Uber is not the only self-driving car game in town. Ford announced in February that it would invest $1 billion over five years in Pittsburgh-based Argo AI. The company has been quietly building its team, with the goal of putting self-driving cars in production by 2021.
CEO Bryan Salesky spent part of his childhood in southwestern Pennsylvania, and graduated from Pitt in 2002.
90.5 WESA’s Liz Reid sat down with Salesky to talk about his background, the shortage of computer scientists in Pittsburgh and his vision for the future of transportation.
Their conversation has been edited for length and clarity.
LIZ REID: When did you first become interested in self-driving cars?
BRYAN SALESKY: Well, I love self-driving cars and I think it's because there's something really gratifying when you're writing software and you know that it's able to actually, like, control a machine somehow. You know, I was never the type of guy that wanted to kind of write software and have it be a website or an app on the phone. I was always the type of guy that was interested in seeing all the hard work you put into a computer program and seeing it actually get executed on a robot and have it do something really meaningful in the world. It's very gratifying for me.
REID: And you attended the University of Pittsburgh, right?
SALESKY: Yeah. So, I got into the robotics world through a friend of mine. I actually worked at Union Switch & Signal here in Pittsburgh for a couple of years, and it was great sort of training grounds for how to learn how to write kind of production quality software. And a friend of mine said, "Hey, you know, I know you're kind of like me, you always wanted to get into this robotics thing. There’s this really cool place called the National Robotics Engineering Center in Lawrenceville and they're doing some pretty big things and I think they can make use of people like us." And I was lucky enough to get hired in there and it's been awesome ever since.
REID: And so were you working on technology sort of tangential to the self-driving car there?
SALESKY: Well, yeah. So I started there in, gee, I think it was 2004. And you know this is way before self-driving cars was cool. It was very much research in nature, trying to see you know how can we apply artificial intelligence and computer science and all these kind of new things that were coming out, how can we apply that to a military context and be able to take soldiers out of harm's way and put robots in instead? And it's sort of the seeds of those programs that developed a lot of the fundamental techniques and algorithms that we use today that is able to make a car drive itself.
REID: How many folks do you guys have on staff right now and are you still hiring to fill out your Pittsburgh team?
SALESKY: Yes, we just crossed into triple digits, which is cool. And we're hoping to have 200 by the end of the year.
REID: So, are you hiring?
SALESKY: We are very much hiring. We need more people with the backgrounds required to work in this field. And that means electrical engineers, mechanical engineers, software programmers, you know all different disciplines. It's what it takes to build something this complex.
REID: We've heard some people in the self-driving car industry say that their presence is going to bring jobs to Pittsburgh for Pittsburghers. Are there enough Pittsburghers with this skill set or are you having to draw from all over the country?
SALESKY: We're drawing from all over the world. It's really a skills shortage and you know we talk to universities in the area and say, "Hey how can we help, how can we support you because we need to get bigger pipeline of students that know how to do this type of work."
REID: Would it make sense for even high school kids to start getting an education in this?
SALESKY: Yeah, robotics is a great way to engage high school students because, kind of how I got interested in it, there's just something incredibly gratifying when you write software, which is a very virtual like non-human thing, right? You're at a computer terminal, you're typing text into a screen. And then to be able to take that and somehow you know learn all the ways in which that can be embedded into a computer that then moves something and does something useful in the world. That is just it's amazing and I love it. I've worked with high school kids and I love when I see how they just light up when they see this this program that they created actually you know come alive.
REID: So, you're working on developing what's called a Level 4 self-driving car. What does that mean?
SALESKY: I know, it sounds very technical. So NHTSA and also the Society of Automotive Engineers they have sort of different guides as far as.
REID: What's NHTSA?
SALESKY: The National Highway Traffic Safety Administration governs all of the motor vehicle safety in the United States. So those two organizations have published guidelines that are pretty similar in terms of what the different levels of automation are on a on a vehicle. And so it starts with no automation at all and goes all the way up through various driver assistance methods. So you still have to pay attention, in many cases you still have to be driving the car but maybe it will adjust the speed for you, and that's called adaptive cruise control. That's more of a Level 2 system. Level 3 is the vehicle can drive itself and it needs to be able to alert you when it's time to take over if it can't handle a certain situation. And that has a number of issues with, well if you don't have to pay attention but then the car has to tell you to take over ... drivers are not very good monitors and they're also not very good at switching from like reading a book to "Oh now I have to reengage and start driving again." Humans are not good at that switching. So we think Level 3 is tough. And then Level 4 is completely autonomous, so no human driver required in the loop, the computer safely navigates the vehicle through the world.
REID: You and others have talked about how autonomous vehicles will be safer than human operated vehicles. But I'm wondering how self-driving technology addresses one of the biggest challenges related to car culture which is pollution and the burning of fossil fuels. Do autonomous vehicles have a place in a greener cleaner transportation future?
SALESKY: Well, absolutely. I think as part of how self-driving cars can transform a city, it can reduce congestion by reducing the amount of personally owned vehicles in a city. And so what you end up with is a shared network of vehicles that are being run all the time, and that can also be sort of scheduled to recharge whenever necessary. So now, a lot of the reasons why electric vehicles are not necessarily adopted is because of a concept called range anxiety, where people are thinking, "Well gee, how do I know I can get to where I want to go and still be on a full charge and I need to take several trips?" And so on.
REID: So, are you saying the vehicle might go charge itself between passengers?
SALESKY: You got it. The vehicle will go charge itself and there will always be a vehicle available when you want to go somewhere. That's the idea, right? And so now you don't need to have this expensive thing that consumes gas and sits in your garage 90 percent of the time. This is this is how self-driving vehicles, electric vehicles, these are the seeds of a much bigger change that's happening that will happen in the future.