He led WIRED Magazine from 2001 to 2012, making the internet cool and mainstream and even coined the term “The Long Tail.” Now Chris Anderson is the Founder & CEO of 3D Robotics, DIY Drones and DIY Robocars, an autonomous racing series anyone can afford to enter. So how do you build an autonomous vehicle in your garage? And how did amateur drone enthusiasts revolutionize the industry forever? Anderson details the genius behind DIY robots and explains how they can help solve problems encountered by those on the forefront of autonomous vehicle development.

Listen On
Apple
Google
Spotify
iHeart Radio
Share

View Episode Transcript

Episode Transcript

Alex Roy
Hello and welcome, again, to the No Parking Podcast. I’m Alex Roy here with my friend and cohost, Bryan Salesky.

Bryan Salesky
Hey, Alex.

Alex Roy
This week we have, really, they’re all interesting but this guy, was the editor-in-chief of Wired Magazine, who really pushed it over the line into when it became like a big deal, Chris Anderson, who was the editor chief, I mean, for, with the early 00’s and a lot of big changes happened in the technology sector when he was in charge. Today he is the founder and operates DIY Robocars, which is a competitive race of small format… How would you describe it?

Bryan Salesky
Well, it’s kind of hobby robots, but also far more capable than say Lego Mindstorm. I mean these things have embedded electronics where you can do quite a bit with them. It’s really cool to see.

Alex Roy
And he has a quality that I know. Bryan has, and I like to say that I have, just, he can sniff out BS, and call it when he sees it.

Bryan Salesky
He must have learned that at Wired.

Alex Roy
He must have.

Bryan Salesky
What did you think of the toys he brought by?

Alex Roy
I wish I had spent more time studying and building robots and things as a kid.

Bryan Salesky
It’s neat. We could all have a lot of fun with this things. Doesn’t he race them? He has a little league that he gets together in SF?

Alex Roy
Let’s find out. Let’s roll right into it.

Alex Roy
See, you mean to tell me, Chris, that you had no idea that Bryan Salesky was going to be here?

Chris Anderson
I failed to read my emails correctly, and so I had no idea that Bryan Salesky was going to be here. But a guy who introduced himself as Bryan, who I assumed was the audio guy, which is to make sure that the mikes worked, and now, the penny drops.

Alex Roy
Well, Bryan, this is what happens when you spend all your time actually building things and not trying to sell them.

Bryan Salesky
I find it entertaining.

Alex Roy
That’s really great. Let’s get right into this, Chris. What is the best form of AI for a self-driving vehicle?

Chris Anderson
That’s a trap.

Alex Roy
It is.

Chris Anderson
Well, we haven’t found it yet, that’s for sure. I mean, we have a number of problems with your phrasing. Basically, every word you used there is the problem. AI, I don’t know what that means. Self-driving vehicle, don’t know what it means, best, don’t know what it means. I think what, I think I know what “what” means.

Alex Roy
That was about it. Complete fail. As someone who prides himself on precision of language I’m thrilled that you pointed out that was just pure bait with no value.

Bryan Salesky
I was watching Bloomberg TV last night just for five minutes in my hotel room, and Woz was on there. And the hostess asked him, “What do you think about big data?” And it was just, and poor Woz was like, “Could you say that again?”

Chris Anderson
I know.

Bryan Salesky
And the person said, “What do you think about big data?” And it was just, where do we even go with that.

Chris Anderson
So, we’re amongst friends. I don’t use the letters A and I because I think it’s just marketing, and then and then you add…

Bryan Salesky
Why do you think we put that in our company name. It’s marketing. Absolutely.

Chris Anderson
Okay. Fair point okay. That was my fail number two.

Bryan Salesky
This is fantastic, I’m high-fiving you right now.

Alex Roy
Looks like I can leave for the first time. I could just walk out of the room. Take it away.

Chris Anderson
Is it because they didn’t have an ML domain?

Bryan Salesky
Exactly. I think you got it.

Bryan Salesky
We did not put the word deep in our company name.

Chris Anderson
Thank you.

Bryan Salesky
We…

Chris Anderson
And also CNN is kind of ambiguous, isn’t it?

Bryan Salesky
Yeah it very much is.

Bryan Salesky
Do you know what CNN is?

Alex Roy
Some complex new network.

Chris Anderson
Let him dangle.

Alex Roy
Let me go.

Chris Anderson
Just twist in the wind.

Alex Roy
Convolutional neural network.

Chris Anderson
Yeah. There you go.

Bryan Salesky
Well done.

Chris Anderson
Technical interview. Now you have to get a whiteboard and do a link list.

Bryan Salesky
Now you need to write quick sort on a whiteboard, and you could become a software engineer.

Alex Roy
Look…the opening question was intended to bait you into what is happening at the next DIY Robocars event. You told me flat out, you’re going to have three forms of learning compete.

Chris Anderson
Correct. Should we dive in to this?

Alex Roy
I’d like to dive in, here’s why. I’m on Twitter too much. But on Twitter, there are people who do not know what the words they use mean. Debating the merits of different methods of teaching machines to perform tasks. And so, the most common example of this are people saying, “Oh my God, end to end, neural networks and big data are why Tesla’s in the lead.” And when I go out to dinner people say, “Oh, you’re with Argo, but of course you know, Tesla’s in the lead.” And it’s always, “Neural networks!!, yada, yada…” So, I want you guys to explain to our friends in the audience what these actual different methods of learning mean? How do they work?

Chris Anderson
Totally. I’m coming from the idiot-doer side, and Brian’s coming from the genius, well, we do it at mass scale with randos, and you do it probably at less mass scale with smart people. So, our approach would be, think of it as like it’s like in the early days of science where you had citizen scientists who are Charles Darwin, biologists, astronomers who were just gathering data. And then you had like proper scientists who were academics who did proper controlled trials. I think that’s a little similar. We’re doing kind of a citizen self-driving car thing here where we are, we recognize there is no right answer to the best. And we also recognize that a good way to tell which is better is to have them compete.

Chris Anderson
And so, we’ve constructed a way by which people, regular people, brilliant and otherwise can try different algorithms wheel to wheel in a kind of classic racing environment, and see which one performs better in that scenario. So, what we have is basically every category you can imagine. We have different kinds of deep learning. We have behavioral cloning where you sort of drive like a human, it learns, it watches your inputs and outputs and tries to abstract some way to derive from that.

Alex Roy
What could go wrong?

Chris Anderson
Exactly. Well, what actually goes wrong is in that particular case, is that we can get to follow your behavior on that track. So, you on that track. But that means that it would work on a different track. Could you be exposed to situations that you had not already shown it, and it would work et cetera? So, that’s behavioral cloning.

Chris Anderson
Then you have reinforcement learning, where you basically give it some criteria, what’s good performance, what’s bad performance. And it tries to optimize around that. That’s what Amazon’s doing with their DeepRacer. And then you have simple supervision where you basically take a lot of pictures, and sort of say if I was a driver, I’d be going that way, and that way. And this is what your destination should be. So, like when you do a captcha online, it says click on the fire hydrants et cetera. That’s a little bit of supervision. And we do the same thing. Just train it with some supervised perspectives on where it should go and then learn from that.

Chris Anderson
Those are the three deep learning approaches. Then we have traditional computer vision control theory ones, which are based on standard computer vision, or lidar and things like that. And those are, they don’t learn, they don’t get better over time, but they’re extremely fast. So, we have one thing called cone slam, we put cones around the track, which is really the only structure we have. We’re indoors on a carpet. It’s not like we have buildings, and other things, features, and so, we put cones on there, It then sort of figures out where it is on the track based on the cones it sees. And then gets a sense of the whole track. And now it can look ahead, and now it can say, “Hey, I not only know that there’s a turn coming, but there’s another turn coming after that.”

Chris Anderson
And then it has kind of like god like powers of the track, which is another kind of advantage. These and a couple other techniques. But those are the ones that are competing against each other. All have pros and cons. And the end day of the day, in a track environment, the fastest time wins. That doesn’t necessarily means the best solution for the real world.

Bryan Salesky
Out of all of that in all the experiments you’ve seen…what surprised you the most? Which techniques?

Chris Anderson
One I’m most surprised right now is visual slam which is basically visual dormitory with a fisheye lens which just has one camera, looks up with 180 degree fisheye. It has kind of, sees all around it. And initially we were looking at the cones, and it’s like, that’s cool. But we don’t always have cones. And then they realize that indoors we have the lights above, and the lights are a distinct feature. It’s like almost a full ceiling fiducial.

Chris Anderson
I love it.

Chris Anderson
And now it localizes off the ceiling.

Bryan Salesky
That’s so cool, isn’t it? That’s really neat.

Chris Anderson
So, fisheye lenses, it’s getting totally non-intuitive. Everyone’s been thinking, oh well, we should just have cars drive like humans. It’s like eyes looking forward, et cetera. But a fisheye lens is looking up. It’s like an inverse god perspective. It’s completely impossible for a human to interpret it. But of course, this software has no problem to translate that distorted fisheye lens into the equivalent of a top down perspective.

Bryan Salesky
There’s actually algorithms that will attempt to un-distort it, and de-warp it and turn it into something that’s actually quite marvelous.

Chris Anderson
Turns out it’s just built into openCV. Anyway, 10 years ago this would’ve been impossible, 5 years ago it would have been a PhD project. And now it’s just a couple of lines of openCV on a Raspberry Pi.

Bryan Salesky
This is what’s incredible is that I think I’ve mentioned this in the past, but if I was a kid growing up today with everything that’s at your disposal from Arduino, Raspberry Pi, openCV, stuff that Nvidia is doing. You can put these things together now into platforms like what you brought for us to see today. And if we go to your website, we can see it. It’s absolutely, absolutely awesome and incredible what you can do now.

Chris Anderson
We did this 10 years ago with drones. Prior to 2007, drones were military industrial, sensors, gyros, accelerometers were big mechanical things and it was really impossible. And then there was a glitch in the Matrix called the iPhone in 2007, and it had in it a couple of enabling technologies, MEMS sensors, sensors were chips, not mechanical. GPS, fast wireless, ARM core processors. All this kind of stuff. And those took these really tough problems and turned them into chips, and software that you could buy off the shelf.

Chris Anderson
And that led to the modern drone era, and drones went from a million dollars, to 1000 dollars, to 100 dollars. But that was really solving a simpler problem, which is where is down, and where am I in the world which, was G.P.S. Indoors, with cars, you can’t count on GPS.

Bryan Salesky
Even outdoors you can’t count on GPS.

Chris Anderson
Even outdoors you can’t count on GPS, and also you’re dealing with computer vision, which was harder than the simple inertial problems we were dealing with. So, it took another 10 years for all the things that Bryan mentioned. And what we’re looking at right here is a $200 car. It’s Nvidia Jetson Nano which is CUDA cores, that’s a GPU format. It’s got a kind of a Raspberry Pi style camera, and lots of amazing software including things like TensorFlow, and beyond. A year ago this was impossible. The Jetsons Nano…

Alex Roy
For those who can’t see it, what are we looking at here.

Chris Anderson
It looks like an RC car. A seven inch wheelbase RC car, it kind of costs like a hundred bucks on Amazon. A computer on the top which is called Jetson Nano, a little camera on a stock in front of it, and some 3D printed parts. It’s called a jet racer, and it’s really Nvidia’s, Nvidia’s can always do two things at a time. One thing is always going to be computer game, video games, And the other was bitcoin mining for a while. And now it’s autonomy and the dread letters AI, et cetera. Now they’ve always done that. But this has become the second priority.

Alex Roy
I don’t know What AI means, but…

Bryan Salesky
You can have a lot of fun with these. These are also really super educational platforms. Do you have schools using these things?

Chris Anderson
It’s been difficult to integrate these into curriculum.

Bryan Salesky
Why is that?

Chris Anderson
Teachers are overwhelmed, and underpaid.

Bryan Salesky
Yeah that’s been my conclusion too. We sponsor various robotics teams in the cities where we have people in offices, and we’ve talked to a lot of teachers, and some of the school districts that they can afford to hire people who think about the next generation of curriculum. There’s other schools, but most school districts cannot, and they’re just overwhelmed, and I don’t think they realize how many resources are actually available and at their disposal now that can make this pretty low friction to bring something like this in the classroom.

Chris Anderson
We are having more success actually with the girl scouts. Girl scouts are already kind of organized outside the school process. So, we have engaged parents, and at least in the California girl scouts some robotics and technology is a big part of it. So we’re going to be doing an all day session with the Girl Scouts in spring, and having them do self-driving cars.

Bryan Salesky
What kind of exercise, or how do you get them started? Because even as symbols we’re making it sound hard here. Chris, you and I are experts at this. How hard is it to actually get started?

Chris Anderson
We’re starting something really simple. It’s also on the DIY Robocars website, and it’s called the Minimum Viable Racer. And it uses a single board camera computer called OpenMV. And it costs like 50 bucks, and it’s basically a camera that has enough processing onboard to run Micro Python. And then a kind of an optimized version of OpenCV, and it’s got a fantastic IDE, development environment. And what do is we provide them with the cars themselves. You can put them together for about 80 dollars. And we provide them with each team, they call team of three or four girls will have a car, will have a track which will have a red, green, and a blue Lane.

Chris Anderson
And all the cars come ready to drive on the, let’s say the outside one, which is going to be let’s say the blue line. So, they’re all trained to run on the outside, and then they race against each other.

Alex Roy
I just had a daughter. She’s under a year old, and I am obsessed with how I can give her an advantage in learning about robotics at the earliest possible age. So, what is the path to being ready to begin learning what you describe with DIY Robocars? Even the earliest stage. Am I buying DUPLO, am I buying a Lego Technic robot kit?

Chris Anderson
I should tell you, I am 100% a failure with this with my own kids. Not a single one is interested in robotics in the slightest. Possibly because I’m pushing a little too hard, but we’ve done everything. We’ve done like Mindstorms, I was on the advisory board. I mean, the drone started with like a Mindstorms. Our first drone was a Lego Mindstorms driven drone. Right now I find the ones the kids like the most, depends on the age of course. The ones they like the most are these little balls called Ozobots which you basically, they are the line following balls if you will.

Chris Anderson
And you program them with different colors. So, like a black line, and then if you have like a blue stripe, and then a yellow stripe, and then a red stripe that means, do U-turn, et cetera. So, they’re quite like the tactile elements. I don’t know whether they’re learning anything about robotics, but it’s fun to play with, and they understand a little bit of sort of procedures. But in terms of teaching robotics, I only really have one rule which is doing is better than just hearing. So, any kind of hands on work, seems to entice them.

Bryan Salesky
I think that is the key because there’s so much you need to know. The best way to do it is just dive in, and get to know the tools, get to know the language, get to know all the three letter acronyms…

Alex Roy
CNN.

Bryan Salesky
It can be daunting, but you’re right. You just got, it’s so much more interesting to do than to just read…

Chris Anderson
Also I think there are kids that actually learned a lot about robotics without doing robotics. Like for example, one of my sons is really into a video game called Arma. And Arma is, you basically get to program it. These are objects, soldiers tanks, whatever have properties, and you script the properties et cetera. He thinks of that as being a cool video game that he gets to design his own levels, but he’s learning some proper programming techniques. So, I think that over time, robotics becomes less about mechanical, and more about the software elements, and kind of any software, any engagement in programming is going to get them to robotics eventually.

Alex Roy
So, what was the birth of DIY Robocars? And what did the events look like in the beginning?

Chris Anderson
It was about two and half years ago I guess. I had done DIY Drones, and that got quite big, and created a movement, and then I ruined it by creating a company out of it. So, I industrialized my hobby. Always a big mistake. And then not only that, but I also really didn’t do any hands on work anymore because I was running a company, and so we largely spent five years solving really hard robotics problems with drones, and then solve really hard problems about utility, and ease of use, and all that kind of stuff. And I found myself abstracting more and more from hands on. I was looking for projects that were cool. And I think it must have been, what turned me onto cars?

Chris Anderson
I think it must have been maybe Raspberry Pi 3, when that one came out. It was the first one was really capable of doing proper computer vision. And I was like, okay, it is time. If you’re in the tech industry, every now, and then the signal goes out into the universe, something has changed, it’s going to be AR kit, or it’s going to be like a Raspberry Pi 3, or it’s going to be just a nano. And suddenly something you’ve heard about TensorFlow becomes something you can do. And when that signal comes out, can’t just abstractly learn about it, you need a project. And so, I’m like, okay, I think computer vision, I just like lane following, but actually thinking about roads, and tactics et cetera is now possible.

Chris Anderson
And once again, I did the same thing I always do, which is I learned just enough to ask stupid questions, and then other people answer my questions, and then this liberates them to ask their own stupid questions, and it turns out that a lot of people were thinking about the same thing too.

Bryan Salesky
And a community starts to form too so people can teach one another in a much more kind of friendly accessible way.

Chris Anderson
The one thing that’s been common, now I’m everything about everything that we do, is that you take something you’ve read about in the newspaper, but can’t do, and you let them do it. So, the letters DIY are super powerful because really we should be doing it together, not do it yourself. But DIY means that you can participate. And I learn by doing. I think lots of people learn by doing. And we’ve been reading about self-driving cars, but unless you happen to be lucky enough to live in one of those places where you can hop in the back, and experience it, even then you still can’t really do it. And that struck me as a gap.

Bryan Salesky
I think it’s cool. Also teaches you about teamwork, right? When you see kids work on these types of platforms, whether it be First Robotics, or they have different age groups. When you see them all working toward a common goal, and they’re splitting up the work, and they’re running project management even though they wouldn’t call it that, but that’s kind of what it is. They’re all working towards that common goal, and helping each other out with how to program something, or how to fix a mechanical problem with the hardware. They feed off of each other, and it’s fun to watch I think.

Chris Anderson
I did First Robotics judging and mentoring for a while, and I love the teamwork. I love the, they had cheerleaders. It’s quite a big thing.

Bryan Salesky
It really is. It’s huge.

Chris Anderson
What I was a little disappointed by is that they’re not really doing robotics. Almost all of the stuff is tele-operated, manually driven. None of the kids, it’s very hard to get someone to actually read the sensors.They really just want, it’s like go forward three seconds, and then take a left. That kind of stuff. Very scripted. And it just didn’t feel relevant to modern robotics.

Bryan Salesky
I mean, well, let’s face it, they’re learning an awful lot about the very basics of programming electrical mechanical systems, the basics of control system theory. Very basics, right?

Chris Anderson
Exactly.

Bryan Salesky
And I think what I, and I was going to segway to what you are doing, I think what’s neat about what you’re doing is first off, it can be done in a fraction of the cost of what gets plowed into these First Teams, because these robots are, especially the upperclassmen, they are fielding very expensive systems. There’s a lot of money spent on 80, 20.

Chris Anderson
Absolutely. $10,000 is not out of the realm.

Alex Roy
That’s like the entry fee. And so, what starts to happen is that it tends to be some of the actually wealthier areas they are able to actually be successful. I see this as a way that we could actually really democratize for the first time.

Alex Roy
How do you feel about, there’re some of these autonomous race series, Roborace. I mean, these are, what do you think of those guys?

Chris Anderson
I don’t know a lot about them. I certainly know I don’t have one. I can’t do it.

Alex Roy
No one has one.

Chris Anderson
Yes. It looks super cool, they seem to have gotten some awesome designer…

Alex Roy
Actually the designer is David Simon. You know him?

Chris Anderson
No.

Alex Roy
I think you do know him. He’s very cool. He designed the vehicles in Tron: Legacy. And he’s designed a lot of various future concepts, and they brought him in, and that’s why they look like that.

Chris Anderson
It feels like sort of the worst of both worlds, which is that it’s not participatory. I can’t do it. And it’s actually probably a little dull as a spectator sport, because they don’t like crashing, and they’re not really going wheel to wheel, et cetera. So, I don’t quite get it. Also I think they cost a million dollars plus each.

Alex Roy
Probably more.

Chris Anderson
Yeah, more. But they are using the Nvidia stack, and I guess so are we. We’re using something called ISAAC, which is kind of the replacement of ROS. There may be some similarity, but they represent these sort of unattainable pro side, and we’re the attainable amateurs side.

Alex Roy
But you said that you don’t think it’s a spectator sport, and yet DIY Robocars is a spectator sport.

Chris Anderson
We crash a lot.

Alex Roy
Okay, so people are there for the crashes?

Chris Anderson
And.

Alex Roy
And the learnings. So, what is it about having a, I can understand it from the learning standpoint, it makes sense to have very cheap machines moving around that you could damage. But do you think that there’s a market for a sport of machines competing against machines?

Chris Anderson
Sure. I mean, DJI the drone giant out of China has just released a DIY Robocar with a gun called RobotMaster S1.

Bryan Salesky
What kind of gun?

Chris Anderson
It’s a airsoft gun.

Bryan Salesky
Oh, thank goodness.

Chris Anderson
Yeah. But it shoots these like, it’s like a gel pellets. Anyway…

Bryan Salesky
Do they explode?

Chris Anderson
Kind of. I mean, they splat.

Bryan Salesky
I don’t want anyone listening to this to hear this from our office because they’re going to start shooting these indoors. I know. They have nerf guns everywhere. And did you know they have Nerf guns with like high volume magazines, I’m sorry, high capacity magazines?

Chris Anderson
When my children decide not to do robotics, they decide to do a lot nerf gun.

Alex Roy
I’ll tell you, the thing that I object to at high capacity children’s toys of any format is that if you think that by trying three or four times you’re to learn something, you learn accuracy by having to plan what the trajectory that, I mean, no, no seriously. I remember being a kid, and there were lawn darts.

Bryan Salesky
I remember lawn darts.

Alex Roy
These big heavy things, and now they’re illegal.

Bryan Salesky
They became things I threw at my cousin.

Alex Roy
Well, they became things that I learned not to throw unless I knew what was on the other side, which is a metaphor for life. So, I’m curious, have you seen the movie Real Steel?

Chris Anderson
I have.

Alex Roy
What? Why Bryan, have you seen Real Steel?

Bryan Salesky
I have. Carry on.

Alex Roy
You know what I’m going say, right?

Bryan Salesky
I think so.

Alex Roy
I feel like the movie Real Steel is an unappreciated gem with lessons for us about… What? Bryan, why are you making a face?

Bryan Salesky
I just, your movie references. You pick the most terrible movie.

Alex Roy
No, but…

Bryan Salesky
Go ahead.

Alex Roy
Real Steel is a movie that has a script that you can see where the edits were made by hints. These guys, for anyone who hasn’t seen it. Hugh Jackman is a Dad is, he’s a single dad, and his son. But in the future, boxing is outlawed. It’s considered human on human boxing, immoral and violent. I mean, you start some parallels here with driving, because human driving can be argued it’s dangerous, and in some ways immoral. And so, they’ve replaced human on human boxing with robot and robot boxing. And Hugh Jackman as a result, and out of work boxer, much like some people might say about some truck drivers might be, who knows. There’s all these metaphors.

Alex Roy
And then, so he acquires a cheap training boxing robot which has some basic foundational combat skills. It has a defensive AI, but it’s not good at offense. And so, they throw it into the arena with better trained robots, and they’re losing obviously. And in order to win, Hugh Jackman has to use like a neural net interface to remotely control it for offense to do things the machines basically AI can’t do. And so, when Robo Race reached like an inflection point the business where people did not really seem to want to watch too expensive vehicles go round a track, and they weren’t even racing each other. They had predetermined paths which is not interesting.

Chris Anderson
And not even that fast.

Alex Roy
No. And the race craft, there’s no actual race craft, there’s no sport to it. And so, they replaced, I remember talking to someone involved, and I say, “You guys should leave the basic path planning and motion planning and path following. Leave that. And have a human in the pit with like a pass button, the humans decide when to make the pass. The machine decides how to make it, but the decision points to the human.” They’re like, “No, we got a better idea.” So, they hired human race car drivers, and they let the AI drive half the race, and they put the human in the car for the second half, and the average the time. I said, “You guys really don’t understand how to optimize a combination of human and machine for the future.” And that is…

Chris Anderson
One of the reasons that First Robotics and Dean Kamen, the founder still do mostly teleoperation, is for theater, it’s actually more fun to watch humans…

Bryan Salesky
Supper fun.

Chris Anderson
Yeah. Do this. Autonomy is just actually quite boring to watch. Do you guys remember a toy called Anki Drive?

Alex Roy
Yeah. For sure.

Bryan Salesky
Yeah, we know Boris.

Chris Anderson
Anki Drive is exactly what you described, which is that they go around the track on their own, but you press a button, and it changes lanes, et cetera. They went bust because perhaps it wasn’t fun enough.

Bryan Salesky
Boy, I never thought of it that way. We thought it was the neatest thing ever.

Alex Roy
I think it’s cool.

Bryan Salesky
All the folks in the industry I know they would buy it. They would look forward to the next generation for Christmas and stuff. That’s a sad way to look at it, but it’s true, isn’t it?

Chris Anderson
We’re going to do an experiment.

Bryan Salesky
This soul crashing.

Chris Anderson
On September 21st, which is our next event. We’re going to do human vs. robot.

Alex Roy
In Oakland?

Chris Anderson
In Oakland. And unlike in the past where we had sort of terrible humans and terrible robots, and then we got terrible humans and good robots, now I’ve actually found a good human to drive. But we wanted to make it really apples to apples. The human is going to drive in first person view. So, both cars have camera in the front. One of them has a Jetson nano behind the camera, and the other one has Ross behind the camera. And that probably gets closer to what you’re looking for. Now, I suspect in the early days, the human is going to crush the robots because our robots don’t really have any race tactics, but we will learn and then someday Ross is going to lose, and we will all be lost with him really.

Alex Roy
But I mean, I would, even though in Formula One, the cars are all within like a second to each other, and so they’re fighting over milliseconds. There’s still sport, there’s races that are born, these races that are amazing. And I feel I think it’ll always be true as long as there is no absolute best. The targets always moving.

Chris Anderson
Take chess, or go, or one of these other games the computers get good. So, first there was like, computer will never beat a human. Then it was like, oh shit. Then it was like, a human will never be the computer. And now they have these human computer duos. What are they called? There’s a word for, so, what the computer’s really good at is understanding the total sum of possible moves. And what the humans… That’s what the computer is good at. What the human is good at is coming up with a strategy that has been time tested, et cetera. So, in the case of board games, what you described is now the case which is you have these human computer combos that can beat either humans or computers on their own.

Chris Anderson
We don’t know what that quite means. With a computer, one thing we’ve been thinking about is drifting. When you’re drifting as a human driver, you kind of feel the motion with your butt, I mean, you drive with your seat if you will. When you’re driving remotely a small car, you’re not in it, so you’ve lost that. So, can we use the IMUs, the sensors in the car to essentially handle that sort of seat physics, while the human does the track tactics?

Bryan Salesky
What I think is interesting is if you look at, Stanford did some work where they were actually drifting cars, and sort of, what’s the word? Power sliding into a parking space. You know the trick to that?

Chris Anderson
Mm-mm (negative).

Bryan Salesky
Is to do an open loop.

Chris Anderson
Oh?

Chris Anderson
That’s cheating.

Bryan Salesky
I know. But sometimes the simplest solution is the answer. At the end of the day, all those fancy sensor measurements are kind of garbage when you’re…

Chris Anderson
That sucks.

Bryan Salesky
At the edge of the envelope of what the thing can do, and so just…

Chris Anderson
What Bryan just said is actually really important. Let me just see if I could repeat what you said. What you said is when you’re at the edge of performance, the sensor readings are garbage. And as a driver, you probably know that. At a certain point you can’t even explain why you do what you do. You’re just operating on some kind of lizard instinct. It’s not your eyes, or your, it’s just something deeper. And that’s open loop.

Bryan Salesky
And that’s where it’s hard. How do you build that intuition that we don’t even know how we have.

Chris Anderson
We haven’t reached that humility yet. We still believe or not at the edge of sensing. We think faster processors, better sensors, et cetera. We can keep closing the loop closer and closer to nanoseconds. But we’re probably wrong, but we haven’t gotten there yet.

Bryan Salesky
That is right.

Alex Roy
You said when we spoke once before that there’s a, and I quote, a crash vacuum.

Chris Anderson
Yes.

Alex Roy
In self-driving cars. What did you mean by that.

Chris Anderson
What I mean is that you, am being a little flip as always, but in the history of automotive innovation has happened through racing from the beginning, whether it’s power trains, suspension, aerodynamics. It all happened in racing. On the grounds that the crashing and losing was an acceptable price to pay for innovation if you’re a Porsche, or Ferrari, et cetera. For some reason that heuristic has broken with self-driving cars, and that we’ve decided that safety is so paramount that you guys are not willing to push the envelope, and crash, and embarrass yourselves on a public stage.

Bryan Salesky
That’s correct.

Chris Anderson
Right. Now, it’s not obvious to me why that wasn’t true 50 years ago. Safety has always been paramount. Maybe that’s why Volvo doesn’t have a great racing team.

Alex Roy
That’s one of the best things I’ve ever heard.

Bryan Salesky
It’s pretty good, yeah.

Chris Anderson
Brian, let me ask you why don’t you guys race?

Bryan Salesky
Well, I think you also put it that some of the greatest inventions up to this point in automotive history come from in fact racing. In fact, I remember we had a company meeting at one point, and the team had an opportunity to talk to Bill Ford, and they asked Bill, they said, “Why do you do all this high performance racing stuff?” And he made the comment that look, a lot of the huge inventions that have fueled consumer automobiles actually come out of the racing side of things. I think I’m not smart enough to know what inventions or next big thing could pop out of like robot racing. It just isn’t obvious to me. But maybe we should take it up.

Chris Anderson
Well, I presume it’s never obvious to anybody until they do it.

Bryan Salesky
Until they’ve gone full circle.

Bryan Salesky
Yeah, maybe we should just do it.

Chris Anderson
I mean, you go out with a naive strategy, and then you lose, and then you ask, yourself why did I lose? How could I not lose next time? And then the strategies emerge. If you look at my toy here on the table and you say, “You’re a grown man playing with toys, what’s wrong with you?” And I’m like, “It’s a hobby.” What I really believe in, because we did this with drones. We believe that by taking a different evolutionary path towards self-driving, one that involves nimbleness and aggressiveness, rather than slowness and safety. One that involves, orders of magnitude more people. One that involves much higher risk. We might discover techniques that would lead to safer cars.

Chris Anderson
So I mean, there’s two ways to be safe. You can be slower, you can be nimble. Right now, self-driving cars, when they’re in doubt, stop. Whereas a racing car, when in doubt, act, switch lanes, do something aggressive.

Bryan Salesky
I mean, the goal is also to win it. And if you give up then you’re not winning. It’s a little different when you’re trying to get from A to B.

Chris Anderson
Exactly. So, let’s say that there’s two ways to avoid an accident. One of them is slam on your brakes. And the other is to quickly change lanes. You could argue that the second is the better one because you’re not endangering the traffic behind you.

Bryan Salesky
Absolutely.

Chris Anderson
Is that the way you think as well?

Bryan Salesky
It is absolutely. And we do think of those sorts of things. And until we can program in the capability to change lanes as an emergency avoidance maneuver, stopping is a good solution.

Chris Anderson
Right.

Bryan Salesky
But we absolutely need to think about all the different sort of fallback maneuvers that you might use.

Chris Anderson
To answer to your question, I’m just kind of hand waving here because we haven’t done it yet, but I suspect that a racer, a racing mentality would lead you to that, and go faster than the current one.

Bryan Salesky
It’s perhaps, I mean, it’s an interesting thing to think about sure.

Alex Roy
Recently I went through a test specialist driver training with Argo, was what Argo calls a safety driver, but it’s different from this.

Bryan Salesky
How well did you do in that program Alex?

Alex Roy
Not very well. But one of the most difficult things and I failed at it repeatedly was doing what’s called fault injection testing. So, you’re driving, or actually the vehicle is under autonomous control on a race track, and then the instructor in the right seat will enter a fault. The steering might juke to the right or left, vehicle might accelerate or brake, and the test specialist in the left seat has like milliseconds to do something to save the vehicle from following through with a jerk left or right, or the braking acceleration. And the solution is absolutely, for the steering is as easy, just grab the wheel, and hold on, keep it straight.

Alex Roy
But if it’s in a turn, then you have to carry through with the turn, and it might be an increasing or decreasing radius turn, so it’s not always just holding the wheel in the same position suppose that you want to be in, but the thing that was totally counterintuitive, and I could not pass this part, if the vehicle accidentally braked, you’re supposed to throttle through it to maintain speed to prevent a strike from behind because you might be in traffic. And if it accelerated, you’re supposed to brake but maintain, but brake just enough to maintain speed at a consistent speed, because among the calculations that you’re supposed to make is that around you are human actors, and they all anticipate what another human actor will do, but they cannot anticipate, and then the emergency maneuver of another human or machine.

Alex Roy
And so, that maneuver has to look as much as possible like normal operation so as not to induce panic, or other behaviors. And so, I realized that in testing, the optimal behavior of the vehicle is vastly superior to what my optimal behavior would be in the normal world. Whenever anything goes wrong, I want to stop the car in place, turn to the person next to me, and say, “What just happened? Are you okay?” I’m never thinking about getting off the road right away. I assume a tow truck will come, and paramedics will arrive. And so, that really reset my own understanding of what machines can do better than people. Even at the state in which they are right now, and I drove home, and I have a Tesla, I like my own grandmother, went home at 24.5 miles an hour every night after training.

Chris Anderson
Let’s imagine that you had a car that was supper grasshopper nimble, and could always avoid an accident, but at the risk of freaking out everybody around it, would that be a net negative for traffic as a whole?

Alex Roy
Well, if everybody had one, then there’d be a behavioral equilibrium…

Bryan Salesky
But they won’t.

Alex Roy
… And expectation as well.

Chris Anderson
Right now, we have a scenario where you have 90% human, and also, 99% human, and then 1% self-driving car, and it’s important self-driving car act like the human. So, that was your lesson from the test. Now, it could be worse than a human, which is slam on the brake, or it could be better than a human, which is sort of nimbly always, sort of like avoid accidents. The good news about that last one, nimbly avoiding accidents, is that they wouldn’t have accidents, but it might worry everyone around them, so they start to be feeling like the traffic’s less predictable, and maybe everybody gets a little sort of stressed out.

Alex Roy
Well, Bryan I think that’s what you’d call naturalistic behavior.

Bryan Salesky
That’s exactly…

Alex Roy
Programmed in that varying market to market.

Bryan Salesky
That’s exactly right. I guess I have to ask, are you a believer in self-driving cars?

Chris Anderson
Yes, absolutely. I mean, first of all, I’m a distracted driver. So, I’m not a believer in human driven cars.

Bryan Salesky
Driving is the distraction for you?

Chris Anderson
It is. I mean, I don’t like cars. I drive a…

Alex Roy
Tell us.

Chris Anderson
A Toyota Highlander thrashed, covered in dirt. I like texting. I don’t like driving. So, I mean exactly what you don’t want to have on the roads. Also someday, I’m going to get older, and I already have parents who are older. I’m really worried about them driving, and myself driving at that age. So, if nothing else the demographic shifts in this country, we only have two choices with old people, you either let them drive dangerously, or you take away their independence.

Bryan Salesky
That’s right. Especially in some cities where there just is, are no good options, and these poor people are just, they’re anchored to wherever they reside. And it’s tough.

Chris Anderson
Totally. I think you guys all know Voyage, and I’m a big fun of that. I don’t think that a self-driving car, I think of that as elder care. So, if you define self-driving cars narrowly as in like a particular sort of environment, huge believer. Everywhere, anytime, that’s harder. But that’s the way I feel about robotics in general.

Bryan Salesky
We agree. If you’re trying to boil the ocean, the product is never going to launch. So, what do you think the industry needs to do differently? Self-driving car industry.

Chris Anderson
Pick environments where it’s just a win win. This environments where you’re not going to get public opposition, environments where they have a clear problem that you’re solving. Things like that. We have, I live in Berkeley right now, and we have self-driving cars delivering our lunch every day, and they’re called Kiwi Bots, and they’re the size of a couple chairs, and they’re on the sidewalk, and they have cute faces, and little flags, and everyone loves them because they’re bringing lunch. So, that’s that’s kind of, and by the way, we had exactly the same problem with drones.

Chris Anderson
It’s like, we can darken the skies with flying robots, and everyone’s like, privacy invasion, noise, big brother. And so, we don’t do that anymore. Now, we pick just the places where we’re wanted. Mines, construction sites.

Bryan Salesky
And this speaks to not just building tech for tech’s sake, but building tech for the consumer, for the user.

Chris Anderson
Yeah.

Bryan Salesky
I don’t know why this is a novel concept to talk about but, it seems like it is.

Chris Anderson
Phase one is always exactly to say, it’s like, I can create this amazing technology, I want to spread it far, and wide, and then the world slaps you in the face, and says, listen up, that’s not what we want. And then phase two is figuring out where you’re wanted. And technology goes where it’s wanted. Invariably, it starts as supply driven, and ends up as demand driven. And I think right now we’ve been focused on the supply side of self-driving cars, and it’s time to really think hard about the demand side.

Alex Roy
You mentioned earlier that you your hobby became less fun when you started a company, that company is 3DR?

Chris Anderson
3DR. Yeah.

Bryan Salesky
And so, what is the product that, what does 3DR sell?

Chris Anderson
Well, we used to sell drones. We were America’s biggest producer of drones. And hundreds of thousands of them. And then that commoditization as it tends to. And they’re all made in China right now, and now we sell the data, we don’t sell the data, we sell the software product that uses the drones to do useful things. And our big markets are the things that either weren’t measured before. Our job is, I think I actually have this kind of somewhat grandiose theory of Silicon Valley, which is that all of us both in this room, and on the other side of the microphone, we’re blessed to live in, to have been born in an era where we had the internet.

Chris Anderson
And the internet is just, it’s the electricity of our day, and we have a holy mission, and that is to spread the mission, spread the internet far and wide. We should spread it to, we started with spreading it to our homes, and now our arms, and our cars, the streets, and some people took it to the sky, and space, and beyond. And when you spread the internet, when you take the Internet to the sky, which is essentially what a drone does, it does two amazing things. One is that you’re able to measure the world, and bring it back so you can manage it better. And the second is that you had the intelligence of the internet, and you get to spread that out to the devices.

Chris Anderson
So, the devices get smarter because they’re connected to the internet, and the internet gets smarter because it’s connected to the devices. So, it’s the internet of flying things if you will. That was great, took years to get to the point where we could just push a button and do that. Now the question is which part of the world needs to be measured best. And we kind of go where like Landsat went 50 years ago, whatever satellites were measuring, we can now measure at higher resolution both temporal, and spatial under the clouds cheaply, et cetera. So, today most of what we do right now is what used to called geospatial. So, it’s things like, right now we take parts of the world that are not measured, like construction and mining, et cetera, with measures so you can measure better.

Chris Anderson
But when you look at this, what were satellites measuring before, they measure water, environment, and population flows, this notion of specializing data, that recognizing all the world’s data could be put on a kind of a map, and it makes more sense on a map than it does on its own. We’re just populating the map.

Alex Roy
Do you have a pilot’s license?

Chris Anderson
I don’t.

Alex Roy
Because you seem to know a lot about aviation.

Chris Anderson
I have no interest in aviation.

Alex Roy
I see.

Chris Anderson
I know. I don’t have any interest in cars either. I have a real interest in spreading the Internet. The reason I pick the air is because there was no one else doing the air. And the reason I picked cars is because it was known that no one else, or not enough people doing self-driving cars. I’m really interested in the Internet.

Alex Roy
What is the one question you wish people would ask you that they never do? And since you were the editor in chief of Wired magazine. You must think about this all the time.

Chris Anderson
About the cars in particular?

Alex Roy
About the questions that people don’t ask home. I mean, I’m always, people ask me why do you, people think I’m against self-driving cars. I actually think there’s a huge use case. And what people ever ask me, why?

Chris Anderson
I find why to be a somewhat threatening question. I mean, I have a why. And I can answer that question. But a lot of people find it, parts of the world is not okay to ask people what they do for a living. Is like, you don’t define.

Bryan Salesky
It’s true.

Chris Anderson
Don’t define me by my job, et cetera. I think that one of the things we learned from the hacker community is that why is sometimes, the answer why sometimes, because, I can.

Bryan Salesky
And sometimes is very personal.

Chris Anderson
Exactly. It’s could be that they’re learning something. It could be that they feel a sense of satisfaction. They don’t always… We had Burning Man this next week. Why does Burning Man exist. I mean, if you can answer that question, the answer is as many reasons as there are attendees.

Bryan Salesky
I don’t get that one.

Alex Roy
People always ask me, “Why don’t you go?” And I’m like, “What do you mean?” And they like, “Alex, it seems made for you.”

Bryan Salesky
It does seem…

Alex Roy
And my answer to them is, “If you have to ask me why I don’t go, you don’t know me.”

Bryan Salesky
Before I knew you, totally wondered why you would not be going.

Alex Roy
Screw that.

Chris Anderson
I have reasons why I would want to go. And they involve things like music, or kinetic sculpture, or things like that. I’ve reasons why I can’t go, which is that my wife won’t let me. But yeah.

Bryan Salesky
She’s wise.

Chris Anderson
In general, why you have your vocation, your avocation. So, your vacation is your job, and your avocation is your hobby. Job’s the why is paycheck. It could be otherwise. But start with paycheck, and avocations, they more subtle. It’s because I have a sense of self, a sense of personal growth. Things like that.

Bryan Salesky
Well, I know the question that you really wanted everyone to ask you. So, I’ll just ask you right here. Chris, what do you think about big data?

Chris Anderson
Woz said define your terms. Was that the phrase?

Bryan Salesky
He did in fact. I felt for him because he struggled, it was just like, what is this even? Could that question even connect to…

Chris Anderson
By the way, I do get the question why a lot. And I think the answer you can’t ask why is kind of flip, and arrogant, so I don’t actually say that. Instead, I try to understand why people are asking the question. I think it’s either that they can’t believe I’m wasting my time with these stupid hobbies, or they’re generally, there’s like, Chris has typically, he’s been the editor of Wired, etc. Typically, there’s a reason he follows these things because there’s something there. What does he see that we’re not seeing, and I didn’t have an answer, sometimes it’s just instinct.

Chris Anderson
And so, what I’m trying to do is construct why do sort of sub-scale, inexpensive, democratized self-driving cars, where are they taking us. That’s that’s relevant and important. And over every race, I get closer to an answer. I don’t think I’ve done anything yet. I think my community’s done anything yet that would impress Bryan. nothing at all.

Alex Roy
I think you’d be surprised.

Chris Anderson
Yeah, I’m not sure about that. Look there’s, I’m big on wanting to find ways to help educate people on the technology, and to get folks into this field, and I think what you’re doing is hugely impactful to giving them, like I said, in a more affordable way platforms for them to learn and get involved, and that’s important.

Chris Anderson
Okay, thank you. I’ll take that. It feels a little remedial. That’s was not my ambition, that’s a side effect. That’s what I say when I’m trying to be, trying to get the girl scouts on board. But why I’m really doing this is that I want to someday discover some, when I say I, I mean, we the community. I want we to discover some technique algorithm approach that you Bryan hadn’t thought of, that you realize is the world speaking to you as a way to change your own company.

Bryan Salesky
To me that would just be gravy. But we’re going to be watching to see if that happens. Watching very closely.

Chris Anderson
Give us an assignment. Gave us a couple of problems that, my perception of, loaded word. My sense of the industry right now is that perception is largely a solved problem. And I could be wrong, but you my sense is that that’s kind of where we are, where we’re trending, but that the decision making, the path finding is an unsolved problem.

Bryan Salesky
I don’t know that I would frame it exactly that way, but what I would say is that the perception problem is vast, and it includes a lot of elements to it. The piece that is still extremely difficult is the prediction.

Chris Anderson
Right. That’s what I meant.

Bryan Salesky
Humans are incredibly good with such few observations. We’re able to extrapolate with a very high level of accuracy what a pedestrian, a scooter, you name your object, what it’s going to do next over the next several frames. Teaching a computer to do that is extremely hard.

Chris Anderson
Totally. Prediction is a better word than path finding decision. But getting back to the board games that have been beaten by computers. So, when Deep Blue beat Kasparov, or where AlphaGo beats Go, everyone’s like, oh my God. That’s it. They’re sentient , Skynet, et cetera. And it’s like, no, that’s completely unrealistic because the board is perfect information.

Alex Roy
Also, if you ever hung out with a guy who’s really good at chess, they’re horrible people.

Bryan Salesky
Real pain. Yes I have in fact, yeah.

Chris Anderson
So, basically what they did is there’s two variables. One is like, what’s the state of the world today. And the second is, what shall I do about that. Right now with cars, both were variables, but if perception becomes solved, you’re getting close to perfect information. And then you can reduce the problem sets to, okay, given the state of the board, what shall I do. I think we’re closing it down now. That’s orders of magnitude harder in the real world than it is even in Go, but at least we’re closing into I think that problem set. And then if we can shrink the problem said even smaller to rather being the world as a whole, but rather maybe just a racetrack, maybe we can complete your prediction problem on a very very small scale.

Bryan Salesky
I think that’d be pretty fantastic.

Alex Roy
Before we wrap up this conversation which I would love to have going all afternoon. I know Bryan and I and a lot of folks often debate like what it will take to earn trust, to compel people to trust a machine whereby previously they were in control, and now they’re no longer in control. How do you come to trust your kids to put in a self-driving car? What has to happen? What do you have to see?

Chris Anderson
We think about this a lot since we’ve been doing this with drones. So, the first thing was we didn’t put people on them. It’s a robot aircraft, but it’s small, and the risk factors are low. And then the next point is, okay, at what point do you feel it’s reliable enough to put a human in, and will we ever get there? And the answer is of course, we will get there because we got there with elevators. And your airplane, your 777 is a drone for 99 percent of the trip. But we have this theater around it with the airplane. We have two pilots, not one, we have two pilots.

Chris Anderson
We’ve got flight attendants and all this kind of stuff. And so, what if you, we know that these aircraft can taxi, take off, land, get back to the gate on their own. What if we went from two pilots to one pilot? What if we went from no pilot in the cockpit, but teller operated pilot with like an inflatable Avatar, or something like that.

Bryan Salesky
Like that movie Airplane II?

Chris Anderson
Exactly. I think we have to bring something bring some theater in to reassure people.

Alex Roy
Aren’t we in the theater phase? We’re entering the theater phase.

Chris Anderson
Well, you’re a safety driver.

Alex Roy
Test specialist.

Chris Anderson
Test specialist. Exactly. Right now it’s still RD. But at some point, it’s going to be theater. And the question is how long, so, how long do you have to be a physical person there? Could you then just be a screen like one of those teller presents bots, et cetera. I think it’s really about managing people’s sense that there’s always a human within reach.

Alex Roy
Maybe it could be as simple, I can’t believe I’m saying this, as you have someone in the AV, and then eventually after some period of time, you allow the passenger to choose whether they’re willing to pay more to get the person not in the AV. And that’s when trust has happened. You let trust just, the market decide like the market decides.

Chris Anderson
Do you think… We have trains, of course, they don’t have drivers and airport shuttles and things like that. Was that ever controversial?

Bryan Salesky
I mean, at one point it was controversial taking the operator out of the elevator.

Chris Anderson
On safety grounds?

Bryan Salesky
Yeah. That’s right.

Alex Roy
Was it controversial to remove the flight engineer from a commercial airliner?

Bryan Salesky
I’m not sure.

Chris Anderson
Yeah.

Alex Roy
Because at one point they had like four, five people in the cockpit.

Chris Anderson
So, it’s definitely going to get, it used to be four, then three, then two.

Bryan Salesky
Society’s gone through these inflection points. I think what’s different about the world right now though is that tech fatigue is setting in. And so, I think people have a heightened awareness of what’s going on than maybe has been in the past.

Chris Anderson
Well, I think the answer is going to be teleoperation. I think that we’re going to go, two pilots, one pilot, teller operated. And the question is I think with cars, we’re probably going to go one driver, tele-operated, et cetera. And the question is, how do you present that tele-operated.

Bryan Salesky
Well, there’s a lot of problems with teleoperation though just because if the cell, the link to your outside world goes away or if the latency shoots up, compensating for all of that. It’s really, it’s pretty tough.

Chris Anderson
In with drones we call them optionally piloted. This will be optionally teller operated. That’s why they’re called Kiwi bots that we have all day. They can drive down the sidewalk on their own, when they get to the corner, a dude in Columbia with 16 screens …

Bryan Salesky
So, it’s from a system perspective absolutely.

Alex Roy
We have to go back, I really want to ask one more question, Brian May I?

Bryan Salesky
Sure.

Alex Roy
You were the editor in chief of Wired during a critical error in the growth of technology in society era. And so, you were the editor in chief of Wired during a critical era, technology in entry society. What was the worst, have we been through a worse era in terms of misinformation around an emerging technology? It seems to me that autonomous vehicles are really horribly portrayed on both sides, both by the companies, and the media itself. As anything like this happened before?

Chris Anderson
Always invariably happens. The telegraph was portrayed this way, electricity was portrayed this way. This is completely normal. So, I was, by the way, I think that my predecessors at Wired from 93 to 2000 were really the ones who were there at the critical age. The reason I came to Wired was largely, I was trained as a computational physicist. And at the very very worst time. Basically the last interesting experiments that could be done in high energy physics were done like two years before I graduated. So, there was like, at this point, there was no way to validate your theories anymore. And when theory and experiments starts moving apart in time, theory turns into philosophy, fiction, et cetera.

Chris Anderson
So, we lost our groundings, and it sucked. However three great things came out of physics. Number one the internet. That’s what the internet was forced to connect physics labs, super-computing labs. Two was the actual web created CERN, physics lab. And three was big data. There we are. Back at big data. Physics labs had the first big data, and so, all the statistical techniques came out of that. So, I didn’t know that at the time because I was just a physics nerd. But then this weird magazine called Wired came out in 93. And it was like, this is not just a tool for physicists. This is this is going to change the world. This is a social revolution, political revolution, et cetera. And I’m like, holy shit, I’m in the right place, at the right time, and didn’t know it.

Chris Anderson
The tools are more interesting than the science. And so, I went off and decided to make my career around the internet. So, that was thanks to Wired back in the day. And then when, that was at The Economist, and elsewhere, and then when the opportune came over to edit Wired in 2001, remember this is right after the dot com crash. And the dot com crash wasn’t just bad economically, there was a huge contingent people out there who said, “This internet hoax, is the subprime mortgage of its day. It is time that these arrogant bastards in Silicon Valley were shown to be the snake oil salesmen they are.”

Bryan Salesky
What era is this?

Chris Anderson
This was this was 2001.

Bryan Salesky
One, yeah, exactly.

Chris Anderson
These are the telecoms companies who’d lost control, the networks, et cetera. And they’re all like…

Alex Roy
Maybe this Internet will go away.

Chris Anderson
This internet will go away. We knew it was a scam. And I had to make a bet then that the crash was due to a stock market bubble, not a technology bubble. And then the fact that the Internet was transcendent, and real, and would only rise which turned out to be right. But that was way worse than this.

Alex Roy
If we want to learn more about Chris Anderson, where can we find you?

Chris Anderson
Well, don’t go to my Wikipedia entry because I’m scared to even look at that.

Alex Roy
That’s where you and I differ.

Chris Anderson
Diyrobocars.com is the site where we have the links to all this kind of stuff. We have 10,000 members, and various meet ups, I think 70 meet ups around the around the world. You can find links to all those. We have a bunch of projects that you can participate in, things like donkey car, the jet racer, the Amazon DeepRacer. There are guides to get started, and I’m on Twitter all the time as well like you.

Alex Roy
What’s your Twitter handle?

Chris Anderson
It’s @Chr1sA with the I being a number one.

Bryan Salesky
Clever.

Chris Anderson
Yeah. You can already imagine the other Chris A was already taken.

Alex Roy
That guy sucks.

Chris Anderson
But now I’m number one.

Alex Roy
A fun episode. But what was your favorite part Bryan?

Bryan Salesky
The part where he had no idea who I was, I always enjoy that.

Alex Roy
The part where he thought you were the IT guy?

Bryan Salesky
Yeah. I don’t know what he thought I was. I think he thought I was like security?

Alex Roy
Isn’t it great to know that your reputation so totally dwarfs you?

Bryan Salesky
I completely enjoy that.

Alex Roy
Yeah. It’s cool. I heard a rumor that when you go to Dearborn or Detroit, and you hang out in the lobby of a hotel, that some people ask you if you’re the Argo guy?

Bryan Salesky
You’re the guy.

Alex Roy
And what do you say?

Bryan Salesky
I say yeah. I’m part of that company.

Alex Roy
I would anything to come and hear you say, “No, I’m not the Argo guy. What makes you think I am? Do I really look like him?”

Bryan Salesky
There’s another one that’s just like me.

Alex Roy
The other guy. No, I’m the IT guy.

Bryan Salesky
All right. To be clear, it only happens in that one hotel. It’s right there on Ford campus.

Alex Roy
Well, the next time that happens, if I’m around, I’m going to quote, I see, I’m going to look at them and say, “And who the F are you?” All right. If you would like to learn more about the No Parking Podcast, check us out online www.noparkingpodcast.com. You can check at us on Twitter @noparkingpod. Brian doesn’t do social media, so don’t even ask.

Bryan Salesky
Where do we find Chris’ stuff?

Alex Roy
Diyrobocars.com. And I’m AlexRoy144 on all platforms. And if you want to be a guest on our show, or know someone who’d be a good guest, or suggest a topic, you can send us a message at guests@noparkingpodcast.com. See you next week.