>>Live from Orlando, Florida. It's the cube covering Microsoft ignite brought to you by Cohesity. >>Welcome back everyone. You are watching the cube. We are the cube, the ESPN of tech, and we are here at the orange County convention center for Microsoft ignite. I'm your host, Rebecca Knight, sitting alongside of my co host Stu Miniman. We are joined by Dave Cahill. He is the principal PM Bonzai at Microsoft. Thank you so much for coming on the cube. Thanks for having me. It's been a while. Has been by your back. That's right. So you are now, you were the COO of Bonzai. You are now part of Microsoft. There was an acquisition about a year ago. Tell us a little bit about bonsai. It's the AI business system. Got a shout out from Satya on the main stage yesterday. Tell us a little bit about bonsai and then about the transition about now being part of Microsoft. >>Yeah, sure. So the, the big vision for Bonzai from the founders, Mark and Keene was how do you build a set of tools? This makes AI more accessible than to just data scientists. How do you open up, ended up to developers and subject matter experts. And so from day one they've been focusing on building this abstraction layer of platform set of tools. They really enables more than just data scientists access to the low-level mechanics of machine learning, of deeper enforcement learning. Um, everything we've been working on really they've been working on for four years prior to the acquisition was, uh, building out that tool chain. And from my side of the world it was where did we figure out where to point that? Where do we, where are we seeing the strongest traction and adoption for the tools? Early days, uh, from a go to market perspective. And so while they worked on the technology, we really found a pocket of interest, uh, in these real world, often industrial systems. Uh, and so inside Bonzai that's a lot of the work we were doing was taking that platform to market. Um, as part of Bonzai. And then, you know, of course post acquisition, we're doing a lot of the same >>thanks so accessible AI. I love the concept, but what does it really mean? So this is so that someone could be a subject matter expert in an industrial company and be able to still program. Can you explain a little bit, give us an example of, of what bonsai was? >>Yeah, sure. And I mean there's a lot of low level mechanics and machine learning, the algorithms, the toolkits, et cetera that are, that are difficult for just anyone to pick up and start programming. And so the idea here is how can you write an obstruction layer above that? And in this case, it takes a foamer for programming language that allows a developer or subject matter expert to break down the concepts of the problem they're trying to solve in, in, in business terms, right? And so if you think about a wind turbine or a drill or um, a baggage optimization system, it's not the data scientists that intimately understands the behaviors of that system and how it works. It's the subject matter expert that can practically stand next to it and understand or hear that it's starting to fail. Or they know the, the way to turn the knobs most optimally to figure out how to program that system. Now if you just took a of data and threw it at infrastructure, eventually it would figure it out the patterns and how to optimize that thing. But you have a subject matter expert inside the four walls of your organization that readily knows how to solve it like that. And so why not empower them with a, a programming language, really a mechanism to outline the core concepts that you want the AI to learn because they've spent their entire career, uh, trying to figure them out. All right, >>so yeah, Dave, yesterday, Satya Nadella talked a bit about the autonomous systems and if I got it right, he said, we're allowing those engineers to really build systems, become the teachers for what's going on there. So help help frame this a little bit as to where this fits into kind of the broader AI discussion that Microsoft's having with companies today. >>Yeah, I think there's, there's a obviously a massive AI portfolio at, at Microsoft and there's lots of different applications and systems and use cases that are fit for more and more intelligence in the form of AI and machine learning. What we've seen is that an opportunity in the real world and the physical domain that requires a different set of tools and techniques than maybe in the logical, you know, our data centric domains. And oftentimes in the press you see a lot of emphasis on supervised and unsupervised learning and very data centric use cases for the logical world, right? For, for databases or CRM systems or things like that. We believe there's this massive opportunity in the physical world. And when you get into the physical world and these vast practically infinite state spaces, you need different sets of tools and from a machine learning perspective, different sets of techniques. And so I think Microsoft looks at the entire portfolio and says, you need the right tool for the job. Um, as opposed to hammer nailing everything. And that's really the autonomous systems piece is really our effort in real world systems. So >>David, you know, when I'm listening to what you're saying there reminds me of some of the discussions we've been having the last five years or so about the industrial internet. A lot of the OT systems here, which really outside the domain of traditional it or though some of the same challenges that your your team's facing. >>Absolutely. So OT, it's interesting you bring that up. Um, oftentimes the teams that have time inside an organization to pick their head up from their day job to look at new emerging technologies aren't in operations. They're not in the business because they're running the business. And so you have to be able to bridge the gap between central technology, central and innovation teams and those that are actually running the business. And I view OT as kind of the, the kind of mortar between those two bricks oftentimes as the one that has to accept this technology and figure out how to deploy it. And that's just not technically that it works, but also kind of commercially and from a safety risk, trust perspective. So OT really has a, a big role in this. And understanding, not that it just solves the problem technically, but it actually can be deployed, um, in ways that fit within corporate security requirements, data privacy requirements, trust, et cetera. Um, it's not, you know, there's a, there's a, there's a lot of gaps to be bridged there. So I saw this, this, this, like autonomous systems have been projected to grow to more than 800 million in operation by 25. Right? >>That's a big number. So what are you doing within Microsoft to do prepare for that? >>Yeah, so I think I view autonomous systems. It's not a product, it's an endpoint, right? This is like 2000 when VMware came out and said, listen, you're on the journey to the virtual data center. Right? And their customers were in physical data centers trying to go virtual. The journey towards autonomous systems is kind of that we're on that same path. And really it's about providing customers the tools to, but I them along that journey from where they are today to kind of full autonomy, full autonomous systems. And it's a, it's a, it's a maturity, right? You start out, you know, just managing that system, you're maintaining it, then you're, maybe you're, you're optimizing it and you're, then you're controlling it a little bit better, but there's always a human in the loop and then you're at full autonomy. And I think along that path there's lots of different pieces or tools and technologies that we can bring to bear to help them on that journey. Um, technically, commercially. And then also from a safety and trust perspective. And so a lot of the work we're trying to do is build out that tool chain and, and we think Bonzai is a core piece of that actually at the, at the center of what we're trying to do. >>So how, how when you're talking about the human and the aluminum, I'm, I'm imagining a subject matter expert who is working in concert with you developing whatever, whatever tool it is that is going to automate something that they are the subject matter expert, as you said, can fix it like this. Calibrate the buttons and know when a system's about to fail. So how, how trusting are they in terms of, Oh, so this is no longer something I'm going to be doing here. How, how, how do you work with them and, and helping them understand? No, really you can trust this. >>I think it's really about, um, augmenting and scaling the work of the, of the experts and, and oftentimes in every customer engagement we have the subject matter experts are excited because they're literally caudifying their expertise and then figuring out how to scale it. Right? Those experts are frustrated because they are the subject matter expert by definition. They're the problem solver for that problem for everybody in the organization. And so the ability for them to take that expertise in scale, it means more time for them to do what they really want to do, which probably isn't solving problems tactically for everyone. That's not at the expertise level. They are at the executive level. It's about scaling that quality of work so that your expert, you know, your best expert for tuning this turbine can then be scaled across the organization and you're reducing, you know, training costs and other things because you can scale that expertise more effectively. >>Yeah. So Dave, what are some of the big challenges that customers are having? Is it the availability of the expertise and hiring the right people? Uh, you know, we, we've looked at, uh, you know, the, the big data wave, uh, you know, half of those deployments failed for, you know, so many different reasons there. You know, why, why, why, why will this be different? >>Yeah, I mean it's certainly not without challenges. I mean, I think the, one of the things where we run into, you know, data readiness, like I naively thought because we use simulations, we got, we got over the cold start problem that, you know, we don't have data, we'll just use a simulation instead, I think to get around the idea that simulations, there's this idea of a simulation, which is where we train our environment in. And I can kind of go into that in detail, but that's very different than a machine learning ready simulation and having a simulation that runs. It can be parallelized, it can run on Azure that works fast enough to train. These are all impediments to just getting to train these models before you even get to the actual model working in the real world. And so I think the pipeline for training these models is as intense in some cases as you know, data centric training environments. >>Once you get that model trained, it's been about deployment and you have a whole different set of challenges and that's where OT comes into play is starting to figure out, okay, how do we operationalize this model? Is a human in the loop? Is there a a mechanism to to stop the AI and defer to the human right. And we see a maturity model there as well where customers are starting with decision support, which means you know, the AI is not controlling the end system. It is making a recommendation and then a business analyst would then implement that in real time. But walking through what those procedures look like is something that most customers haven't done yet until they're like right at that last step ready to deploy to saying, wait, who's going to watch this? What, what is our safety procedure for deploying a drill, an autonomous drill? It usually doesn't exist in an organization today. >>Yeah, it sounds, it's a little bit different as to, as opposed to, you know, just your regular it operations and you kind of say, here's the five step model. Oh wait, I've always done this. You're, you're attacking some new challenges here. So are they a little bit more likely to move a little bit further and let the autonomy take over? Is that the case? >>Um, I mean, I think so, and it's, it's certainly lines of business, right? This is not, it is there to kind of manage the transition as needed and kind of watch over for security and privacy concerns. Um, I don't, I don't see the hesitation around the autonomous nature of it from the business users. It's, it's people around the periphery, whether that's security or compliance or safety that is most concerned about that. And organizations I think are still trying to get all of those people in the same room and develop policies around that. And oftentimes for better or worse, we're the, we're the forcing function to get them all in the same room and say, okay, what is this going to look like? But, but I, I see the businesses as really driving for the smarter and smarter and increasingly autonomous systems and excited about those pieces because the, the efficiencies to be gained from, from that are so significant. >>And a lot of these use cases I want to ask you about innovation. So this is, you are part of Bonzai and now you are part of Microsoft, which as big tech companies go is, is a rather mature company. We've had some guests on this week who've said that Microsoft actually feels like a lot like a startup. Yeah. I'm interested to hear the, the approach to innovation, the mindset that your new colleagues have and how you are keeping that, that more startup agile approach and that inclination in this big company. Yeah. So I can certainly speak to our experience with Bonzai. It's been pretty neat. I think as having been acquired a few different times by different companies, the way that Microsoft has landed this technology has actually been quite interesting. And we sit within a team within Microsoft research called business AI and business AI's entire charter is to incubate either required or organically developed technologies to the point that they're ready to graduate and scale across the organization. >>Up until that point in time, they're trying to figure out, you know, almost product market fit, but inside a larger organization, leveraging the tools that you know at their disposal that is the broader Microsoft, whether that's the field of the marketing engine or things like that. And then you seeing bonsai be able to take advantage of things like that. The keynote was Satya and, uh, you know, our access and collaboration with the Microsoft field, but we're still in that incubation mode trying to figure out exactly how the technology goes to market. Um, let be continuing to build out and mature the technology and figure out the right home for it. Um, the right partner for it. If it's a business unit or you know, whatever that may be. Um, and I think in that scenario, we're, we're a bit standalone in that regard while we figure this process out. >>So it's, it's, I think oftentimes you see innovation gets stymied when you, you, you force a premature integration of technologies like this and you almost kind of determine their destiny before even knowing really where they're trying to go. And just letting us breathe a little bit for a pointing for, for a period of time, I think allows a better outcome than if you tried to guess ahead of time. Cause at this early stage, you don't know the answer, right? You're still trying to figure out what is the ideal application, what is the ideal target audience? What is the ideal, um, port part of the portfolio where they should sit? Right? Those, those aren't, I think, guessing those up front, even a year ago when the acquisition closed would have been impossible. So that kind of, I don't know that gestation period is, is I think a key, uh, Dave, take us inside some of the conversations you're having at the show. >>Uh, key takeaways you want people to have of, of your group. Uh, out of Microsoft ignite. >> Yes. Right. I think a lot of the conversations are, you know, this, this big vision that is autonomous systems and that really is an end point. And what you really have to do is distill down, you know, where to get started. And that's not the glamorous kind of use cases are the ones that you see in the press or drones. Um, there are autonomous vehicles, right? It's, you know, things that likely fly or we saw on the Jetsons. But the reality is that like where customers are seeing the strongest business opportunity is, is drills, it's turbines, it's air conditioners, it's a extrusion process for some food that you've probably consumed right while you've been here at the conference. Um, that's, and so really kind of, I think dialing customers into surface level use cases that are a fit for deep reinforcement learning is refreshing because a lot of people come at it saying, well, I don't have an autonomous vehicle and I don't have a drone, so I must not be for you. >>And that couldn't be further from the truth. All you need is a control system. Right? If you have any sort of system run by a PID controller or model predictive control, you can optimize that system further with deeper enforcement learning and bonds as a mechanism for making that significant more accessible to your teams. So I think bringing it way back to like, Hey, I saw this big vision on stage, where do I start? It's just really been a bit of a, you know, a search inside their organization for the types of applications that are good fits >> AI. It's not just for the Jetsons anymore. That's right. Great. I'll take it. Dave Cahill. A pleasure having you on. Thank you so much. Yeah, thank you both. It's good to be back. I'm Rebecca Knight for Stu Miniman. Stay tuned for more of the cubes live coverage.

>> Narrator: Live from Boston, Massachusetts, this is the Cube, covering Spark Summit East 2017, brought to you by Databricks. Now, here are your hosts, Dave Vellante and George Gilbert. >> Welcome back to Boston, everybody, where the town is still euphoric. Mike Gualtieri is here, he's the principal analyst at Forrester Research, attended the parade yesterday. How great was that, Mike? >> Yes. Yes. It was awesome. >> Nothing like we've ever seen before. All right, the first question is what was the bigger shocking surprise, upset, greatest win, was it the Red Sox over the Yankees or was it the Superbowl this weekend? >> That's the question, I think it's the Superbowl. >> Yeah, who knows, right? Who knows. It was a lot of fun. So how was the parade yesterday? >> It was magnificent. I mean, it was freezing. No one cared. I mean--but it was, yeah, it was great. Great to see that team in person. >> That's good, wish we could talk, We can, but we'll get into it. So, we're here at Spark Summit, and, you know, the show's getting bigger, you're seeing more sponsors, still heavily a technical audience, but what's your take these days? We were talking off-camera about the whole big data thing. It used to be the hottest thing in the world, and now nobody wants to have big data in their title. What's Forrester's take on that? >> I mean, I think big data-- I think it's just become mainstream, so we're just back to data. You know, because all data is potentially big. So, I don't think it's-- it's not the thing anymore. I mean, what do you do with big data? You analyze it, right? And part of what this whole Spark Summit is about-- look at all the sessions. Data science, machine learning, streaming analytics, so it's all about sort of using that data now, so big data is still important, but the value of big data comes from all this advanced analytics. >> Yeah, and we talked earlier, I mean, a lot of the value of, you know, Hadoop was cutting costs. You know, you've mentioned commodity components and reduction in denominator, and breaking the need for some kind of big storage container. OK, so that-- we got there. Now, shifting to new sources of value, what are you spending your time on these days in terms of research? >> Artificial intelligence, machine learning, so those are really forms of advanced analytics, so that's been-- that's been very hot. We did a survey last year, an AI survey, and we asked a large group of people, we said, oh, you know, what are you doing with AI? 58% said they're researching it. 19% said they're training a model. Right, so that's interesting. 58% are researching it, and far fewer are actually, you know, actually doing something with it. Now, the reality is, if you phrase that a little bit differently, and you said, oh, what are you doing with machine learning? Many more would say yes, we're doing machine learning. So it begs the question, what do enterprises think of AI? And what do they think it is? So, a lot of my inquiries are spent helping enterprises understand what AI is, what they should focus on, and the other part of it is what are the technologies used for AI, and deep learning is the hottest. >> So, you wrote a piece late last year, what's possible today in AI. What's possible today in AI? >> Well, you know, before understanding was possible, it's important to understand what's not possible, right? And so we sort of characterize it as there's pure AI, and there's pragmatic AI. So it's real simple. Pure AI is the sci-fi stuff, we've all seen it, Ex Machina, Star Wars, whatever, right? That's not what we're talking about. That's not what enterprises can do today. We're talking about pragmatic AI, and pragmatic AI is about building predictive models. It's about conversational APIs, to interact in a natural way with humans, it's about image analysis, which is something very hot because of deep learning. So, AI is really about the building blocks that companies have been using, but then using them in combination to create even more intelligent solutions. And they have more options on the market, both from open source, both from cloud services that-- from Google, Microsoft, IBM, and now Amazon, at their re-- Were you guys at their reinvent conference? >> I wasn't, personally, but we were certainly there. >> Yeah, they announced the Amazon AI, which is a set of three services that developers can use without knowing anything about AI or being a data scientist. But, I mean, I think the way to think about AI is that it is data science. It requires the expertise of a data scientist to do AI. >> Following up on that comment, which was really interesting, is we try and-- whereas vendors try and democratize access to machine learning and AI, and I say that with two terms because usually the machine learning is the stuff that's sort of widely accessible and AI is a little further out, but there's a spectrum when you can just access an API, which is like a pre-trained model-- >> Pre-trained model, yep. >> It's developer-accessible, you don't need to be a data scientist, and then at the other end, you know, you need to pick your algorithms, you need to pick your features, you need to find the right data, so how do you see that horizon moving over time? >> Yeah, no, I-- So, these machine learning services, as you say, they're pre-trained models, totally accessible by anyone, anyone who can call an API or a restful service can access these. But their scope is limited, right? So, if, for example, you take the image API, you know, the imaging API that you can get from Google or now Amazon, you can drop an image in there and it will say, oh, there's a wine bottle on a picnic table on the beach. Right? It can identify that. So that's pretty cool, there might be a lot of use cases for that, but think of an enterprise use case. No. You can't do it, and let me give you this example. Say you're an insurance company, and you have a picture of a steel roof that's caved in. If you give that to one of these APIs, it might say steel roof, it may say damage, but what it's not going to do is it's not going to be able to estimate the damage, it's not going to be able to create a bill of materials on how to repair it, because Google hasn't trained it at that level. OK, so, enterprises are going to have to do this themselves, or an ISV is going to have to do it, because think about it, you've got 10 years worth of all these pictures taken of damage. And with all of those pictures, you've got tons of write-ups from an adjuster. Whoa, if you could shove that into a deep learning algorithm, you could potentially have consumers take pictures, or someone untrained, and have this thing say here's what the estimate damage is, this is the situation. >> And I've read about like insurance use cases like that, where the customer could, after they sort of have a crack up, take pictures all around the car, and then the insurance company could provide an estimate, tell them where the nearest repair shops are-- >> Yeah, but right now it's like the early days of e-commerce, where you could send an order in and then it would fax it and they'd type it in. So, I think, yes, insurance coverage is taking those pictures, and the question is can we automate it, and-- >> Well, let me actually iterate on that question, which is so who can build a more end-to-end solution, assuming, you know, there's a lot of heavy lifting that's got to go on for each enterprise trying to build a use case like that. Is it internal development and only at big companies that have a few of these data science gurus? Would it be like an IBM Global Services or an EXIN SURE, or would it be like a vertical ISV where it's semi-custom, semi-patent? >> I think it's both, but I also think it's two or three people walking around this conference, right, understanding Spark, maybe understanding how to use TensorFlow in conjunction with Spark that will start to come up with these ideas as well. So I think-- I think we'll see all of those solutions. Certainly, like IBM with their cognitive computing-- oh, and by the way, so we think that cognitive computing equals pragmatic AI, right, because it has similar characteristics. So, we're already seeing the big ISVs and the big application developers, SAP, Oracle, creating AI-infused applications or modules, but yeah, we're going to see small ISVs do it. There's one in Austin, Texas, called InteractiveTel. It's like 10 people. What they do is they use the Google-- so they sell to large car dealerships, like Ernie Boch. And they record every conversation, phone conversation with customers. They use the Google pre-trained model to convert the speech to text, and then they use their own machine learning to analyze that text to find out if there's a customer service problem or if there's a selling opportunity, and then they alert managers or other people in the organization. So, small company, very narrowly focused on something like car buying. >> So, I wonder if we could come back to something you said about pragmatic AI. We love to have someone like you on the Cube, because we like to talk about the horses on the track. So, if Watson is pragmatic AI, and we all-- well, I think you saw the 60 Minutes show, I don't know, whenever it was, three or four months ago, and IBM Watson got all the love. They barely mentioned Amazon and Google and Facebook, and Microsoft didn't get any mention. So, and there seems to be sentiment that, OK, all the real action is in Silicon Valley. But you've got IBM doing pragmatic AI. Do those two worlds come together in your view? How does that whole market shake up? >> I don't think they come together in the way I think you're suggesting. I think what Google, Microsoft, Facebook, what they're doing is they're churning out fundamental technology, like one of the most popular deep learning frameworks, TensorFlow, is a Google thing that they open sourced. And as I pointed out, those image APIs, that Amazon has, that's not going to work for insurance, that's not going to work for radiology. So, I don't think they're in-- >> George Gilbert: Facebook's going to apply it differently-- >> Yeah, I think what they're trying to do is they're trying to apply it to the millions of consumers that use their platforms, and then I think they throw off some of the technology for the rest of the world to use, fundamentally. >> And then the rest of the world has to apply those. >> Yeah, but I don't think they're in the business of building insurance solutions or building logistical solutions. >> Right. >> But you said something that was really, really potentially intriguing, which was you could take the horizontal Google speech to text API, and then-- >> Mike Gualtieri: And recombine it. >> --put your own model on top of that. And that's, techies call that like ensemble modeling, but essentially you're taking, almost like an OS level service, and you're putting in a more vertical application on top of it, to relate it to our old ways of looking at software, and that's interesting. >> Yeah, because what we're talking about right now, but this conversation is now about applications. Right, we're talking about applications, which need lots of different services recombined, whereas mostly the data science conversation has been narrowly about building one customer lifetime value model or one churn model. Now the conversation, when we talk about AI, is becoming about combining many different services and many different models. >> Dave Vellante: And the platform for building applications is really-- >> Yeah, yeah. >> And that platform, the richest platform, or the platform that is, that is most attractive has the most building blocks to work with, or the broadest ones? >> The best ones, I would say, right now. The reason why I say it that way is because this technology is still moving very rapidly. So for an image analysis, deep learning, very good for image, nothing's better than deep learning for image analysis. But if you're doing business process models or like churn models, well, deep learning hasn't played out there yet. So, right now I think there's some fragmentation. There's so much innovation. Ultimately it may come together. What we're seeing is, many of these companies are saying, OK, look, we're going to bring in the open source. It's pretty difficult to create a deep learning library. And so, you know, a lot of the vendors in the machine learning space, instead of creating their own, they're just bringing in MXNet or TensorFlow. >> I might be thinking of something from a different angle, which is not what underlying implementation they're using, whether it's deep learning or whether it's just random forest, or whatever the terminology is, you know, the traditional statistical stuff. The idea, though, is you want a platform-- like way, way back, Windows, with the Win32 API had essentially more widgets for helping you build graphical applications than any other platform >> Mike Gualtieri: Yeah, I see where you're going. >> And I guess I'm thinking it doesn't matter what the underlying implementation is, but how many widgets can you string together? >> I'm totally with you there, yeah. And so I think what you're saying is look, a platform that has the most capabilities, but abstracts, the implementations, and can, you know, can be somewhat pluggable-- right, good, to keep up with the innovation, yeah. And there's a lot of new companies out there, too, that are tackling this. One of them's called Bonsai AI, you know, small startup, they're trying to abstract deep learning, because deep learning right now, like TensorFlow and MXNet, that's a little bit of a challenge to learn, so they're abstracting it. But so are a lot of the-- so is SAS, IBM, et cetera. >> So, Mike, we're out of time, but I want to talk about your talk tomorrow. So, AI meets Spark, give us a little preview. >> AI meets Spark. Basically, the prerequisite to AI is a very sophisticated and fast data pipeline, because just because we're talking about AI doesn't mean we don't need data to build these models. So, I think Spark gives you the best of both worlds, right? It's designed for these sort of complex data pipelines that you need to prep data, but now, with MLlib for more traditional machine learning, and now with their announcement of TensorFrames, which is going to be an interface for TensorFlow, now you've got deep learning, too. And you've got it in a cluster architecture, so it can scale. So, pretty cool. >> All right, Mike, thanks very much for coming on the Cube. You know, way to go Pats, awesome. Really a pleasure having you back. >> Thanks. >> All right, keep right there, buddy. We'll be back with our next guest right after this short break. This is the Cube. (peppy music)