 Live from Orlando, Florida, it's theCUBE. Covering Microsoft Ignite, brought to you by Cohesity. Welcome back everyone to theCUBE's live coverage of Microsoft Ignite. I'm your host, Rebecca Knight, along with my co-host, Stu Miniman. We are joined by Jeffrey Snover. He is a technical fellow, O365 Intelligent Substrate at Microsoft, most famous for being the father of PowerShell and one of the key architects of the Windows server. Thank you so much for coming on, for returning to the show. Yeah, thanks, it's great to be back. So first of all, Defant, you're relatively new to this role. Yeah. So tell us a little bit about what you're doing and what is the Intelligent Substrate? Yeah, so you know, a lot of people get this confused. There's the Intelligent Substrate, there's O365, the Microsoft Graph, and what I do is I say, hey, the best way to think about this is an analogy to an operating system. Operating systems are complex, but at the end of the day, they're really, really simple. They only do three things. They manage and protect resources. They provide services for developers, services, APIs and common controls, and then they provide a base set of applications and a way to get additional applications. So in Windows, manage CPUs, memory, the services, Win32 APIs, and then the applications like the browser, et cetera. So O365 can really be viewed as an operating system. This sounds strange, why? Because most operating systems have been operating systems for devices, an operating system for a phone, an operating system for a PC, an operating system for a server. This is an operating system for people and organizations. So when we think about those three responsibilities, resources and protecting and managing resources, these are the resources for people and organizations. So it's their identity, their emails, their chats, their documents. Services for developers. These, where there's Win32 for Windows, we have MS Graph. That's our public API. But then we have services to be able to create, collaborate and communicate documents and interactions. And then the applications are things like teams and Outlook, et cetera. And so then, oh sorry, then the substrate. The substrate's sort of at the core of it. That's one of our core services. It is storage and then a set of services to manage that and set of services. So the storage is basically a planetary scale, no SQL data store. So every time you create a chat, an email, document, whatever, it gets stored in the substrate and then three additional copies are created, one of them at least 250 miles away. That's why our data availability and the high availability are one thing. So everything gets stored there and then that allows us to do common services like search against it. Does that make sense? So, Jeffrey, one of the biggest challenge people have is when you learn about something and then it has changed an awful lot. I think back to the first time I used Microsoft Word, Microsoft Excel, it wasn't connected to the internet. Exactly. Let alone talking about the era of global scale and AI and all of these things that you can do in. So maybe give us a fresh as if I'm a brand new person and I don't have all of the legacy history with the Microsoft Office family. What is the new people OS that you're talking about? Yeah, so I like to think of it as back to the original Office 1.0. If you remember the original Office 1.0, it had Word, Excel and PowerPoint. And I like to joke, I say it was integrated with the advanced technology at that day called Cardboard, right? We just took the floppy disk from each one of those products, put it in a cardboard box and said it's a suite. But then it was a vision too, a vision of how things should work together to help the individual. And then after that version one, then we reorganized the organization to have common technology teams. And that's when we started to get common controls, common user experience, et cetera, common file formats. And then it became a true integrated suite. Same thing happened when we went to the cloud. We had all these products that would have a front end coupled to a back end, another front end coupled to a back end, another front end coupled to a back end. Each one would have one or more SDKs, et cetera. And when we first brought them to the cloud, it was the same sort of thing, integrated with an offering and a name. But there was a vision there and then that vision drove the reality. And what we did was we said, hey, let's figure out how to have a common storage for these things, common back end, a common way to communicate, a common way to do messaging. And then that took a number of years, but that's what drives this consistency. And so that's why when you go and you say, I don't want to like to search for something, you'll find that term, whether it's in your word documents or it's in your emails or your team chats or anything. It's that commonality. Did that make sense? Did that answer your question? It does. So, right, it's, I think about the era of collaboration and there were competitors to Microsoft that came out that were built on the internet and delivered those solutions. So this week we've talked to, we haven't dug deep into Teams, but everyone we've talked to that's using it is like, no, no, really, this is a really great product. And almost like, forget about some of the things you might have remembered through some of those iterations and changes and things not working together. Teams has been built and is allowing some great collaboration communication with remote workers, smaller businesses and the like. So it's tough because especially if you were using one tool and you've gone over to some other tool set, it's like, why would I go back to that? But it's a very different Microsoft productivity suite today than we might have used in the past. That's exactly correct. And then the intelligent substrate is this layer of AI on top of the substrate, right? So part of that is search, but then we're also doing natural language processing. So basically imagine you store a file in OneDrive. That gets stored in OneDrive and a workflow gets kicked off. And that workflow then goes and analyzes the contents of that file and creates search terms, et cetera, so that we then have common search. And then we've got natural language processing that'll go and find, hey, what are the key points for that document? How do I summarize that document? So then if you see it somewhere, you can say, oh, show me the file card. And I'll say, here's this document. You don't have to read the whole thing. Here are the three key points about it. And so to answer the question, why would a platform guy be working in office? It turns out that to build this AI infrastructure, it's really sort of a platform play. There's key advances that need to be made in AI, but actually when you get involved in AI, what you realize is what we really need is more engineering than more science. We need more science, no doubt about it. But boy is there a need for engineering. I need to figure out how to get three to five to seven orders of magnitude, more volume of AI going through the system. So when you talk about these key advances in AI that need to be made in terms of applying them to O365, describe them for us and talk about how they will change the future of work and the way we collaborate with our team members and the way we communicate with our team members and in our productivity. Yeah, so this is where I get so excited about Microsoft's play, right? Because when I decided at the end of last year that I was going to make a new change, I had a number of opportunities both inside and outside the company. And so the thing that really made me say this is where I want to go was, well one, it was most important new technology, AI, on our most precious business asset, our customer's data. So that was very exciting. But what really got me over the edge was Microsoft's approach to AI. Microsoft takes a very different approach to AI than our competitors, right? The heart of most AI is trying to figure out you and you to achieve some result. Now our competitors do that to try and get you to click a button to buy an ad or to buy something you don't need or subvert some government that they want subverted, right? That's none of our objectives. We want to understand you for exactly one reason to make you successful, right? How do we, like in the past, people would throw the rocket at Microsoft and say, oh, you know, when I use Microsoft products, I got to understand the Microsoft org chart, you know, you ship my org chart. What they're really saying is that they have to understand the tools to get their job done. They have to navigate the tools. What we're trying to do is have the tools understand the person to help that person get their job done. So there was this great show, I think it was called The Remains The Day, the movie with Anthony Hopkins. He played a butler and in that he did some research and he talked to the butler of Buckingham Palace who had been there for 50 years. And he said, the essence of a great butler is that he makes the room emptier when he enters. What's that mean? Well, when someone sits down, the magazine that they want is there, the drink that they want is there, it just all works out. Well, that's not my experience with computers today. I mean, how many times do you end up at the end of the day and your spouse says, what'd you do to your day? You're like, I don't know, I don't know, I'm just exhausted. Well, it doesn't have to be that way. What we want to do is to have the computer understand you, understand your objectives and not have some big splashy AI, it just, oh, things just work. Oh, I'm coming to this meeting. Oh, the information I need for that meeting is just there. Oh, it prepped me. It knew that I had a few minutes and so it gave me a few minutes worth of prep and things just flow. And at the end of the success will be when you end the day with more energy than you started. Like that's a big tall effort, but that's what we're going for, the gestalt. Yeah, we found that the word that has summarized this week for us is one that Satya said over and over again and it was trust. So in today's day and age, there's a lot of cynicism and especially looking at big tech companies. You did a presentation talking about AI and social responsibility. You teach out a little bit of it there as to why you believe Microsoft is well-intentioned with AI, but maybe share a little bit more about that vision for social responsibility and where we need to go with AI as an industry as a whole. Yeah, exactly. So there's kind of two key points. First is, I think there's a very vast misunderstanding of the state of AI, okay? It really is best understood as software 2.0 and we've been at software 1.0 for about 75 years and I don't think anybody thinks we're doing a particularly great job at it. I think we started to make progress starting around the 1990s with the core principles of the World Wide Web. That's when we started to really make some progress but we still have lots and lots of problems. So we're at software 2.0, we're at the very beginning of the beginning of the beginning. Now here's the point. The innovators set the field. The innovators set the path. And in AI, it's important for Microsoft to be one of the key innovators here because of our approach. Because we're standing up and saying, wait, there's great promise, there's great challenges, right? There are privacy challenges, there's data bias challenges, there's inclusivity challenges. There are things that really need to be addressed by governments, local legislation and global governments. Brad Smith has been particularly vocal on this and the need for a digital, the only way you're going to solve the problem of autonomous killer robots, which is a real thing, is by a digital Geneva Convention. Microsoft can't solve that. IBM can't solve that, Google can't solve that. Governments need to solve that. And so Microsoft's being very proactive in engaging the communities around these problems. For myself, for instance, I've been working with some of the security researchers to say, okay, well, software 2.0, how do you do threat model on machine learning? Nobody knows. Like literally nobody knows. And so we've been working over the course of the last year to produce a taxonomy of attacks. Now this is the initial thing, but it sparks a conversation as we've shown it to various government people and other competitors. They're very excited about this, about trying to join this in to identify the class of attacks. Because once you can understand the class of attacks, then you begin understanding, well, how do I defend against those? But literally it doesn't exist. So. So talking about autonomous killer robots, I'm very worried now. So how do you, Jeffrey Silver, you're talking about Microsoft's more measured approach. And as you said, you are working with governments and reaching out to policy makers and regulators to talk about these things. Maybe unlike some other technology companies that aren't doing that, how do you, are you a tech optimist at the end of the day? Or does it keep you up at night? Nope, not at all. Not at all. No, I'm a wild tech optimist. People are very pessimistic and I just like, yeah, you know, no. Like let me give you an example, right? There's this thing that says, oh, an autonomous car turns the corner at a high speed and it has to decide between killing two old men and a woman in a baby carriage, right? And that's what, this is a philosophy problem called the trolley problem. Oh, a trolley driver has to pull a switch, and that was like over a hundred years old. In the a hundred plus years that that's positive, there's been exactly zero trolley drivers ever put in this position. It's just not an issue. Look, there are real issues. We do have to work these things. I'd say the biggest worry is not these killer robots or the autonomous cars going wild. It is complacency. It is overconfidence. It says, oh, I got something to work. Let's just ship it. Like there's a lot of brittleness in these AI systems, right? Oh, this works and it can be spectacular, but then this is a complete disaster and that's a complete disaster. So how do we get that taxonomy of like, hey, when do we know when we're done? How do we test these things? How do I have like a secure supply chain for the data models as well as the code itself? You know, so I think that Software 1.0 does not provide us any of the answers to the challenges of Software 2.0. But I do believe that Software 1.0 and its challenges tell us the areas that we need to apply our mindset to. And that's what we're doing. So, Jeffrey, before we let you go, we do need to get the update on PowerShell. I have to say, ever since I've first talked to you, I feel like more and more when I go to shows, I hear people just talking about how it's helping their career, helping their business in doing it. I don't know if it's just because, you know, it was brought to the front of the mind and it's like, oh no, I'm used to seeing that car model out there, but give us the latest on PowerShell, even though you're no longer in that group. Yeah, if I continue to meet with them all the time, I'm very active in PowerShell. So we took PowerShell and made it cross-platform to run analytics, we've talked about that. And I don't know where we were when we talked about that, but basically, we sort of did it for our own purposes, right, we need to manage the world's estate and so we want to have a common infrastructure for doing that. And the joke was at the point is like, look, we're not confused, we don't think that the Unix people are going to greet us as liberators, like, oh, thank heavens, you know, I've been dying under this bash and such. Thank God, Microsoft came to save us, right? There's no confusion. Well, surprise, we shipped it and then the vast majority, the numbers are crazy, how many Linux people are using PowerShell? It's just insane and we don't really understand it, we're out there talking to people, but they just love it. So anyway, so PowerShell version seven is coming out, they'll come out officially at the end of the year, beginning of next year. And this really is the tool that then you can use to manage everything, both Windows and Linux, we have parallel for each so you can do massive scale, but that's the one that really just brings all the pieces together and gains the critical mass. So we're very excited about it. Jeffrey, always a scintillating conversation when you come on the show. Thank you so much for coming on. Thank you. I'm Rebecca Knight for Stu Miniman. Stay tuned for more of theCUBE's live coverage of Microsoft Ignite.