A company doesn’t have to be cool to make an impact — just ask Microsoft.
The tech giant is, depending on the day, either the most-valuable company on earth or close to it. And it has done so via a radical reinvention that embraces building the tools to help others do their best work.
“I really like what (Microsoft CEO Satya (Nadella) says over and over again: our job is to make other people cool,” Microsoft CTO Kevin Scott said in an interview with GeekWire. “We don’t need to be cool ourselves.”
Scott has been around the block in the technology world. He took the job of CTO in January 2017, following Microsoft’s acquisition of LinkedIn. Scott led LinkedIn’s engineering efforts during his six years there, helping take the company public.
LinkedIn’s acquisition wasn’t even the first big money deal that the technology executive has been a part of. He was the vice president of engineering and operations for mobile advertising company AdMob, which sold to Google for $750 million in 2010, a deal that remains one of the tech giant’s largest.
Scott is also an angel investor and has his own foundation focused on hunger, early childhood education and women in technology. Scott is part of a group of tech executives who have banded together for a plan to bring tech jobs to rural Iowa and other places that have felt the pain of slowdown in recent years and decades.
Today, as CTO at Microsoft, Scott says he helps set the table for the company’s overall technology strategy. Scott works with a small team that tracks the company’s major initiatives. He has his own “three to five things” that occupy his personal attention, though he played coy on what exactly they are.
The company’s push in artificial intelligence tops Scott’s priority list. Microsoft is infusing the technology into its own hardware and software offerings and building tools for others to use. The company has also committed millions of dollars of work to various AI for Good initiatives to implement the technology in areas like accessibility and environmental protection.
Scott recently sat down with GeekWire to talk about how he spends his days at Microsoft, the company’s bet on AI and why diversity should be a top priority for tech giants. The discussion has been edited for style, clarity and length.
GW: So what precisely do you do at Microsoft and what are your main objectives right now in your job?
KS: After the LinkedIn acquisition closed I was looking for my next thing, and simultaneously, Satya had this interesting problem that he was trying to solve. At Microsoft there’s tens of thousands of engineers and 130,000-plus employees. He was the catchall for too many technical things. He was the one trying to spot a duplication of effort on technology or a hole or an error in a technology strategy in our portfolio.
He was looking for someone who could help with two things. His term is doing the “left-to-right scan” across the entire company and trying to make sure we have a sensible technology strategy, where we’re building the right sets of things to empower the products and experiences for our customers over the next several years. And then making sure that over a three-to-five year time horizon we aren’t either doing things or failing to do things we’d regret doing or not doing down the road.
The stuff that’s going on here is so vast and broad. It’s everything from building a quantum computer to making video games and everything in between. And so what I built is a small team of technical experts who have vertical domains of expertise and work with the teams across the breadth of Microsoft deciding “this is the data strategy, this is AI strategy, this is how we’re thinking about this emerging intelligent edge computing platform.” And then quarter by quarter, I reassess the three to five things that are most urgent for me to be paying direct attention to.
GW: What are your current three to five things?
Most of them I can’t say, but a bunch of it is around artificial intelligence, making sure that we have the right set of AI investments and anticipating these big trends that are coming in. When you believe that there are 10s of billions of IoT devices coming online over the next several years, all with increasingly high-fidelity sensors and increasingly cheap silicon so that you can do inferences on relatively complex models right there at the edge, it changes how you build user experiences and what the programming model is for software in that particular environment.
We think we’ve got a lot of data right now. It’s nothing compared to the volumes of data that we will have relatively shortly as these IoT devices start, in high-fidelity, mapping the real world on to digital, and the problems you have are going to be different. Say, I’ve got a bunch of cameras in this environment that I want to make sense of, to do something valuable and compelling for customers. But with the frame rate of all of these cameras, I can’t ship it all back to the cloud. I’ve got to do something with it here on the edge. Because of bandwidth restrictions I may have intermittent network connectivity if I’m on an oil rig, or in the self-driving car, or I have really serious legitimate privacy concerns, and I just don’t want this data coming out of my environment.
GW: What’s your elevator pitch for why people should take Microsoft seriously in AI? There are so many companies in this realm fighting for talent. Why does Microsoft matter in AI?
KS: The first thing that got to human parity with image labeling, the first thing that got to human parity with professional translators from Chinese to English, the first thing that got to human parity with question answering, we were the first ones there on a huge number of these things. We’ve got this long history of pushing the boundaries of what’s possible with machine learning.
The thing that we are really well situated to do — given our developer ecosystem heritage and our platform heritage — is build AI so it’s not just for our own benefit, going only into our own products, and it’s not even just a set of tools and infrastructure that are for the expert AI developers. We’re really trying to build a set of things that lets everybody participate. It’s in our DNA, right? You build Visual C++ for the professional developers building these high performance applications. You build Visual Basic and macros in Excel to let people be more productive and do more automation.
GW: It’s funny; I was going to say yes, this is Microsoft’s DNA. but to me it’s more like Windows and Office. I think that raises a fascinating question as you’re looking at this, because you can see where Microsoft is, through its cloud technologies and developer tools, replicating that platform for the cloud and AI. But I don’t think it’s obvious yet, at least to the outside world what the killer app will be, like the word processor or the spreadsheet at work for the personal computer. Will Microsoft be the one creating those killer apps or will you be the one enabling others to create them?
KS: Yeah, I think it’s going to be a mix of the two. We have some really exciting things coming down the pipeline for productivity.
GW: Under your top three to five things you’re focusing on right now?
KS: There’s some really interesting stuff — we rearchitected the entire back end of Office 365 to support these collaborative intelligent experiences — it’s a bunch of goodness coming that I’m super excited about. I get pleasantly surprised all the time as the outsider coming in. One of the biggest surprises to me is how sophisticated the technology stack is that Bing has built, which is increasingly powered by really sophisticated machine learning.
We have this beautiful cloud platform that gives us the opportunity take all the learning that we get from building our own products that we’re selling to consumers and packaging that infrastructure up in a way that we can give to every developer everywhere so that they can build their own thing and have less of a struggle with it.
Even though I was not by training a machine learning expert, that was the first thing I worked on out of grad school. It has remained challenging to build these systems in a world class way. It’s incumbent on us to build these tools and enable infrastructure for folks, especially given how scarce that expertise is in the world.
GW: It’s obvious in the stock market that Microsoft is valuable again — and I think Microsoft is relevant for some segments of the population. How do you make Microsoft cool again?
KS: I don’t know that that’s even the objective. I really like what Satya says over and over again: our job is to make other people cool. We don’t need to be cool ourselves. We need really good technical folks to understand that there are amazing things that we’re trying to do here at Microsoft, and you can come here and learn and have huge amounts of impact and grow your career with us. We as a company are really happy when somebody can take our technology and then themselves make something that’s great and interesting and valuable in the world.
GW: How do you diversify your engineering leadership?
KS: We have to create as many opportunities as humanly possible up and down the stack to get folks involved. I have a true, representatively diverse team that I got to build from scratch. My right hand is a woman. I don’t know what the exact percentages are, but we tried to be about 50/50 inside of the office of the CTO in terms of gender balance.
We have a bunch of underrepresented minorities inside the team, and it’s fantastic having that diversity. We tried to bring in people who were already super collaborative by nature because you can’t get anything done at a job like this if your first order of business isn’t “I’m here to help you all be successful. I’m here in service of what you’re doing.”
Having that diversity inside of the team has been magnificent. I run this program I inherited from a member of my team who created it called LEAP. What LEAP does is it tries to provide an on-ramp into Microsoft and into a technology career for people who have non-traditional backgrounds. You may not have the set of things on your resume that will get you through a normal technology company interview process. We bring you in and put you in an apprenticeship program and give you on-the-job training that we believe you’re going to be able to absorb and adapt really quick. Just focusing on that and not even paying attention to diversity metrics, you get more diversity that way.
GW: What are the practical implications of the diversity that you have on your team?
KS: You get better conversations. When you are building a bunch of enabling technology for everyone in the world, it’s really good to have people who come at it from different perspectives and points of view. It’s really easy if you are a computer scientist, and you’re someone who looks like me, to get really myopic about what would be a good thing to do in service of customers. When you have diversity in the team — and it’s not just gender diversity or ethnic diversity, it’s diversity of experiences — your path to the field is different. You get all sorts of delightful, inspiring, revelatory conversations happening. I don’t have a set of metrics where I can tell you this is 73 percent better because of this. I just feel it in my bones. It’s the most delightful team I’ve ever worked with. And it’s just fun to solve problems with them.