Kevin Scott CTO of Microsoft

(Joel Beasley at 00:00:01) Hello, my friends. I'm starting today's episode with an update. I'm now visiting some of our past guests in San Francisco, and I've been thinking about how grateful I am to be the host of Modern CTO. I want to thank you for listening and joining me this year as we've talked to these unbelievable leaders in technology. I am so grateful to wake up every day and bring you value.

(Joel Beasley at 00:00:24) I've spoken with over 135 CTOs in the past year, both on and off the podcast, from startups to billion-dollar companies. And with this knowledge, I've learned a tremendous amount about teams, leaders, and individual contributors. The one thing that I found to be true for everyone is that learning is not a spectator's sport. You can read all the books, listen to all the podcasts, take all the courses, and talk to the mentors, but change and experience will only take place when you take action on knowledge. This realization has spawned a new endeavor, and I've been working behind the scenes for the past two months with some amazing CTOs, and we have created leaderbits.io, leadership and professional development for technologists.

(Joel Beasley at 00:01:12) You gain experience, develop professionally, and become more valuable through weekly leadership challenges. Each challenge is designed to take 10 minutes and results in real, tangible experience. Isn't gaining experience with a specific skill much better than learning about a skill and never using it? The experiences you've gained are tracked through entries in the Reflect DB. You can track your growth and improvement over time, and you can even invite your team in to participate.

(Joel Beasley at 00:01:39) To sign up, head over to leaderbits.io, where you can find us on the homepage of moderncto.io. And get excited right now because today's guest is going to knock your socks off. Today, we are talking to Kevin Scott, the CTO of Microsoft, and we discuss the resurgence in computer architecture, his optimism when it comes to the future of AI, and Microsoft's Intelligent Edge and Intelligent Cloud initiative. All of this right here, right now on the Modern CTO podcast. Here we go.

(Joel Beasley at 00:02:17) This is the Modern CTO podcast. Kevin.

(Kevin Scott at 00:02:29) Hey, Joel. How are you doing?

(Joel Beasley at 00:02:31) Oh, man. I'm absolutely fantastic. How about yourself?

(Kevin Scott at 00:02:34) I am doing pretty okay.

(Joel Beasley at 00:02:38) They decorated the office for you?

(Kevin Scott at 00:02:40) No, not for me. So I'm in Microsoft's Channel 9 studio space. And so the room that I'm in, they recently upgraded all of the lighting and grip stuff in here. And it looks like, to give it some ambiance, we've got a copy of Windows XP Home Edition, $97.

(Kevin Scott at 00:03:08) Yes. That's pretty cool. Yeah.

(Joel Beasley at 00:03:12) So I listened in on—I watched a couple things. I watched this really cool hologram of you explaining the Intelligent Edge.

(Kevin Scott at 00:03:23) Yeah. That one was interesting. I was very worried that would come off as sort of overly pedantic, but I think it was okay.

(Joel Beasley at 00:03:35) No. I liked it. I really did. I thought it was like a series. I really enjoyed it. I was like, I want to hear Kevin Scott explain more things to me.

(Kevin Scott at 00:03:45) Well, thank you.

(Joel Beasley at 00:03:47) And then I read your article on LinkedIn about technical debt, and it made me want a book from you.

(Kevin Scott at 00:03:57) Yeah. I am in fact working on a book right now that is—it's not really an engineering leadership book. It's called Reprogramming the American Dream, and it's about I have a contrary point of view about what automation and AI is going to do for all of us in the near term. And particularly, what impacts AI is going to have on the working class. I'm far more optimistic than everyone else is, and we've been out and about in all sorts of places.

(Kevin Scott at 00:04:38) I went to Memphis a few weeks ago to talk to entrepreneurs there. I went back home to Campbell County, Virginia, and talked to a bunch of folks who are doing agriculture and light manufacturing and whatnot. And it really is inspiring to go out and talk to these folks who are using technology in all sorts of interesting ways. And then imagining this world where the thing that I'm spending a bunch of my time on at Microsoft right now is to make sure that all of this technology that can sometimes seem almost near magical isn't just a thing that is in the hands of elite engineers, that we, as quickly as humanly possible, democratize access to it so it can just become yet another building block in the toolkit that we make available for people to use to create technology. And I think if we do our job right, it not only isn't just a more democratized thing in the sense that all developers can use artificial intelligence as well as the most highly trained developers.

(Kevin Scott at 00:05:58) But I think you may see an order of magnitude more people being able to, quote-unquote, "program" because, with machine learning in particular, we may be making this transition right now from the sort of paradigm where programmer stares at machine, understands all of its gnarly idiosyncrasies, and then has to do all of the mental gymnastics that are required to transform a problem into a fairly brittle set of instructions that this machine or distributed system or what have you has to carry out to solve the problem, to this mode where we just sort of show the computer what problem we're trying to solve, and then it sort of figures out how to solve it on its own. And I'll give you an example of that. I saw this amazing thing this week, a demo from a designer who was using a new machine learning tool. And they had a cistern at their house to collect rainwater. And they were like, I want to have a system that tells me how many gallons of water I have in the cistern and sends an alert when it gets too low.

(Kevin Scott at 00:07:31) And so they rigged this pulley up with a float on one end that's inside of the cistern and a little stick, piece of wood, tied to the other end of the rope thrown over the pulley. And so when the stick is low, that means the float is high and the cistern is more full. And when the stick goes high, the float is low, which means the cistern is getting empty. And they took a bunch of pictures of this and sort of notated how full the cistern was at a bunch of different positions of the stick and then just fed it into the system. And then it had a fairly accurate map.

(Kevin Scott at 00:08:07) The machine learning system was able to learn a fairly accurate mapping from positions of this stick to how many gallons of water in the cistern. And so this guy didn't write any code at all. He literally showed the system how to solve his problem. And it's completely different from how I would have approached it. I would have put sensors on the thing and done a bunch of math. You know, it's crazy.

(Joel Beasley at 00:08:30) What the next generation is going to be able to do with these technologies. I was speaking with my cousin last week, and the fact that her son can go online, they can go on Amazon and get an Arduino, a little kit, and start programming and do that. That was not possible when I was a child.

(Kevin Scott at 00:08:49) Yeah. Well, I mean, Raspberry Pi, for instance. Again, you go on Amazon, you can pay $30 for this thing, $30, $40 bucks for a current generation Raspberry Pi. And you can plug it directly into your TV with HDMI and hook it up to a keyboard and mouse, and you've got a fully functioning Linux workstation for $40.

(Joel Beasley at 00:09:19) It's crazy. It's the future, Kevin. We're here, man. We're in the future.

(Joel Beasley at 00:09:24) I love how you're optimistic, especially about AI. When you started talking about that, it reminded me of in the late '90s when everybody was scared that the internet or email would make the postal service not do anything or go out of business, but then it did the exact opposite because everyone started doing business and shipping stuff to each other.

(Kevin Scott at 00:09:45) Yep. You know, and I think it's important not to be Pollyanna about it because it's not like there won't be disruption. But the thing that we have to remind ourselves of is that technology and what it does isn't some sort of predetermined outcome that we're all forced to walk this path and whatever it does we sort of have to accept, whether it's good or grim. We get to decide what we want to do with this stuff. And so if we're optimistic and what we want to have happen with AI is that it creates inclusiveness and opportunity and abundance, then that's exactly what we will go do.

(Kevin Scott at 00:10:28) And if we believe that it's going to create further concentration of economics and deprive people of jobs and meaning in their life, then that's sort of what we do. We have to fight against this negative tendency, I think, which is real work on our part.

(Joel Beasley at 00:10:53) Yeah. It's not super easy to always be optimistic, but one of the great traits of a leader is optimism and to be pragmatic in times when you need to step up.

(Kevin Scott at 00:11:06) Yeah. Absolutely.

(Joel Beasley at 00:11:08) So Reprogramming the American Dream, that's next year or...

(Kevin Scott at 00:11:13) Yeah. It's probably going to come out towards the end of next year.

(Joel Beasley at 00:11:16) So there's going to be no talk in there about technical debt?

(Kevin Scott at 00:11:20) Actually, we'll—I'll talk a little bit about it, but it's not a book about how to—my take on how you should run an engineering team. Maybe someday.

(Joel Beasley at 00:11:32) That would be awesome. You can get in on my next book if you want.

(Kevin Scott at 00:11:36) Awesome.

(Joel Beasley at 00:11:36) I'm calling it Everything But Code.

(Kevin Scott at 00:11:39) Oh, cool.

(Joel Beasley at 00:11:40) So the first book was called Modern CTO, and that's what spawned this podcast. And then the next book—I learned how to write a book because everybody, when you write and then put a book out there, everybody hears different things when they read the chapters. So it's really interesting as I was getting feedback from the writing. I was like, now I know exactly how to structure a book, what people are expecting, the type of feedback I'm going to get, the different collections of individuals and what they expect from books. So I began writing the second book, Everything But Code, and that's just about technology leadership and helping the next generation up.

(Kevin Scott at 00:12:18) That's really awesome. Are you having fun?

(Joel Beasley at 00:12:21) Oh, man. I never thought I—I didn't do super well in school. So I didn't ever think I liked writing because they forced us to read the newspaper and write these things called current events. And that was my extent of writing. And so when I got into a habit of writing and started with my 300 words a day, I started to very much enjoy it and find it an excellent medium to express myself. And now it's just something, a habit I do.

(Joel Beasley at 00:12:48) So you know a couple of my friends?

(Kevin Scott at 00:12:54) Oh, do I?

(Joel Beasley at 00:12:55) Yeah. You know Fred at Sift Science?

(Kevin Scott at 00:12:58) Yes, I do.

(Joel Beasley at 00:12:59) Yeah. And then you know Kit Colbert at VMware.

(Kevin Scott at 00:13:02) Yes.

(Joel Beasley at 00:13:03) Yeah. They're awesome people, aren't they?

(Kevin Scott at 00:13:05) They're super awesome people.

(Joel Beasley at 00:13:07) Are you in San Francisco or Seattle?

(Kevin Scott at 00:13:11) I still live in the Silicon Valley. I live in Los Gatos, which probably, topologically speaking, is about as far from San Francisco as it is from Seattle, given Bay Area traffic.

(Joel Beasley at 00:13:30) So what does your day look like? Do you travel a lot?

(Kevin Scott at 00:13:34) I do. I spend a fair amount of time up here in Redmond, and increasingly, I spend a bunch of time on the road chatting with all sorts of people all over the world who are doing different technological things, which is a big change because when I was at my startup AdMob and when I was at LinkedIn, I rarely traveled at all. In fact, my big joke at LinkedIn is I did that for six years. I think I took four business trips in six years, and one of those was for the IPO to go ring the bell on the New York Stock Exchange.

(Joel Beasley at 00:14:17) Was it a digital bell, or did you—it was an actual bell?

(Kevin Scott at 00:14:21) No, I think they actually rang a real bell.

(Joel Beasley at 00:14:24) Oh, it's so exciting.

(Kevin Scott at 00:14:25) It's amazingly old school.

(Joel Beasley at 00:14:29) Well, you want that haptic feedback. You want to be able to ring the bell. You know, you don't want to press a button.

(Kevin Scott at 00:14:35) Yeah. Correct.

(Joel Beasley at 00:14:37) Yeah. When I talked to Fred and heard—he was on the show two weeks ago, and we got along super well. And now I'm going to fly out there and be out there next week, so I'm going to have lunch with him, I think, Tuesday, so in four or five days. What Sift Science is doing is unbelievable.

(Kevin Scott at 00:14:56) Yeah. It's really great.

(Joel Beasley at 00:14:58) And then I looked up—we're doing the show prep, and I saw that you aren't just connected with him on LinkedIn. You're involved too.

(Kevin Scott at 00:15:06) Yeah. So one of the cofounders of Sift, a guy named Brandon Ballinger, worked for me at Google, and he's one of those guys where I've actually invested in both of his companies. And if he goes off and creates five more, I'll invest in those as well. They're just some of those people who are just so awesome, and they have such good ideas. And more importantly, they're just able to make stuff happen. And Brandon is one of those guys.

(Joel Beasley at 00:15:35) Oh, that's awesome. Maybe I'll have the chance to meet him next week.

(Kevin Scott at 00:15:39) Yeah. He's off at another company. He's doing this thing called Cardiogram right now. So they're building a system that takes EKG, or I should say heart rate data from smartwatches and fitness bands and applies a bunch of modern deep learning to the data streams that they get to try to diagnose serious cardiovascular conditions. And they're doing super well.

(Joel Beasley at 00:16:08) I got a notification that says you're going to have a heart attack in 3, 2, 1. That wouldn't be good.

(Kevin Scott at 00:16:14) Actually, it's sort of—you know, it's better than that in the sense that a bunch of these cardiac conditions that you can have, you can recognize them in data very, very, very far in advance of when you're symptomatic. And that's what we want, I think, in general with medicine because the problem we've got right now is people go to the doctor when their body is sending them a signal that says you're sick. And oftentimes by that point, you had a bunch of biological damage done and you've got sort of second-order effects of whatever started your cascade failure. And it's super expensive to get yourself back to health. And in some cases, you've hit a point of no return, and you have an incurable condition. And so being able to—in this world that we're increasingly living in where your body itself is more instrumented and you've got these sophisticated AI techniques that can look at this data and help to very, very early tell you when you're ill. You know? So this Cardiogram stuff's one thing. There are a bunch of folks, including us here at Microsoft, where there are a bunch of things where you can, from a blood sample, do a full PCR screening of all of the biological material that's in a small vial of blood and then use that genetic information to do things like, oh, we understand what your T cell distribution is inside of your body.

(Kevin Scott at 00:18:02) And, you know, if we can learn the mapping between T cells, which are the individual components of your immune system that fight particular diseases, you may be able to detect from blood draw when you very, very, very early—like, when you have cancer or other serious diseases. So all of this stuff's really exciting.

(Joel Beasley at 00:18:26) Yeah. My brother and stepmom are both physicians. So they're always asking me about technology and giving me ideas and saying, "Hey, you should make this." And I'll Google it, and I'll be like, "Yeah, there's six companies funded that are making it right now. It's on its way."

(Kevin Scott at 00:18:42) Yeah. But that's good. That's the thing where you want a little bit of overinvestment. I would rather have a hot competitive marketplace in health care rather than, you know, a hot competitive marketplace in some latest picture-sharing thing.

(Joel Beasley at 00:19:04) My job.

(Kevin Scott at 00:19:05) Those are the—

(Joel Beasley at 00:19:05) Those are the things I want to hear about. I want to hear about the problems solved or the problems we were making progress on that really help us as a people, which I'm not surprised at all that Microsoft is involved in working on big picture stuff because that's what Bill Gates has chosen to spend the rest of his life doing. Right? Yeah. That guy's awesome.

(Kevin Scott at 00:19:27) And it's—I mean, you spend a little bit of time talking with Bill about some of the things that they're discovering. He gave a talk to the—Microsoft does this CEO summit every year where we bring the CEOs of a bunch of big companies. And one of the things that Bill does every year there is he gives a little talk. And, you know, it's really—I find this whole notion of engineers and technologists being optimistic to be the funniest thing in the world because I was this cynical little grump when I was growing up. And I still am, just, you know—as an engineer, your job is to look at the world and identify everything that's broken or suboptimal so that you can go fix it. And so, you know, the fact that we're able to summon any degree of optimism always strikes me as funny. But, you know, Bill was talking about—they are doing foundational work right now in mapping the microbiome of the human gut. And they have discovered a bunch of really interesting things.

(Kevin Scott at 00:20:40) Like, we in the developing world may no longer have a nutrition availability problem. It could be that—and this isn't universally true, but in a bunch of places, there is sufficient nutrition available to people. And the problem that we've got now is that various conditions in the gut caused by the composition of your microbiome can affect nutrition absorption and cause all sorts of, you know, things like growth and development issues. And so, you know, they are hot and heavy pursuing both the basic research and then figuring out, like, how—if these research hypotheses prove to be true, how do you go off and solve that underlying gut microbiome problem so that we can have a healthier, more productive world.

(Joel Beasley at 00:21:37) Was that, when you joined Microsoft, was that your first time getting to meet Bill, or had you met him before?

(Kevin Scott at 00:21:43) I had never met him in person before. So, funny enough, I was an intern at Microsoft in 2001, and I went to a couple of things when I was an intern at Microsoft Research where Bill was speaking. And in fact, one of the funny things when Microsoft bought LinkedIn, we were doing some sort of presentation to the Microsoft executive team. So, you know, Satya is sitting in—we go into this room and there's a little stage and there are a couple hundred chairs, and Satya is sitting in the front row and then there are all of the vice presidents of the company—just a hundred or so people. And I'm up on stage talking about LinkedIn, and I have this moment where, "Wow, this all feels really, really familiar." And then I realized that I had been in that exact same room fifteen years prior at an intern dinner that the company had hosted and where Bill had been on the stage in almost the same position that I was standing in, talking to the interns. I'm like, "Holy crap. This is really weird."

(Joel Beasley at 00:23:00) You're giving me goosebumps, man.

(Kevin Scott at 00:23:02) Yeah. It was so crazy.

(Joel Beasley at 00:23:05) So, Satya—I find you. That—go ahead. That guy's transformed Microsoft, man. He's done unbelievable things. I put out to my audience, you know, "Send me questions because I'm talking with the CTO of Microsoft." Very excited. And I got back so much, but one of the things I got back a lot was the topic of how well Microsoft has transformed in the past, you know, decade going from where they're at now to completely embracing the open source community all the way to acquiring GitHub. That's just been amazing.

(Kevin Scott at 00:23:42) I love it. Satya.

(Joel Beasley at 00:23:44) Thank you, I guess.

(Kevin Scott at 00:23:46) Yeah. No. You're entirely welcome. I, you know, and I'll pass that along to Satya. I think he's done an incredible job as well. I've known him for seven or eight years now, and he's just done an extraordinary job.

(Joel Beasley at 00:24:04) So are you the first CTO of Microsoft?

(Kevin Scott at 00:24:07) I think, technically, I'm the second CTO of Microsoft. So the first was Nathan Myhrvold, who was an early, early employee, and he left in 1999. And we had folks playing, you know, arguably the same role, but with a different title in the interim. So there were a few people who were chief software architects of the company. So Bill played that role for a while, and then Ray Ozzie was chief software architect when Bill retired. But, yeah, I think, you know, as far as the exact question, I'm the second CTO in the first one since 1999.

(Joel Beasley at 00:24:51) I'm gonna say you're the first. I'm gonna say you're the first modern CTO. How about that? Oh, man. So the video that I had saw about where you were cartoonified, right, if that's a word, they were talking about the intelligent edge and the intelligent cloud. And then I saw a little Google alert for myself that Microsoft Build conference, you know, the CEO addressed and spoke to developers about the edge and the cloud. And so can you give me a child's like explanation of what this is? Is this a new direction for Microsoft?

(Kevin Scott at 00:25:28) Yeah. So well, let's just talk first about where the technology is headed over the next few years. Perfect. So one of the really interesting things that's happening right now is silicon is getting much more capable, particularly the type of silicon that you need to build AI models and to perform inference in real time. And so it's really crazy, I think, right now. When I was in undergrad and when I was in grad school in the nineties, you know, high-performance computing was still a thing. Like, Cray was selling these crazy big computers to all sorts of folks and, you know, you had this fixation inside of the media on who had the fastest supercomputer in the world and, you know, we were building bigger and bigger machines all the time. And there was a bunch of innovation happening both at the microprocessor level as well as at the distributed computing level. And then, you know, we sort of hit this phase in computer architecture where big, loosely coupled, reliably networked systems of commodity computers with commodity components were the best way to solve problems. I mean, this is basically how the Internet's been powered for the past twenty years or so.

(Kevin Scott at 00:27:12) But we're seeing this resurgence of innovation in computer architecture right now that's really incredible and at least as interesting as the stuff that was happening in the late eighties and the nineties. And what will happen very quickly over the course of the next few years is this power in silicon is gonna get to be really cheap. So you're gonna find it in devices that have consumer price points. So that's one trend. And then this other interesting trend is that there are just more devices in general connected to the Internet every day. So, you know, pick your study. The Gartner one is sort of interesting in that their data predicts that we're gonna get to somewhere on the order of 20 billion IoT devices by 2020.

(Joel Beasley at 00:28:11) Oh, no.

(Kevin Scott at 00:28:11) Which is more or less double from where we're at right now. And so, you know, if you think about that, we've got a billion PCs, you know, two and a half billion smartphones, and we'll very shortly, in fairly early days, have 20 billion IoT devices, which because of this, you know, interesting AI silicon trend are gonna be smart devices. And then you've got—because of these smart devices, you know, IoT has been this weird thing for a while in that when you want your IoT devices to do something in the world, you have to go instrument the world. So it's very expensive. You know, if you wanted to build an IoT irrigation system, for instance, to help you manage how much water you're dispensing on a crop, you'd have to go put hydrometers in the field and, you know, do all sorts of things. Right? But, you know, perception capability that this AI silicon is gonna give devices relatively cheaply means that computer vision can solve a bunch of these instrumentation problems. So instead of instrumenting the field, you just point a camera at it or fly a camera over the field, and then you use a model to build a hydrology model of the field. You know? So that's super interesting. And, you know, so if you think about how things are playing out right now, most IoT devices—so these things live on the edge, the very, very edge of your network. They're the things that directly touch human beings or touch the physical world. Most of them are like these vertically integrated systems. So somebody's had to build something from scratch, like, whether it's your home automation system controller or your smart thermostat or your intelligent speaker or, you know, the smart system that's in your refrigerator.

(Kevin Scott at 00:30:29) So all of these things, you know, somebody had to decide, "I'm gonna build this into my product," and, you know, "I gotta pick the device form factor," and "I gotta pick my peripherals. I gotta pick my operating system. I've gotta integrate all this stuff together," and then "I'm gonna have this bespoke software stack that I'm gonna build to implement my application and then to manage its ongoing operation in the customer's environment." And it's like, all of this stuff is gonna sort of collide in a really messy way unless we, the tech industry, come in to provide the software stack for this new world that is emerging. The world's emerging whether we want it to or not. This is just sort of what's happening. And so we have to figure out how to make these systems fundamentally more secure. We have to figure out how to make developing these systems more accessible to programmers. We have to make sure that the capabilities that get built into these devices are more useful to customers. And so that's, you know, sort of—that was far and away not the short explanation for the intelligent edge, but this is the label that we put on this phenomenon. And we, Microsoft, are actively doing research and selling a whole bunch of products into this space right now to help solve the set of problems that are going to emerge over the next several years as this ecosystem becomes more and more populated with more and more compute and AI and sensor capability.

(Joel Beasley at 00:32:26) Yeah. It's smart. I mean, I'm always a fan of the business model of, you know, building the underlying technology or developing the standard. Right? Because if you develop the standard, that's—you're solving the big problem and people adopt it. And is that a way for me to look at the intelligent edge as sort of like an IoT standard way of interacting with devices?

(Kevin Scott at 00:32:47) Yeah. I mean, so I think it's gonna be a bunch of different things. But, like, so for instance, one of the things that I think this ecosystem needs is you need an intelligent edge runtime, because the things that an intelligent edge application needs to do is probably gonna be relatively consistent across a broad range of applications. And so rather than have everyone write the same thing, you know, over and over again, let's provide this thing. Let's make it open source. Make it as generally useful as possible. And this might power something like a brand new event loop for doing software development. You know, the traditional event loops are, like, wait until the key is pressed or wait until somebody clicks a mouse. And the new event loop might be, you know, streams of data being exchanged between nodes in your edge network where you are doing inference in real time over the streams. And then the events are the inferred events across this raw sensor data. So it might be, "Oh, I just recognized the person," or "I just detected a speech intent," or, you know, "I've detected a gesture or a policy violation," or there's just a whole new rich layer of semantic events that you might architect your application around. And for some of these things, you shouldn't have to build your own model for doing this stuff, for speech recognition and object recognition and a bunch of other stuff. Like, the runtime just ought to provide these for you, so you're not having to worry about, like, "Oh, crap. I gotta go find the data and train the model or find some pre-trained model and hope that it—" No. That should be in the runtime.

(Joel Beasley at 00:34:38) Yeah. You don't want the project to be creating the feature. Right? You want to just grab the feature and then apply it and solve bigger problems.

(Kevin Scott at 00:34:45) Yeah. We want the engineers to solve the problem beautifully for their customers and let the runtime, let the infrastructure sort out the tedious repetitive bits.

(Joel Beasley at 00:34:58) Yeah. I've been seeing a lot of this low-code movement. Have you been hearing about that?

(Kevin Scott at 00:35:03) No. Actually not. Tell me about it.

(Joel Beasley at 00:35:05) I have a junior understanding of it. But I've had a couple of guests in the past week or so, and they've been a part of these low-code companies. And I had them explain it a little bit to me. But, essentially, it's—there's such a shortage of developers. They're building these really sophisticated softwares that allow just generally intelligent people to build out basic business process applications. Yep. Yeah. And I saw that—you know, my conversation with them was I saw this happening, you know, fifteen years ago with these basic interfaces that would allow you to write code a little bit easier by structuring components. And I'm like, "Well, this is just the way more advanced version of that."

(Joel Beasley at 00:35:52) It was very smart.

(Kevin Scott at 00:35:54) Yeah. I look. And it's totally unsurprising. The tech industry itself is growing at a crazy clip, which, you know, hoovers up more and more developers.

(Kevin Scott at 00:36:07) And the interesting thing, we looked at the LinkedIn data a couple weeks ago right before the GitHub announcement. And, you know, the data in LinkedIn shows that the number of developers outside of tech is growing at a strong double-digit percentage and faster than the number of developers, than the rate of new developers coming into the tech industry itself. And so it's just sort of crazy. The non-tech companies and, you know, Marc Andreessen predicted this years ago—every company's becoming a tech company.

(Kevin Scott at 00:36:48) Development is becoming a core part of how you do your business across all types of businesses. And there just aren't enough developers in the world to satisfy all of that demand. So it doesn't surprise me in the least little bit that you're starting to see these companies where folks are trying to build high-tech without needing as many developers.

(Joel Beasley at 00:37:14) Yeah. That's one of the recurring topics that comes up on the show is when the companies get into that sweet spot of 30 to 50 engineers where they, you know, they started out as eight or one core team. When they get to that 30 to 50 and they start having the teams of teams issue, then it becomes, how do we convert our subject matter experts or our lead engineers to be leaders in the sense that they can run a team or a team of teams? And so that tends to be a shortage that's also occurring in the industry, is getting people out of the subject matter expert and into the leadership.

(Kevin Scott at 00:37:50) Yeah. Absolutely.

(Joel Beasley at 00:37:51) Which is why the podcast has done pretty well. People come out here, they're like, let's listen to the CTOs and see what type of leadership information they have. I'm excited because, you, my friend, you have a podcast starting, and I am your first subscriber.

(Kevin Scott at 00:38:08) That's awesome. Thank you.

(Joel Beasley at 00:38:09) You have a fan.

(Kevin Scott at 00:38:11) We're, I think we're gonna have our first episode go live on June 26. So the podcast is called Behind the Tech, and the basic idea is I'm gonna talk to really interesting people who are doing great things in tech and, you know, in some cases, who had long careers of doing really interesting things in tech, who, in a whole bunch of different ways, have had enormous impact on their lives, but they tend to not be the people that get mainstream media attention or the same flavor of hero worship that, you know, maybe a founder or a successful entrepreneur might. And it's not like I love founders to death. Many of my very dearest friends are founders. But, you know, we also have people like Anders Hejlsberg, who is the subject of my first podcast, who perhaps wrote the very first integrated development environment.

(Joel Beasley at 00:39:24) What?

(Kevin Scott at 00:39:25) When he was in school and went on to join Borland, and he built Turbo Pascal and Delphi and a bunch of their really awesome developer tools, and then came to Microsoft and was the language designer for C Sharp and was the principal architect for the C Sharp language for many years and who now is the lead on TypeScript. And, you know, this guy, for me, I chose him as the first guest because the first computer science class that I took, the real class, I got lucky when I was a senior in high school and got into this thing called the Central Virginia Governor's School for Science and Technology, which is in Lynchburg, Virginia. They took two kids from each of the local high schools in a three or four county area. And I went there when I was a senior, and I took a real computer science class, and they taught it in Turbo Pascal 5.5, which Anders wrote. And it literally changed my life.

(Kevin Scott at 00:40:33) I don't know that I would have had a career in tech if it hadn't been for this thing that Anders made. And, you know, that to me is worthy of admiration because he did the same thing for hundreds of thousands of other developers who use his tools to make a huge range of products, from the payroll application that I wrote in Turbo Pascal 5.5 for my dad's construction company as soon as I figured out how to do something like that to big systems that generated tons of revenue and had impact on tons of customers. So that is going to be the flavor of the podcast.

(Joel Beasley at 00:41:20) Oh, that's awesome. I have a recommendation for you. So there's this guy named Mitch Vandine. He did the, he came on my show because he's, well, he listened to the show, and we became friends, and he came on and we started talking. And I find out halfway through the episode that he was responsible for doing the acceleration of the mouse. When as you move, the mouse moves faster, right? So he wrote the formula for the mouse acceleration.

(Kevin Scott at 00:41:49) That's so great. Yeah. Stuff like that, you just don't hear about all that often. Those are the amazing things that make our tech world work.

(Joel Beasley at 00:42:01) Yeah. He was at Xerox in the eighties.

(Kevin Scott at 00:42:04) Yep.

(Joel Beasley at 00:42:04) They, he was just working on a project, so he didn't know what to do. So he grabbed a giant book and started flipping through documentation, came across, oh, here's the mouse. I can do something with it. Wrote a little program to invert the mouse and installed it on everyone's computer overnight. And then they all came in and their mice were inverted, and he got in trouble for, you know, as a joke.

(Joel Beasley at 00:42:24) And then he dug a little deeper and played with the technology a little bit more, and then he's like, I want the mouse to move faster. You know, Xerox in the eighties, just think about how long it would take you to get the mouse across the screen. Yep. So yeah. He's a really cool guy.

(Joel Beasley at 00:42:39) Thank you so much for coming on and hanging out with us. I do have one last question if that's okay.

(Kevin Scott at 00:42:45) Yeah. It's totally okay.

(Joel Beasley at 00:42:47) Okay. Are you a fan of Elon Musk?

(Kevin Scott at 00:42:50) I am. I admire Elon a lot. I mean, there aren't many folks who can go off and create a space company and a payments company and a company that solves, maybe will solve both the transportation and the energy crisis that we have.

(Joel Beasley at 00:43:08) So let's say that he creates a time machine, right? Yep. He invites you over. You jump in the time machine. You get to go back ten years and give yourself one piece of advice. What would it be?

(Kevin Scott at 00:43:19) That's a super interesting question. So I have learned a ton of painful lessons over the years, and I think most of them I wouldn't want a shortcut because the process of learning them has been just as valuable as the lesson itself. But there are a few things that, you know, I'd give myself some advice on. One is, and maybe this would have been more useful fifteen or twenty years ago than ten years ago, but I've had a number of points in my career where folks told me that I needed to be more patient.

(Kevin Scott at 00:44:04) And it sort of sounds like a reasonable thing. You should be patient. You don't need to run around like you've got your hair on fire if you have hair all the time. But I think I would have advised my younger self to ask the question or try to get to why it is that the person's asking me to be patient. You know?

(Kevin Scott at 00:44:30) So sometimes people, especially when you're thinking about a career development conversation, you want to be promoted, you know, you want somebody to give you the next degree of responsibility or whatnot. And they tell you, be patient. It's just not actionable. There's nothing to do with that other than wait around. And that's not a productive use of your time. And, you know, sometimes people are telling you that just because they're being lazy, and sometimes they tell you that because, you know, they're waiting for you to ask the next question about, like, what really should I be doing while I'm being patient to make myself more ready for this next thing? And sometimes they're just stalling because they've got a problem that they're trying to solve.

(Kevin Scott at 00:45:21) I can't give you this job because it's gonna piss off these other people. And that's not your problem. That's theirs. And it's a real problem, but they should just tell you upfront why they're asking you to wait rather than have you walk around being frustrated. So, you know, that would be the advice to my younger self is don't just accept "be patient" as a prescription for how you should behave.

(Joel Beasley at 00:45:52) I like it. That's great advice.

(Kevin Scott at 00:45:54) Awesome.

(Joel Beasley at 00:45:55) Kevin. Dude, we did it. We made a podcast together.

(Kevin Scott at 00:46:00) It's been a ton of fun.

(Joel Beasley at 00:46:02) Yeah. And now it's complete. It's been just the most amazing time ever with you. I'm out in San Francisco next week, but it looks like you're traveling. So I'm gonna have some lunch with Fred at SIF Science and let you know how that goes and see all those great people out there.

(Joel Beasley at 00:46:16) But if I'm ever around and you're around in the same town, I would love to see the Microsoft offices, you know, take a look inside the empire. I'm super interested personally in HoloLens. I'm gonna get one the next time I can get my hands on one so I can program up on it a little bit in my free time. But, yeah, it would be really cool to shake your hand, say hello, and see what you're working on in person if our travel schedules ever collide.

(Kevin Scott at 00:46:41) Yeah. For sure. When you're in town, let us know. Somebody on my team, at the very least, would be more than happy to give you a tour of Microsoft even if I'm not in town. But I would love to meet you and maybe go grab a coffee or lunch.

(Joel Beasley at 00:46:54) Sure. Your team, by the way, interacting—I have my Jackie and my Jake interacting with your team. Nothing but the most beautiful amazing things to say. It's been an absolute pleasure interacting with Microsoft from the moment I reached out and until we've wrapped up this podcast. It's just been awesome.

(Kevin Scott at 00:47:11) Awesome. Well, I really appreciate that, and I'm sure they will be happy to hear it.

(Joel Beasley at 00:47:17) If you ever need anything at all, anything at all, just let me know. Send me an email, reach out, and I'll be there, whatever you need.

(Kevin Scott at 00:47:24) Awesome. Thank you so much, and vice versa. Don't hesitate to reach out to us if you need something.

(Joel Beasley at 00:47:30) Fantastic. Thank you so much, Kevin.

(Kevin Scott at 00:47:32) Alright. Talk to you soon.

(Joel Beasley at 00:47:33) See you.

(Joel Beasley at 00:47:39) Thank you so much for listening to the Modern CTO Podcast. Share this. Get the word out. Thank you guys so much. I couldn't do it without you. I appreciate it. You guys are the absolute best.