Trending: Jeff Bezos’ phone reportedly hacked in 2018 by Saudi crown prince via WhatsApp
Andrew Moore, dean of the Carnegie Mellon University School of Computer Science, in his office this week. (GeekWire Photo / Todd Bishop)

PITTSBURGH — Last fall, after we solicited proposals to establish our own temporary “GeekWire HQ2” somewhere in North America, we received submissions from cities around the country to set up shop in their midst and cover their tech communities for a month. But only one came from an educational institution: Carnegie Mellon University’s School of Computer Science.

Yes, the CMU proposal acknowledged, a “parade of journalists” had already covered Pittsburgh’s resurgence. Politico, 60 Minutes, CNN’s Parts Unknown, NPR, The Economist, The New York Times even director Werner Herzog “have pulled the ripcord, touching down in Pittsburgh to get a glimpse of a receding industrial history and a burgeoning tech future, not to mention a suddenly trendy food scene.”

However, the submission added, Pittsburgh “remains largely virgin territory for in-depth reporting; people who come for a day, or three, simply miss a lot of stories. They end up producing the same stories as the crew that preceded them the week before. And that’s why Pittsburgh is the ideal city for GeekWire HQ2.”

We agreed, and after the past month, we agree even more. Over the past 34 days, six GeekWire reporters and editors have done extended stints at GeekWire HQ2 in Pittsburgh, fanning out from our base in the Lawrenceville neighborhood to cover not just the technology resurgence but also the larger societal and economic challenges facing the city, with some valuable contributions from reporters at GeekWire HQ1 in Seattle, as well.

We’ve talked with city, regional and academic leaders, community activists, entrepreneurs, artists, religious leaders and people on the street about everything from the future of technology to the possibility of Pittsburgh landing Amazon’s HQ2, the high-profile project that helped to inspire our adventure.

Pittsburgh Postscript: Surprises, gripes, memorable characters and epic experiences from GeekWire’s month in the Steel City

With that backdrop, we could think of no better person to conclude with than Andrew Moore, the longtime Carnegie Mellon machine learning and robotics professor who established the Google engineering office in Pittsburgh before returning in 2014 to lead CMU’s School of Computer Science as dean. Our questions for Moore were informed by our month of reporting, including Pittsburgh’s strengths and shortcomings, its role in shaping the future and its efforts to adapt to the new economy.

Continue reading for edited highlights from our conversation.

GeekWire: Thank you for having us in Pittsburgh. It has been a great experience, and CMU’s School of Computer Science made it happen through your proposal.

Andrew Moore: We’re really glad that it did happen, and we are also glad that it turned out great.

GW: One of our observations as a group is that the academics and the research here are clearly world-class, if you compare it to other places we’ve covered — there is world-changing stuff happening here. We were struck that we didn’t see that translate into the tech ecosystem to the degree even that we’ve seen in Seattle. First, do you agree with that observation? And second, what can CMU do to change that?

Moore: I agree with it, but very reluctantly. The reason I’m reluctant is I think we are finally getting better at it.

Back around 2008, 10 years ago or so ago, I think we were very sleepy like that. We as Pittsburgh, or certainly the academic folks, would finish up papers or deliver stuff to sponsors, but they wouldn’t necessarily take it out and try to make it massive. Now we’ve hit a different place. Our stuff is getting out there. I’ll give you an example. FacioMetrics, one of the emotion recognition companies here, was a nice startup. And it was acquired very early, and for a very small amount of money. That’s what frustrates me. That’s still better than 2008 when it would have been in a paper, and then four years later someone would’ve picked up the paper in some other part of the world and made a company out of it.

But we’re still not at the point where someone here says, “All right, I’m going to do this, and this is going to be the next Google.”

GW: With your reluctant agreement, how do you then change that?

Moore: One of the things that I’m trying very hard to do is make that step into running a startup much less fearsome. And the fearsome aspect of it is mostly a lack of experience by those who have already done it, who can sort of help you. So we are making it pretty easy for faculty to disappear for three years to do a startup, which is the kind of length of commitment you need, rather than having to have them first ask the question, “Shall I do this for a short amount of time and then quit CMU if it goes well?” This gives them much more comfort there. So we have a few of those things going on.

Capital, I think, is relatively available now. So I’m very happy about that. But it’s that buddy who you talk to every week or two, who’s been through this whole thing and has seen everything, which we’re still lacking here. So we’re importing a bunch of our alumni who have done that, but are now living elsewhere, to come here every week and hang out with faculty and students to talk about their experiences.

We are doing everything we can to get hold of product managers and biz dev people, including growing our own here, so that we have those kinds of folks in the ecosystem as well. But I think that’s what’s really needed. And that’s the biggest difference between here and Seattle, where there’s plenty of product people and second and tertiary serial entrepreneurs, who are important in their own right, but really important because it means you’ve got friend who’s done it.

GW: It feels like there’s a cycle, where if you do then create that ecosystem, it comes back. In Seattle, we’ve seen this happen. People cite, of course, Microsoft and Amazon, but there’s this middle tier of company where it’s starting to bubble up. Zillow is probably the best example, because they’ve bubbled up, and now they’re starting to give back to the University of Washington Computer Science School. When you look at the competition with the other schools in the top tier of computer science around the country, are you worried that this dynamic, if not fixed, could lead other schools to overtake you, like the University of Washington? Is that a concern?

Moore: That’s not the thing that keeps me awake at night. The thing that’s most important is that we relentlessly deliver the world’s best technology. And that, I feel, we’re safely ahead of. If we just stay where we were in 2008, we are kind of dead in the water. And when I arrived here as dean, I had three sort of non-negotiable principles that I needed to agree with before I’d say yes to the job. One of them was that being responsible to the School of Computer Science actually means doing a lot of work to help make sure the rest of Pittsburgh grows, because you cannot have a great academic institute without it being surrounded by technology.

Carnegie Mellon’s Gates Center for Computer Science was made possible by a lead gift of $20 million from the Bill & Melinda Gates Foundation. (GeekWire Photo / Todd Bishop)

But startups which have grown to billion-dollar valuations are a really important one, but they’re not the only thing. So we do have many places, many of the big blue-chip companies have put down roots here. Those are turning out to help the really tiny startups who’ve then got that local relationship.

And we have a sort of investment from folks who want to help real robotics companies grow, especially those in self-driving. And so that makes me feel a little bit safer.
But having said that, what I’m waiting on, and really anxious about, is I do want to see a few billion-dollar valuation companies coming in.

GW: One of the major triumphs of your computer science department over the past couple years is the gender equity that you’re seeing in your incoming classes. It’s remarkable. And it’s not something that’s happening elsewhere. How have you done it, and what can the rest of not only academia, but technology, learn from what you’ve done here?

Moore: The first thing is you’ve got to begin two decades ago. That’s a really hard thing for folks who want to see that recipe. But of course, I believe that folks can do it more quickly. The first thing not to do is try to make the curriculum “pink” in any way, if you see what I mean. And the faculty and the leaders around this were very clear on that. It is very important to have those connections with middle schools and high schools, where all we have to do there is say what we really believe, which is that computer science is problem solving. Computer science is not semicolons and curly braces. And I think that really helps.

But then, here was the inspiring piece. After a while it seems to become a self-reinforcing flywheel. It’s self-reinforcing in the sense that once we were 50 percent more women than the national average, we were a very attractive and interesting place for women to come to. And once that took us up to 75 percent, that was even more so. And now I think we’ve hit the optimal situation here, where it is a really good environment, where no one feels, at least gender-wise, no one feels like they’re the only one in the class or in a small minority within the class. And that was really good news.

Uber test vehicles at the Advanced Robotics Group in Pittsburgh. (GeekWire Photo / Taylor Soper)

GW: There was the high profile case of Uber bringing a lot of CMU folks aboard. How do you coexist peacefully and constructively with all of these tech giants that are coming into Pittsburgh?

Moore: Yes. So we’ve had, every year, many universities’ faculty disappear to go off to industry or to go to other universities. I’m hyper-sensitive about this, because like a lot of top universities, I’m worried that suddenly, a large fraction of my faculty might all disappear at once. And for a while, we were concerned that Uber was portending that sort of thing, because we lost four faculty in a single transaction there (in addition to about 35 technical staff).

The thing that we’ve done, which seems to be working really well, is we’re just relentless about hiring great new junior faculty, and we spend pretty much all of our time in February, March, and April on interviewing huge numbers of candidates, finding the ones who have that combination of crazy thinking and skills that are actually getting stuff done so that they produce results to bring in. And over the last three years, we’ve brought in approximately 45 new faculty.

And I think the reality is we bring in plenty of great faculty every year. We give them visibility, which makes them prime targets for these companies. But the fact that our faculty get so much opportunity later on in their careers is what’s going to make them decide to come to CMU instead of some of our competitor schools, cough, University of Washington, un-cough.

GW: Coming here as an outsider, especially from a town that was built on software, it is remarkable to see the robotics. Coming here, it’s amazing to see the number of robots and then the applications out into the real world, obviously self-driving cars and automation being the biggest example. Do you see CMU and Pittsburgh continuing to bet largely on robotics and the future of robotics and AI in the future? Or will there be something else, something that we don’t even know about that comes along and becomes the thing that defines Pittsburgh and CMU in the decades ahead?

Moore: Both. I know that’s a tricky answer, but the time is really here for physical robots, as opposed to soft-bots, or anything like that. What I like now is that we’re seeing diversification as not just self-driving cars. We have autonomous farming, we have robots designed by Bossa Nova to use at Walmart, which are designed to be completely safe wandering around among humans doing significant amounts of work. And of course, we have all kinds of drone mapping companies. So that, I really like.

I expect that there will be something which we haven’t quite thought of yet. I’ll give you one example of what it might be. The notion of computers really understanding us as emotional beings, who express emotions and can easily be stressed out or made comfortable, I wouldn’t be surprised if in the middle of the next decade that’s actually considered a more important technology than speech understanding or language understanding, which is human understanding. We’re well-positioned to win there.

GW: Emotional intelligence.

Moore: Actually, more than that. Emotional intelligence is where either I as a human or I as a robot don’t irritate people, in a way, or make them feel uncomfortable in the way I discuss them. But then there’s this deeper kind of relationship, where perhaps you talk every night to whatever the emotional intelligence startup comes out of Pittsburgh. And it asks you the question, “Was today a good day?” You talk to it a little bit. And over time, with information from many other folks, and reading your body language, knowing your health statement, it begins to really help you to live the life that you want to live.

So that might be it, but it’s probably going to be something else. I think that when I look at what a few of our faculty are doing around self-improving software, software which rewrites itself and detects its own bugs, maybe that’s going to be the big thing. There will be something else there, as well as the robotics, which is the thing we can easily see.

GW: How are you trying to make sure the next generation of software developers thinks about the ethical issues of robotics and AI? We’ve covered this: jobs, workforce, privacy and all of the issues that surround AI. Are you thinking about those issues as well?

Moore: Yes. So students and the faculty are both demanding that we do classes about this. And I’m happy to say that even in the last two years, we’ve introduced seven classes in this area. There’s one class called “Computers, Algorithms, and Justice” which talks about notions of fairness, and how you can incorporate that into software.
There’s a course about bias removal, or bias detection, in machine-learning algorithms and what you can do about it.

The thing we are uniquely interested in all these classes is, part of the class is, “Look, there’s a problem. Here’s the evidence that there’s a problem.” But then part of the class is always, “And here are the beginnings of solutions to the problems.” So as opposed to a class which, I refer to these as whiny classes, which just say, “There are all these problems. Isn’t that bad?” We don’t educate them just about the problems. In most of them, we do make sure the students come in and actually start to help thinking of solutions.

GW: As we talk about the future lot of us have young kids on the team, ranging from 1 year old to 7 years, and older. What should our kids be thinking about? Actually, and more specifically, how should we as parents be thinking about the future as it relates to computer science and engineering, robotics, and getting them not only into things they love, but into things that will change the world?

Moore: Under about 12, our consensus recommendation would be they’ve got to solve problems together. The fundamental part of computer science doesn’t actually involve a computer. It involves groups of two or three or four kids who, together, solve a problem. One of the best introductions to computer science is getting some dry ice, letting your kids play with it while making sure they don’t kill themselves, and give them a challenge as to something interesting to do with it. And it’s that ability to work together on a problem is what we want to see kids prepared for, and then they can start programming when they’re 11 or 14 or 17 or 19.

GW: We’ve written more than 80 stories over the past month. We’ve had six reporters and editors here. And I think our biggest struggle is there is more to cover here than we could possibly get to in a month. I can’t imagine how a reporter would parachute in here for a day and figure this place out, because it is sophisticated and complex, and there is a lot going on. We’ve missed a lot, I know, even in a month. And I know you’ve been following the coverage. What have we missed?

Moore: Although this is not the right time of year, I wish you guys had gone hiking in the forests and streams and rivers around Pittsburgh. It’s actually kind of magical. Pennsylvania refers to a place covered in trees, and me coming from England, that was one of the things which amazes me about this place. You can go almost get yourself lost in forests, and you suddenly discover some old ghost town area sort of with ivy growing on it, deep in the forests that probably everyone’s forgotten.
And so many huge boulders of granite, and fast-flowing streams, and so forth. Some people get here, and it takes them a year or two to realize that they’re actually in a very special part of the world, geographically. …

I was really pleased that you didn’t just cover the cool bits of technology, but you talked about education and you talked about the social changes in Pittsburgh, so I really liked that.

GW: You helped inform one of those stories, on the Kelly Strayhorn Theater. You observed that there is a divide between Pittsburgh and the tech community at times, and the theater is emblematic of that. How do you change that?

Moore: That is one that I’m less confident about. It may take a while for kids to grow up and just making other connections as that happens, but I would say that is a weakness that I’m concerned about. If there’s stuff we can learn from Seattle about this, we should, but it would be a shame for us to create two worlds in Pittsburgh, which kind of like each other but are almost not communicating.

GW: I want to thank you for submitting the proposal to bring us here. It has been a remarkable experience. I don’t think we could have picked a better city than Pittsburgh. Thank you for having us.

Moore: Fantastic. Thank you so much.


See all of our GeekWire HQ2 coverage from Pittsburgh.

Thanks to everyone who made GeekWire’s HQ2 possible, including: K&L Gates and DQE Communications for sponsoring the project; Carnegie Mellon University’s School of Computer Science (including Mark Power, Byron Spice and Andrew Moore) for submitting the proposal that brought us to Pittsburgh; Butler Street Lofts and Beauty Shoppe coworking for hosting us in Pittsburgh’s Lawrenceville neighborhood; and the organizations and people who organized and joined us for events and gatherings throughout the month, including CMU, the Allegheny Conference, the University of Pittsburgh, Duolingo, Birchmere Ventures, Pittsburgh Mayor Bill Peduto, and many others. Most of all, thanks to the people of Pittsburgh for putting up with a bunch of meddling reporters. Thanks for welcoming us to tahn, yinzers.

Like what you're reading? Subscribe to GeekWire's free newsletters to catch every headline


Job Listings on GeekWork

Find more jobs on GeekWork. Employers, post a job here.