John Bares, a Carnegie Mellon graduate and former professor, briefly led Pittsburgh’s autonomous vehicle revolution before returning to lead Carnegie Robotics, the defense tech contractor he founded in 2010.
His ties to the city’s history and his vision for its future make him well-suited for a discussion on AI, military innovation, competition with China (and Japan) and more.
This is the first in a new Pittsburgh Post-Gazette Q&A series we’re calling "In Conversation.” The following has been edited lightly for clarity and length. You can watch a video of the entire interview here.
In Conversation with John Bares
Post-Gazette: I'm here today with John Bares, known around here as JB, founder and CEO of Carnegie Robotics. John, good to be with you. You guys are trying to get AI on the battlefield with soldiers who, as I understand it, have a small computer that would do some of the same computing that really high-level AI computers do, but that would enable the soldiers to interface with things like satellites, understand their situation on the field. How did we get to that point?
John Bares: So it's interesting. When I think about our history, in 2011 we made our first — I guess, what you would now call an AI-powered product — which was an agricultural machine for sorting strawberry plants. And this was technology licensed from Carnegie Mellon University. And basically what it did was look at these strawberry plants that are about the size of your hand — we looked at about 100,000 per hour on a very high speed conveyor belt — and determined the quality of that plant and then downstream, sorted it with AI.
So although the rest of the world maybe believes that it's only been here a couple of years, AI has been around for a long time. Back in the day it was called machine learning, and it was neural nets that would sort out what's a good plant, what's a bad plant.
Read More
And to bring us to the present, we’re of course applying AI in many more places. One of our newest products is called CardShark, where, as you described, it's a one-pound computer, basically, that the soldier can wear and can help that person basically do what we do on our cellphones. When you do something on your cellphone — a translation or some kind of a map search — it's actually doing most of the processing up in the cloud. A soldier may not have the cloud, may not have any WiFi, Bluetooth, cell, none of that. So we allow some of that to happen right there on their person.
What's exciting about that is you can start to do things like translation, live on the battlefield — translating signs, translating conversations, trying to understand a better situational awareness.
PG: Wow. So, for diplomacy, not just fighting.
JB: Absolutely. And you know, I think a lot of people have said over the years that a lot of lives could have been saved quicker, more in the moment, by translation.
PG: A better understanding of each other, got it. Can you tell us what we're in front of right now? This has been an exciting project.
JB: Our company does a lot of work in the commercial space and in defense. And so we just talked about a device worn by a soldier; behind me are some large trucks called PLS trucks. They're army trucks, and this is the main truck used for Army transport.
These are 50-ton trucks that run in convoys. Convoys are viewed as one of the very obvious, perhaps, and sensible applications, of automation and robotics. So you can have five, ten trucks driving with very few humans compared to what we have today, which is humans in every truck, and you have to protect those humans. You have to sustain them. Once they get there, you have to feed them, take care of them. And convoys are generally sort of sitting ducks for combat theater. So we're working on a program to automate convoying.
It's a very challenging program, competitive. Last year we were in phase one, that was three teams. Now we're in phase two. It's two teams. So we're going hard.
PG: And you're moving fast. Part of what's interesting about you and the leadership team here working on this project is that you were once leading the autonomous car effort here in Pittsburgh. And we've seen that go through ebbs and flows and arrive at this moment: It's an interesting week, the CEO of Uber is saying that we're still 10 years out from cars taking over from human drivers, but we're also in this moment where robotaxis are actually expanding fleets to other cities. What do you make of the moment now, compared to when you were at the forefront of it a decade ago?
JB: So that's interesting. We talked about convoying here, it's a great application for that technology, because it's a much more limited application area: closed-off roads, paths and things like that.
Robotaxis — they're here, right? You can go and ride them in multiple cities, and I think it makes sense. A lot of what's happened in the last decade, maybe not quite a decade, but six, seven years, you've seen a lot of improvements in the technology.
But also, I think regulatory and society have become ready for this. We see that because of the accidents and the deaths caused by human drivers, mistakes. It's only increasing with both fatalities and injuries in vehicles, but also to cyclists and to pedestrians. So I think automated taxis are a great starting point, and I think we can improve that.
Of course, they're going to have accidents — there have been accidents — but when you compare it to the rate of human accidents, I think it's really going to be an exciting new trend.
PG: I had one last question for you: You started your career in Japan, if I'm not mistaken. There has been some interesting geopolitics with Japan and with China in recent years, just recent months, even here in Pittsburgh, trying to figure out who we want to collaborate with in the AI race, and who we feel threatened by — if we see a model that comes out of China that undercuts some of what we're trying to build here.
Do you think that it behooves us as a nation, as the United States, to work with our allies on a variety of technology problems, and do you have concerns about the technical chasm?
JB: I think first that the relationship and issues between Japan and China are quite different, right? Today, Japan is in the news locally just because of Nippon Steel, and what do we do with that purchase [of U.S. Steel]? I think Japan's a very close partner. And in general for technology, I think the more people involved, the better. More diversity of thought, the ability to move faster and find new, exciting applications.
Also, it is no secret on AI and some of the new technologies, we also have to be careful and put up the correct guardrails, and I think again, more is better there, for helping us all figure out how to use this tech.
PG: There is certainly this march that you've seen from here in Pittsburgh and beyond toward more powerful AI, in more machines, doing more for our daily lives. And that doesn't seem to be slowing down anytime soon.
JB: No, and I think that that, you know, there's certainly challenges to that, but I think there's also opportunity for where we can be safer: We can remove soldiers from harm's way, we can reduce deaths of pedestrians or motorists on our highway. So you know there's, there's going to be a balance.
Want to see a conversation with a Pittsburgh business leader? Send suggestions to business@post-gazette.com.






Comments
Welcome to Post-Gazette subscriber-only commenting community. Please review our community guidelines.