(energetic music) >> Announcer: Live from San Francisco, it's the CUBE, covering DevNetCreate 2017, brought to you by Cisco. >> Okay, welcome back everyone. We're live in San Francisco for the inaugural event for Cisco's DevNetCreate, part of their DevNet classic developer community now extending out into the community of open source and cloud native and dev ops world, where applications and infrastructure coming together. It's the CUBE's exclusive two-days coverage. I'm John Furrier with my co-host, Peter Burris, head of WikiBon.com research. Our next guess is Ben Brown, CEO of Botkit out of Austin. Welcome to the CUBE. >> Thank you. >> So we were just chatting before we came on about how open source and how essentially using machines and humans workin' together, that there's a nice evolving machine learning marketplace for having new kinds of re-imagined recommendation engines. Chat bots that actually work. Integrations, again, back to software. >> Ben: Yeah. >> Tell us what you guys do, and how you guys relate to the cloud native, and what your role in open source is. >> Sure. So, it's real interesting, you know. Over the last couple of decades, an enormous amount of progress has been made on AI and machine learning, and NLP tools at these big companies like Google and Microsoft, and they are now giving that away, right? Like, it is free to use Facebook's top of the line machine learning algorithm. But, it's sort of a mystery and unfamiliar territory for developers coming from web or mobile. It's a black box that nobody's ever used before. So, what we do at Botkit is provide tools for developers, mostly developers who are coming from the web or coming from mobile development, and give them semantic, easy-to-use, and customized tools for building conversational user interfaces. And that can mean chat bots, that can mean voice skills for the Amazon Echo or Cortana or things like that, and give them these open source tools that allow them to take advantage of this exciting NLP and voice to text, and text to voice, and all that to build real software today. So what Botkit is is an open source library. It's free to use, it's MIT-licensed, so very liberally licensed, and it gives the developers tools like hearing and saying, right? So it's not about API calls and NLP classification and utterances and all that. It's about how does a robot think and act, and the metaphors around that. >> So I think of Botkit, I think of Webkit, these are languages of developers. So are you guys actually providing bot kits to create bots, or is it more of a platform? How do you guys describe what you do in open source, and how do you guys stay in business and keep the lights on? (laughter) >> Good question. Yeah, so we're a venture-backed startup. We have an open source toolkit and these kits, right? So if you want to build a Slack bot or a Facebook bot, we will give you 90% of the code that you need to bring that bot up and start talking. And that piece is all free. And we do that for Slack, for Facebook, for Twilio, for Cisco, for Alexa, and Microsoft, and a bunch of other platforms. And what we're really hoping is that we can instill in people, or sort of give to people a skill set that is akin to a web master, right? There's a bunch of skills that are interrelated that you need to actually bring this software to life. >> It saves time. It's tooling to save them time and to get acclimated and get working. >> Absolutely, absolutely. And then, on top of that, we have a set of power tools that sort of complete the process. Botkit, the open source piece is a software development library, but you also need deployment management and operational tools and content management and integrations and things like that. So that's where our business is. >> The class freemium model. The first hit's free, as they say. I'm sorry, that's a drug dealer model. (laughter) You get 'em in there but, as they scale, they're already successful, so it's not like you're gouging someone for not getting value out of it. >> Absolutely. I mean, we think about our business model in the same way a lot of other developer APIs do these days. >> Well, let's talk about some of those other developer APIs, because used to be that you used a language, then you would use a data management system, and then we start talking about web services, and that's all good. But where does this end up going, where you have a specialized toolkit for bots that you can add? You made up specialized toolkits for-- Amazon's talkin' about specialized toolkits for voice recognition that you can add. So is it just going to be in the interface? Are there going to be other classes of kits that developers are going to buy, and combine them together? Where do you see this going? >> Yeah, absolutely. I mean, it's just like, you know, all software development that came before, right? Nobody built every line of code for their mobile app. Nobody had to define what a button was for iOS. That was done at a higher level. In the same way, people who are building these conversational apps, or composing their own code with third party services, with open source software and all that combine. So there's really interesting stuff going on. Like I said, there's NLP tools coming down from all of these big players, but also from small players. There are tools like human takeover, which is like a new thing that didn't exist before. You're talking to a bot, you're starting to get angry, IBM Watson can identify your sentiment and say, "Oh, this person is frustrated. Let's bring in "a real operator." So there's third-party services to actually manage that kind of thing. >> Male: I want that job, by the way. (laughter) >> Only angry customers. >> Parachute me in just for the angry customers, yeah. >> Does not sound like a great job, yeah. And then there are almost every kind of component that you might imagine existing in the web stack is being specialized, or the mobile stack, is being specialized for conversational stuff, 'cause it's just different enough, right? So analytics and CRM and push notifications. >> I mean, you don't got to be a rocket scientist to figure out that voice is the hottest app in the market. I mean, you got Alexa, you got Siri, Google. I mean, voice interface is here. That's conversational, to your point. >> Ben: Yeah, yeah, absolutely. >> So now software will evolve. So that's kind of where you guys are betting, right? >> Yeah, absolutely. I mean-- >> John: Not just voice, but conversational software. >> Right. I mean, as I was just saying in my session here, I don't think anybody really wanted to sit down at a typewriter attached to a television. That was just a technology that we had at the time. Charles Babbage or whatever was dreaming about the thinking machine. So we're just much, much closer to that now, and we think that, over the next five or ten years, almost all software will have some sort of conversational element, whether that's in the app, does it mean you're on an Alexa skill that's embedded in the car, who knows? >> It's just never fight fashion, but this is a relevant fashion piece, where we see machine learning get rendered in AI and some of the cool applications like cars and voice and AI. So I got to ask you. You mention that all this free stuff's comin' out. It's like Christmas, it's like a kid in the candy store if you're a developer. How, in your opinion, has that shaped the developer ecosystems because, outside of the young kids who are just green and have no idea that it wasn't like this before. Back in the old days we used to actually program everything. Lot of cool stuff coming in for free from Google, from Facebook, in some cases Amazon. But I mean, what's the impact? >> I mean, people are able to take advantage of much more sophisticated technology much earlier on in the process, right? For the last 10 years, we've been talking about "Ah, machine learning, isn't it great if you're Google, "and you have ten trillion data points?" But nobody else has it, so it's not even worth talking about. But now, it's possible. You can start on day one, and start training your machine learning and models and things like that. And you don't have to actually invest in that technology. And voice to text, things like-- >> It's given them more speed to get to newer high, the higher functioning stuff. >> Yeah, absolutely. And it's bringing that kind of technology that was-- Most of AI has been in academia, right, and in research. And now, all of a sudden, it's on my kitchen counter. My kid now uses NLP technology every day, and that is a big-- Without the independent developers and smaller apps-- >> Well, the IoT's going to be in your wheelhouse, too. As more things get connected, the interfaces will be more human. >> Well, I was going to ask a question about that. Does this technology-- Today, the technology's mainly thought of part of the interface between the machine and the human being. Does this technology end up in between machines? >> Yeah, absolutely, sort of between bot. Inter-bot communication is very, very interesting. And then also-- So yes, absolutely. But also, like being on the other side of the human, or like between people, right? So customer service representatives using AI to have solutions suggested to them that they can pick from and things like that, like translating systems that suggest the response, so that you can use it if you so desire. And it makes your job easier, but it's not actually doing the transaction for you. It's really, really interesting, and that's nothing that the end user would actually experience themselves. >> Final question for you. Cisco has always been the king of networks. I mean, the internet was their wave, they rode that hard. We all know what they've done. Amazing, connecting routes together, routers, MLPS routers, PLSM routers, paths. I mean, they own that. Now they're moving up the stack, so now you're seeing this a gesture of going into the community, bringing apps and infrastructure together, to bring true dev ops. Kind of like what you're doing with your interfaces to software. What's your thoughts on this strategy. So, what's your take and reaction to what Cisco's doing? >> Clearly, the software layer is becoming more and more powerful and prevalent for people, and a bigger part of people's lives. So I think it makes tons of sense. And what Cisco's going to gain by opening these things up is the innovation of the community, like they were never going to be able to do the things that people are going to do with Spark APIs. And the way that things are connected and interwoven to each other, because I have a smart home, I have all these IoT devices. They don't talk to one another. I am left to weave them together. >> Peter: You mediate. >> I mediate, right. And I'm sophisticated enough to be able to do that. But if they're going to make it as easy as plug and play, and drag and drop, it's going to open up all sorts of exciting capabilities. >> It's the quote as saying waterfall versus agile, which one's faster? Agile. >> Well, but that's exactly why I asked the question about bots reconciling, or bots you having mediating between different devices or different machines, is that it could be a way that a human being can understand a set of instructions for how these things engage other stuff, so that it still looks like it's a set of human interfaces while, at the same time, it's operating at machine speed with machine efficiency. >> This is one of the most interesting things, particularly in the IoT space, that I've seen. There's an app called Thing-tin that is like a chat room for devices, and the way it works is like those devices emit machine messages and human readable messages, but they can talk to each other in machine language, but you can read it as a dialogue. >> That's SkyNet. That's SkyNet. I'm tellin' you, it's coming. >> Yeah, if SkyNet only turns your lights on and off. >> Machines talking to each other. "Hey, go kill that human over there." (laughter) >> Somebody's going to have to program it to kill first. >> We need algorithms to watch the algorithms. Great stuff. I think Cisco clearly, this is a move that they have to make. I've been following Cisco for many generations. Past 10 years, they were one of the first in smart homes, one of the first in smart cities, first with IoT, they called it Internet of Everything, the human network, social network. They had the pulse on all the right trends, but could not execute it, Peter. And, to your point, they'll never get there without open source, in my opinion. I think this is a signal that Cisco can do that. Now here's the key: They have the keys to the kingdom. It's called the network, and I think that making that programmable and extensible is a great strategy. >> Well, that's what they have to be able to do. They have to be able to make it, they have to make it obviously available to developers so they can create value on it. And that's something that they're still struggling to do. >> Yeah, so when he does the Botkit and does all this new creative activity going on, the network has to be adaptive and not get in the way, and not for the creativity of the developer, 'cause networking is hard. >> And that's a great point. And so much of what we do at Botkit is try to drain the complexity out of this complex stuff and make it available so that this enormous amount of power is available to the developer of today. >> Power to the developer, developers are in charge, developers are driving the network policy in a dynamic way. Congratulations on your success, great to chat with you. I'm going to check out Botkit. I already have some ways, Peter and I are already lookin' at it for the clips, and then the crowd chat virtually, great stuff, congratulations. Ben Brown, CEO of Botkit. Check it out, Botkit-dot-AI. We are soon to be replaced by bots here in the CUBE (laughter) with talking machines, but that's down the way, when SkyNet takes over. This is the CUBE here at the inaugural event for Cisco DevNetCreate. I'm John Furrier with Peter Burris. We'll be back after this short break. (electronic music) >> Hi, I'm April Mitchell, and I'm the senior--