick. So this has been something that's been featured in the industry for a while, Dave, but give us an example. What's the relationship between volume and AI? So if we take the relationship between AI and volume, AI is going sideways and I would predict that it's going to go sideways in 2018 because every implementation is a snowflake. Until there are solutions out there which can be delivered in volume by vendors, then that'll be the point that things will take off. So an example, for example, automated cars, they are AI, when they start to come out in volume, there'll be volume manufacturers, volume of the sensors, volume of the processors, the ARM processors, volume of everything there will drive down cost and make those implementations. And it's still software, so we're still worried about support and service on a very, very broad scale. That leads to our second quick prediction. Dave Vellante, build on this notion of volume. What's going to be the impact on a lot of the innovative smaller companies in 2018? So, Pierre, my prediction is you've got to go scale or go home, aka go out of business. So we expect massive industry consolidation is going to take place in the next two years, certainly through 2019. The business models of VC-backed tech startups are getting smashed by cloud and to a great extent open source in a turnabout from the historical norms, innovations and cost reductions from the largest cloud players are moving at a pace that's faster than many of the most startups are able to deliver. So finding white space is much, much harder. We see private equity as playing a key role here, providing capital for M&A and doing roll-ups that are going to create scale and large portfolios that can compete. So Neil Raden, as we think about what Dave just said, one of the key things that's happening is a lot of money is being put into some of the new technologies that are intended to provide more intelligence in a lot of different places. One of the large company leaders at indicating or describing how this was going to play out was IBM of its Watson story. What's been going on with Watson? What's our prediction for how that's playing out and likely, what's a likely 2018 scenario for IBM and Watson? Not well, not the sugar coated but Watson's been a dismal failure. And I think that IBM is going to reassess their whole approach to cognitive computing in 2018. Numbers don't lie, let me give you some numbers from 2016, they obviously don't have 17 yet. But these are reliable numbers from some institutional clients of mine. Their goal for 2016 was over 8,000 clients. They achieved 500. Their goals for business partners was over 4,000 and they achieved 329. So the numbers speak for themselves. Watson hasn't caught on. It's a solution in search of a problem. It was a marketing stunt really that someone thought could be turned into a $20 billion per year business. It's not even a product really, it's dozens of subsystems that are linked with APIs. Some of them are interesting but most already are available in the open source world. Well, one of the things we talked about last week, Neil, was the idea that we're going to see more buy as opposed to build and we talked about the volume play there. And then we asked the question, is there going to be more softwares? Are there going to be more services? It sounds like IBM's play to be a dominant player in AI related services has not gone as well as expected. Is that kind of where we are right now? Well, yeah, if you look at one of the more public failures of Watson, which was MD Anderson Cancer Center, they pulled a plug on the project after $62 million. But IBM only got about $20 million at that. The rest of it went to PWC. So how they intend to split that business between global services and their partners, I really don't know. And the failure of Watson and MD Anderson wasn't entirely IBM's fault. A lot of it had to do with PWC's project management. And a lot of it had to do with the people of Anderson who basically started the project by looking at a very well understood type of leukemia that had a well understood etiology and treatment options. So when the authors looked at it, they said, we haven't learned anything for $62 million. And that's been repeated in other projects. So it sounds like this is, again, tied back to the idea of scale, volume and related issues. But it also sounds like there's a lot of question, ultimately, about what is AI? What is an AI? What role is Watson going to play? Is it going to be private data? Is it going to be public data? A lot of questions are going to emerge over the course of next year. But there are domains where AI, ML, DL are likely to have some important success. And George, we've got a prediction about where they're likely to be successful in 2018. What are we thinking, what's one domain where we think at least machine learning is going to have a significant impact in 2018? Well, keying off David's point about volume, volume economics, we think that IT operations management is going to be one of the first horizontal applications that embeds machine learning. It's not about presenting, modeling and tools to developers. It's just a heart of the application. The reason it's important, there's really two, two key reasons. We're building out shared ephemeral infrastructure, which is very different from the dedicated silos that we had for mission-critical applications. And this infrastructure and the application landscape on top of it is extremely hard to manage. And machine learning can help greatly. And I think investment in that will be driven also by a realization that this is training wheels for IoT in the sense that you're monitoring machines through data telemetry that they throw off and you're using models to figure out how they should be operating versus how they are operating. So this is significant implications across IoT, ML, and how we get to volume because it's a controlled and pretty well-defined space. By that I mean, but nonetheless, it's related to the problem space. By that I mean that bespoke applications, whether they're from AI or whatnot, are going to create new needs for new types of monitoring, but the classification of the tools and the classifications of the devices that will be monitored are pretty well understood and they're controlled by the IT industry. So they ought to have pretty good definitions. Is that what we're thinking here, George? Yes, precisely. And the bespoke pieces can be modeled because they fall within a well-known domain. But I just want to add on the go-to-market side that keys off of what Dave Valente said, which is that these IT operations management applications, they can come from cloud vendors, they can come from enterprise software vendors, but especially the ones that are going to be hybrid cloud, they're going to need enterprise sales forces to get them to market. You hear millions of, virtually millions of startups say, I'll go to market strategy is land and expand. That doesn't get you enterprise-wide and for that you need an enterprise sales force, most expensive migratory workforce in the world and startups don't have them. And that's why one of the reasons we will see roll-ups for scale. So we've talked about the need for scale, the impact on startups, the impact on big companies like IBM. One of the domains we think this is going to play out most successfully is an ITOM, IT operations management for some of these new technologies, but underneath all of this is a lot of new complexity because of distribution of function, distribution of data, distribution of application, and there needs to be a new technology concept that allows for that distribution to take place under control. And we talked about this a few weeks ago, but Jim Kabilis, what's our prediction for the role that blockchain or blockchain-like technologies are going to take in facilitating this new distribution of capability around digital business? Yeah, blockchain we're predicting it will be as fundamental to the growth of the worldwide digital infrastructure and digital markets as 30 to 40 years ago, TCPIP was to the growth of what became the web and the internet and why is that? Well, when you look at the basic principles for development of any infrastructure, when there's an innovation on the infrastructure side that is shared or standardized, robust, meaning secure and distributed, it quickly becomes a common bond enabling growth of sharing and teaming and markets and so forth. So really it's a layering process where we had TCPIP and DNS and URL providing a shared address space. Layered on top of that was public key infrastructure, which is the foundation of the security that makes blockchain so strong, you know, PKI and SSL and all that. As an enabler, that's another robust shared common infrastructure. And then on top of that, what we see now is a distributed, robust shared record of transactions. That's blockchain and really blockchain as an enabler for the new generation of digital crypto currencies, such as Bitcoin, enabling a shared robust and distributed currency or means of payment across the worldwide economy. So in many ways, blockchain as an enabler for this new generation of truly robust and shared currencies and transactions with a mutable secure to share record is just going to be a growth accelerator for the world economy in the 21st century going forward. So many respects, technology takes off when network formation occurs. TCPIP was a foundation for network formation for distributed computing. What we're basically saying is that blockchain becomes a crucial feature of how application networks get constructed over the course of the next 10 years. Have I got there right, David, before? Absolutely. That's the key is the guy who sold his first telephone was a genius. The second was easier and it gets easier and easier as that network grows and blockchain is a key contributor to the development of those networks and a one-to-one relationship, many, many one-to-one relationships that can occur from that away from centralization and to a much more distributed environment. So I think we got time for one more prediction really quickly and I'll bring it up and then I want to open it up for conversation because this is an interesting one. We come back to this notion of global network formation, blockchain being what we think or blockchain like technology is being a crucial element of that. But let's talk about how the relationship between technology, the cloud and global economies are likely to evolve. For the most part, when people think about the cloud today we think of US-based companies. Amazon, Microsoft, Google, Facebook, IBM, also in there, but there's some other companies that are going to have a say on how the cloud industry evolves over the course of the next five years. Alibaba, Tencent, Baidu. So our prediction is that in 2018 we're going to see a lot more conversation about the role that China plays in establishing some of the new rules for how cloud application networks and security plays on a global basis and that's going to facilitate the emergence of Alibaba, Tencent and Baidu also on the global stage as cloud computing companies. What are your guys' thoughts? Dave Vellante, let me start with you. Well I think we're going to see the emergence of the, we've seen the emergence of the China cloud and we're going to see that seep through other parts of Asia Pacific. As we discussed earlier as a team in our private meeting Europe is going to be a very interesting pivot point because if China can control at least portions of Europe and use that as a lure for China that's going to give them a leg up on global cloud. So that leads ultimately to a series of questions about what will be the relationship between formation of cloud industries, the evolution of the cloud industries and geopolitical concerns and I think what we need to do guys is dedicate an entire research meeting to that question because it's going to be one of the most important dictators of how the industry evolves over the next few years and ultimately how businesses and enterprises need to start establishing crucial partnerships with their key and strategic suppliers. So look in the last couple of minutes we want to do our action item round. Now what we do here at the action item show is we start off having a conversation and then we go into the action item. What are you going to do differently Monday as a consequence of the information we're talking about? So let's do that now. Hit some action items. What you heard from the five, six predictions that we talked about. David Floyer, what's your action item? So my action item is for CIOs and CTOs is to take a pause on IoT and look for vendors that have solutions which can be put in easily and quickly and span OT and IT in the IoT space. Neal Raiden, what's your action item? Well I think there's a lot of activity around AI and there's going to be an explosion of it in 2018 but most of it's not really going to be AI it's going to be machine learning and machine learning is really just math and floating points. AI is different, right? AI is neuroscience, it's neurology, it's biology and physics and sociology. It's more science. I think that some machine learning is there on that horizon that AI, but it's not. So we need to make sure we're clear about what announcements and what technology is machine learning versus artificial intelligence. Jim Cabela, what's your action item? My action item is to revisit IBM's prospects in the AI market and deep learning going forward and revisit our positive note actually because IBM was actually turned around their cognitive strategy in the last year where they focus on the power AI platform which is really framework agnostic and so forth. And really the AI space that's actually shaping up is different from one that IBM had always envisioned at the start of this decade. And so really in 2018 we're going to see IBM come on strong I believe as a provider of the core, one of the providers of the core, get framework agnostic, deep learning, a development of platforms in the industry. That's my prediction. David Vellante, what's your action item? I think if you're a startup, you really have to take a hard look at your business and the value that you're bringing to market and be honest, if you're not delivering something that the cloud guys can't deliver or don't want to deliver, then I think you really got to think about either pivoting or exiting the business that you're in. And as part of that, I think you've got to find to George's point, distribution channels and distribution partners that can help you with go-to-market at scale or you're in big trouble. George Gilbert, action item. We've been talking about sort of the cloud wars and my recommendation to CIOs and senior IT leaders would be that if you want to hedge your bets, you don't want to be all in on one cloud. It's not dividing a workload across different clouds. Pick a cloud for a workload or for an application because portability is sort of more of a dream than a reality. It's not about moving containers around. You're in an API ecosystem, you're subject to data gravity. So it's almost like if you're going to do the equivalent of distributed computing, you're going to put some part of the application on one cloud and some part in another cloud. So the action item is be smart about the relationship between new style of applications and architecture and cloud choices. Okay, let me summarize the meeting very quickly. This has been a great conversation about predictions in 2018. You expect to see more from us over the course of the next month. It's going to be a major theme of ours in November and into December. So quickly the findings are these. The technology industry made a major mistake with the dot com boom. And the mistake was a presumption that technology change necessarily meant economic change. That is a false assumption. The economists of technology have been pretty well understood for quite some time and they're going to assert themselves even as we go through the significant transformative period in the technology industry. And the economics of volume are going to continue to be important. And we expect that those economics coupled with the three factors of what's driving cloud architecture decisions, the realities of physics, geopolitical concerns and intellectual property concerns are going to lead to some significant changes in 2018 that we've only just conceived of. One, we expect that we're going to see an emergence of true private cloud that will continue to be crucial to how businesses think about their information, technology, overall infrastructure and plant. And that's going to have an impact ultimately on where AI gets developed. More from software vendors based on volume. Two, we expect to see a significant impact on ultimately what happens in the VC funded world as startups which have historically just presumed that there was no need for go to market, that everything was going to be a try and buy and then we'd scale from there, start to hit the business realities of the consistency of the economics of volume. Three, IBM we think is repositioning and some are paradoxically is likely to become more successful as a consequence as a provider of the technologies that make possible some of these new comprehensive, complex AI and related oriented technologies and not just as a service provider. Very importantly, ITOM is going to become increasingly important and we'll see AI, machine learning be an essential feature of that. In fact, one of the places where we learn how to do it right. And the final one is a lot's going on with blockchain but we expect greater distribution of applications, greater distribution of data and the security technologies and the technologies for bringing that together and supporting the network formation of data and applications must be in place and that's going to be a major area of technology innovation in 2018. All right, so this closes out our action item for this week. Once again, I'm Peter Burris. I'd like to as always thank the Wikibon team for participating with me today and we look forward to once again visiting with you from the CUBE studios here in Palo Alto, California on the next action item. Thank you very much.