PITTSBURGH — So much has changed, in this city and in the technology world, since Rick Rashid arrived in 1979 as a computer science professor at Carnegie Mellon University. Pittsburgh in 2018 is defined by self-driving cars on the streets, not by steel mills on the river. And many of the advances for which Rashid and his colleagues laid the groundwork are now becoming commonplace in our emerging world of cloud computing and artificial intelligence.
After working on projects including the seminal Mach operating system kernel, Rashid left CMU in 1991 to lead Microsoft Research in Redmond, including initiatives that formed the basis for some of the most important technologies of the tech giant’s modern era — such as AI, the cloud and machine translation.
We covered Rashid’s work at Microsoft for many years, prior to his retirement 18 months ago, so it was a pleasant surprise when we heard he was back in Pittsburgh recently for an appearance at CMU, which happened to coincide with our month-long project covering Pittsburgh’s technology scene. I hustled to CMU’s Gates Center for Computer Science from GeekWire HQ2 in the Lawrenceville neighborhood to catch up with Rashid before his talk.
Over the course of our 45-minute conversation, we discussed Pittsburgh’s transformation, the technology revolution, Microsoft’s revitalization, and the unavoidable topic of the moment in this and 19 other communities around the country: Amazon’s search for a city to host its second North American headquarters.
Continue reading for edited highlights from the conversation.
Q: What does this place mean to you?
Rashid: This is really where I started my career. This is where I learned what it was like to run and manage a large research program. A lot of my own philosophy of how to conduct research and how to manage a research organization really came out of the philosophy of Carnegie Mellon. My model for Microsoft Research was the computer science department at Carnegie Mellon. You treat people like adults rather than treating them like children. Especially in a research environment you want to give people the license to take a lot of risk and to do things that might or might not work out, but that they believe are the right things to do in their research areas.
You’re in a unique position now, coming back with fresh eyes looking at CMU and the tech scene here. What’s changed?
Rashid: Some things are the same. This is still a much more collegial environment than other places are. I think that’s a key part of what made CMU really unique and interesting even from the beginning. It was part of the reason people would come to CMU versus going to Stanford or MIT or someplace else. This was much more of a community, people helped each other, people worked together, people shared each other’s work, they used each other’s work and they benefited from that.
At the same time, my gosh, it is so much bigger, wow. The growth has just been spectacular. The entire CS and robotics organization was just in Wean Hall back in the old days, and it’s grown so much. You just walk through the hallways, you see so many students now. I think that’s a credit both to the university and to the people running the school of computer science, but also I think it’s a statement that CMU has become a very desirable place for students.
What has changed in Pittsburgh?
Rashid: The number of tech jobs in the area has just exploded. We were driving around in East Liberty, with all the new buildings, and you’ve got the Google facility there now. Amazon, Intel, all have places around here. A lot comes from the fact that you’ve got some strong universities here. They’ve grown but they’ve also brought growth and changed the character. … When I was here, there were some good places to eat but it wasn’t exactly a culinary paradise.
Q: As you look at the things that the CMU School of Computer Science and Pittsburgh are specializing in now, which areas are you most excited about watching in the next 3, 5, 10 years?
Rashid: Personally I’m super excited about where we’re going with artificial intelligence, deep neural networks, reinforcement learning. We’re also starting to see a lot of really good work as people seem to be understanding the intellectual underpinnings of some of those areas, as well. Back when I was here, it was always a very big hotbed of robotics research, but now I think that takes on more urgency as we can do more and more with robots. Our capabilities of using robots for a broader collection of tasks has really changed our thinking about the role computers can play in our society.
Tesla is not really selling you a car, they’re selling you a robot that can take you places. All of these functions in this vehicle are under central computer control and they can be programmed, they can be changed. They can be changed tomorrow if there’s an update. I think we’re starting to see this happening now in other kinds of products, where manufacturers are starting to realize that if they put all of the functions of the device under central computer control, that now these new AI techniques, given the opportunity, add a tremendous amount of value to what those devices can do. That’s a change.
How do you feel about what’s going on at Microsoft these days?
Rashid: I’m very positive. As Satya (Nadella) took over the company, he had the license to make changes, and in particular to start to refocus the company into areas that were going to be the new growth areas in computing. The pivot toward the cloud and the pivot toward mobile devices and the pivot since then toward intelligence are all things that Microsoft Research was really pushing for quite a while. I’m really happy to see that now having that type of an influence.
Also, I think it’s changed Microsoft as a company. Now we’re really looking at what I think are going to be the really important areas for the next 5 or 10 years, in terms of how do we change people’s experiences. What is the relationship between people and the computing resources they use? We’re changing not just the interfaces but we’re changing what is able to be done in fundamental ways that I think are really exciting.
What’s the best example?
Rashid: I’ve stopped really thinking about what devices I use because the only things that are permanent and interesting to me are in the cloud. I could be doing something on my phone or doing something on my PC or switch between PCs and I’m really working on exactly the same documents, I’m really sending the same email, I’m really doing all the same things. Everything is focused on the cloud. All of the adaptations about me are now being pushed down to those devices.
This is something that I was pushing quite a long time back if you look at some of my older talks. At some point the computing devices simply become invisible to you. It’s not like they’re not there, but they’re no longer important. What’s important is the task that you’re performing or the personal state of the things you’re working on, things you’re thinking about, and those all exist somewhere in the cloud. As far as you’re concerned, that’s the place they need to be.
You’ve seen Amazon first-hand in Seattle, you know CMU, and Pittsburgh is in the top 20 for Amazon’s HQ2. Would Amazon’s culture and Pittsburgh’s culture work together?
Rashid: It could work pretty well. There’s a lot of Amazon that has a blue-collar attitude. “We work hard, we make things happen.” A lot of Amazon is about manufacturing processes and robotics. If they wanted to be in a culture where robotics is really built into the DNA of the intellectual community, where AI is very important, it would probably be a decent match.
If you think about where Amazon could have the most positive impact on the community, this would be a really good place for that. If they go to Washington, D.C., or Boston or New York, no one would notice. Yes, they would notice on day one, but after the news went away it wouldn’t have the same impact.