Episode 236 - Regulatory Guerrillas
Published January 12, 2016
SPECIAL GUEST: John Robb (Global Guerrillas). We're joined by JOHN ROBB, author/entrepreneur/inventor/former USAF pilot, for some futurist looks into drones and self-driving cars. How does the current situation with these technologies mirror the early days of the Internet, what possible ways they will transform society, and through it all, how is DEEP LEARNING reshaping our lives. With a side of social networking, including our favorite kicking target FACEBOOK. Recorded on 1/7/2016.
Mike & Matt's Recommended Reading:
John's blog Global Guerrillas, about Networked tribes, system disruption and the emerging bazaar of violence, the future of conflict.
John on Twitter
Global Guerrillas on Facebook
John's Wikipedia entry
Google Brain: "Can we build a centralized machine leaning platform?" Smarter Devices for our Connected Environments, suggested "find out more" link by our Twitter friend, Nicholas Perry
More coming soon...
Alpha: Welcome to another episode of Robot Overlordz, episode #236. On the show we take a look at how society is changing, everything from pop culture reviews to political commentary, technology trends to social norms, all in about thirty minutes or less, every Tuesday and Thursday.
Mike Johnston: Greetings cyborgs, robots, and natural humans. I’m Mike Johnston.
Matt Bolton: And I’m Matt Bolton.
MJ: And joining us tonight is author John Robb. John, thanks for joining us.
John Robb: Oh, my pleasure.
MJ: For the benefits of our listeners, if you had to pick sort of a “Reader’s Digest” version of your background, could you kind of an idea of what that would be?
JR: Oh, wow. I was a pilot in the Air Force, Special Ops, worked with Delta and SEAL Team Six for a while; was an analyst with Forrester—if you’re familiar with Forrester research, I was their first internet analyst in ‘95/’96, so I covered all the internet companies as they took off. And I started companies, mostly software companies, in finance and printing. Did some social media stuff; did most of the early blogging technology, RSS came out of my little company, which then became the kind of thing we saw on Twitter and Facebook and everything else later. Then I wrote a book on guerrilla warfare called Brave New War, but, you know, a next-generation war, and that proved pretty popular. And I write a blog called Global Guerrillas, and I occasionally tweet on @JohnRobb, which is the tweet handle. That’s the nutshell.
MJ: Well, I guess to share a little tidbit there for our listeners, you were actually one of the articles that we used in our very first episode #0. Since you do write a lot about the internet, and certainly within sort of the futurist podcast segment of the world we talk a lot about accelerating change, I guess when you look at sort of the state of the internet, kind of the state of how it’s impacted society or the way the internet has developed, are there any really big problems that you see that are maybe trending and that aren’t really on peoples’ radars yet?
JR: In technology in general, or in regards to the internet in particular?
MJ: I guess I’m thinking the internet in particular. I know you’ve written quite a bit about the social networks and Uber and things like that.
JR: Yeah, I’ve been pretty lucky in terms of hearing the “sound of the cannon” and being able to ride towards that a couple times in my life. Like early on with the internet and being kind of like the top internet analyst covering everyone in ‘94, ‘95, and ‘96, and hearing the finance sector kind of taking off in ‘97 and going over there and starting a company in that space—I mean, back when nobody thought the internet would actually be used for anything other than chatting. We were really happy to see stock quotes pop up on screen, which was really cool. But it became huge. I mean, we had brokerages going 50% of their business in a year transferring to online. And then with social media in 2001, that was WAY before everybody else was really into it. So, I’ve been able to kind of ride from wave to wave to wave for a while and participate in the early stages, and I haven’t really seen much besides a little bit in the sharing economy—I did that three or four years ago, trying to come up with a kind of open source venture kind of framework, trying to reinvent the company, and got involved with some bitcoin companies—blockchain companies, more specifically. You know, the thing I think is really going to zoom here, it’s going to ride on the back of the internet, and it’s something I’ve been kind of involved in in the last year… I got tapped, because of my writing on Global Guerrillas about warfare in general and about technology, I was asked to work with the Joint Chiefs of Staff. The chairman of the Joint Chiefs of Staff was writing a paper on his vision of what the military would look like in 20 years if transformed by robotics and autonomous systems. It gave me a chance to talk to everybody in the space and look at all the technology and see what everyone was talking about, and gave me kind of an amazing overview right at this early juncture, writing this paper for him. What I saw when I got into it was that this is really much bigger than just the kind of drones and robots that we see now. What we’re seeing is a massive, massive wave of change, bigger than the internet, I think, in terms of the overarching impact, due to cognitive machines, mostly through deep learning, and how you can build these cognitive machines and how you can use cloud robotics and cloud systems, where once you teach a machine how to do something, you can share with every other machine via this cloud. And how that can get really, really good very, very quickly, kind of the exponential, kind of “step on the accelerator” amount of change that we saw early on when we started connecting computers. I mean we had all those computers installed and we had all those copper wires going to everybody’s house, and all we did was run a modem; started running these modem farms and start connecting them up, and it went really, really quick. That’s where I think all the big change is going to come, is this internet of things isn’t really internet of things, it’s this connection of smart stuff—cognitive machines, cognitive systems, systems that learn as opposed to systems we program—and connecting them all up, and they’re going to be in everything. These narrow and deep intelligences are going to make these devices, make these systems extremely smart and interconnected. So, it’s different than a lot of people think it’s going to come, in terms of internet of things. Like, connecting up to the refrigerator—that’s not really interesting. Connecting up to smart stuff and having the smart stuff connect, now that’s interesting.
MJ: That brings up an interesting question. When you kind of talk internet of things, it seems like there’s such a disconnect between some of the writing on the theory, or the state of the industry or where it might go for industrial uses, or the military, and what the consumer kind of sees. You mentioned the connected refrigerator—Matt and I continue to kind of, a little bit, make fun of things like that…
JR: Oh yeah, you should.
MJ: …Because lot of the things they seem to use those technologies for, they seem to be solving problems that don’t exist.
MJ: When you look at the deep learning stuff, I guess how would you see that actually making it into peoples’ lives? Or do you have any idea yet, even?
JR: Yeah, okay, so, you’re exactly right: all of the kind of push, all the change they’ve been talking about in terms of internet of things has been stillborn. It really has been crappy. It’s not solving a problem anyone needs solved. And it’s very similar to the kind of way the industry in ‘94, the telephone and telecommunications giants saw the future of networking—they saw kind of interactive television. [laughs] And that was stillborn, that was horrible when this new thing was just getting started, this internet, this uncontrolled thing, this data-driven and interactive thing, as opposed to their kind of delivering content cable plus. What can deep learning do? You’re not interested in a smart refrigerator per se, but deep learning is already transforming our lives in really subtle ways, and the real benefit of this stuff hasn’t come out. For instance, if you’ve noticed, the last two years we’ve made more progress in voice recognition and image recognition than we have in all the other decades prior. It’s gotten that good, that quick. I mean, it’s above human in terms of its capabilities—and that’s using deep learning. I mean, if you dictate into Google now, it will capture every single word, it doesn’t mess up. And just two years ago, it was sloppy and it wasn’t progressing at any appreciable rate. So what that means then is you can start to get to voice-based interfaces that have been promised for years and never occurred; you can start to have systems that recognize pictures and images and do it at a better-than-human capability, which means that you can have a world that’s interacting using just simple characters. Just like Tesla, for instance, their self-driving cars are all focused on using just simple camera systems. They don’t want to use the LIDARs and the expensive equipment. They’re just using that simple imaging and deep learning in order to teach these cars to drive. That kind of stuff can transform everything. The biggest impact that it’s probably going to have—I mean, outside of those two interfaces; image recognition, in terms of the ability, for instance like for DeepFace, which is Facebook’s internal project using deep learning, they’ve trained it to recognize 800 million profiles, people, based on pictures. If you’re in that pool of people, that 800 million, and somebody submits your picture into the system, it will recognize you in five seconds. I mean, that’s kind of a transformative thing in terms of everything from counterinsurgency to police… And the thing is that the whole system can run, once it’s crunched, on a cell phone. So, it can be transformative in terms of how you’re authenticating transactions and the like. In terms of cars, self-driving cars are huge and they’re coming so fast, it’s kind of scary. Nissan, for instance, is going to have ten different models with self-driving capability by 2020 now. Everybody is just going to be launching these self-driving cars that will take all the hassle out of driving in traffic, driving through intersections, and soon you’ll see the kind of pressure on the regulatory agencies to kind of force people out of driving, because they’re just too dangerous. I mean, self-driving cars can save upwards of 35,000 lives a year. And then, you know, if you’re driving by hand, people are going to start to say—you ever hear the “reasonable man” theory in civil law?
JR: I mean, people will start to say, “It’s not reasonable for you to drive, because you’ve killed somebody, or you cost an accident. You aren’t a reasonable man to drive because self-driving is so much better.” And then what happens to your insurance rates when that happens? They become prohibitive, you become liable. It’d be impossible to drive by hand—and transportation, for instance, is, what, the biggest employer in the United States?
MJ: Yeah, we talk self-driving quite a bit, I guess.
JR: Yeah, but that’s deep learning. If you get that last 10%, in terms of the quality of the driving experience, that’s deep learning. And it’s also using cloud robotics, meaning that when Tesla has these guys out training on the Model S right now and they’re training those systems how to drive, they’re training deep learning systems, and they’re also sharing data on roads and on technique. About a million miles a day are being accumulated up into their central systems. They’re getting good at it. And you can see this kind of scale over a lot of different things—Toyota has invested, what, a billion dollars, or they’re in the process of investing about a billion dollars into autonomous systems. Half of that will go into cars. A guy who’s running the whole project for them in the States is the guy who used to run DARPA, Gill Pratt, up until just recently, up to this year—last year, actually. The other half is going to be on domestic robots, and what I mean by that is teaching these bots using cloud robotics and simple protocols of interconnection and coordination and deep learning to do domestic duty, like take care of elderly people, which Japan is full of, China is going to be full of, we’re full of, providing that kind of extra care that everyone is going to need when they get older and they just don’t have the body necessary to take care of them—which is huge, it’s a huge industry.
MJ: Yeah. When you look at all of those changes, do you think that society itself is having the right conversations already to kind of get ready for them and get out in front of them? I mean, I think of things like the self-driving car, and obviously all it takes really is looking around you when you’re on the road, and all the people that are either texting, or they’re on their phone, or they’re messing with their radio, and it becomes pretty hard to defend people driving, certainly, that right.
JR: Oh, they’re terrible. [laughs] Terrible drivers. They shouldn’t be driving at all.
MJ: Do you think that we lose something by at least not thinking these through? That we’re giving something up? Obviously we’re gaining something, but what are the downsides?
JR: In terms of driving? None. Okay, let’s break this down: are there any conversations we could have now that prepare us? I’m not sure there are any kind of pre-planned or structured conversations that we can have. People can cover the area and write about it and put it in fiction and discuss it on programs like this and get people ready. That’s about it. We could have a little more insight in terms of how to set up the regulatory environment to make this transition smoother, because right now it’s kind of similar to what we saw in ‘94/’95. With the FCC, there’s a government agency that’s actually kind of gating the progress that we’re going to have. In ‘94, it was the FCC, in ‘95 it was the FCC. They could’ve shut down the internet really early on by letting the phone companies control access. They said, “It was our copper. Let us own this. Let us deliver the service,” which would’ve driven up the cost and really slowed it down, and then they would’ve controlled the content and the whole net neutrality thing would’ve been thrown out, would’ve not even emerged at all. It would’ve just been slow. We’d be 15 years ago right now in terms of how the internet progressed. But we didn’t do that. We made it possible to kind of zoom the whole process and now the US has more internet companies than anywhere, and we benefit from that as a country. I think the whole world benefits from that, from us leading the way into this internet age. Now, in this case, in terms of robotics and deep learning, the DMV and the FAA are gating the whole process. They’re putting on rules and slowing it down, and they’re just doing it to protect existing interests at the expense of future interests, they don’t see the potential of what could happen and what is possible. And by slowing it down, they risk damaging long-term economic prospects in the US and making sure that many more of the companies and much more of the capability that would be developed over time is developed overseas and not here, which will damage everyone—security, everything, across the board.
MB: Do you think with self-driving cars there’s going to be some pushback? Especially from police, which you wouldn’t think, but there are a lot of towns and municipalities and things across the country that fund almost their entire government on the traffic tickets, parking tickets, those sorts of things, and if you do away with all of that, I would think that there would be some kind of pushback for that.
JR: Yeah, there’s going to be a little pushback, but I don’t think they have much of a choice. I mean, the savings in lives, the benefit to society… Here’s the transition on this thing, at least in cars: you get the electric cars, big reduction in parts, big reduction in manufacturing cost…a lot of drive down in cost structure on that. The fuel is three cents a mile once you go electric—it’s much more in gas, and it fluctuates a lot more. Self-driving in addition to that, which then makes it possible to do fleet cars or fleet auto-driving cars or driving 24/7, so they’re never stopping. So the fleet costs go down, and they go down and they go down and they go down, and pretty soon you’re just doing everything via fleet. You’re doing it as a service, you’re driving as a service, you’re not owning a car. And once you get to that point, there’s a potential here, and I think this is highly likely given the way these big companies work at scale, is that they’ll offer driving for free in most metropolitan areas because the value of you inside that car is a lot more than the cost of charging you for that distance. So, would the municipalities try to stop that? I don’t think they have a chance. [laughs] What you can do, though, once these things are all connected is you could start to do toll roads, and tolls become really, really easy. If you see the transition, they’re already now in, like, Massachusetts and other places, you went straight from fast lane only, being pushed back to just a couple lanes and they still had those by-hand toll takers. Now they’re just available on every lane and now they’re saying, “Well, we don’t even really care whether you do fast lane or not. We’re just going to charge you based on a picture of your license plate. So, we’ll take the picture of your license plate, we can see you in the car.” There’s almost no error because the learning programs are so good in terms of recognizing license plates. That means you can actually put tolls anywhere, and all it has to be is on the electronic map that says there’s a toll here and your car will figure out whether it’s cheaper or more expensive to go down X, Y, Z road. You can get to that kind of libertarian paradise where, I don’t know if this would work, but it is certainly possible to start charging for the cost of the road repair; every single person that uses it actually pays for that road.
MB: I think it’s in Oregon, they’re trying to roll out a system where you will pay per mile. You have to put in a GPS and they’ll monitor and then, you know, send you a bill at the end of every month for the miles that you’ve driven.
JR: Yeah, you could do that. I think it’s even more powerful or more likely that kind of ground-up, grassroots kind of thing, where towns implement it on a per-town basis, and they just put the cameras up. It’s so connected now—I think we had a late registration of one car, okay. What happened is that the cops don’t even look at your license plate anymore, they just set up a scanner using the terrorist money, scan the car, or one of our cars if somebody else is driving, and it threw up all sorts of red flags and the cop rushed out trying to catch, you know, this license holder with this late registration on another car. And, you know, I wasn’t in the car obviously, because I didn’t get the ticket. The thing is it’s all interconnected, everything down to even the parking ticket can cause you to not be able to, you know, re-up your license plate.
MB: With all of these things being interconnected this way, and it’s obviously getting more and more and more, how concerned should we be with the security of all of this stuff? Because I saw like this week Ukraine had their power grid hacked into.
JR: Yeah, Russians tend to do that with their neighbors.
MB: Yeah. [laughs]
JR: They did that with Estonia… And the thing is that they’re actually much more vulnerable to this. I was always kind of wondering why the Georgians and the Estonians—the Estonians are fantastic programmers—even spend any money on the kind of physical defenses when it would be so much easier to have a couple of Tiger teams or Bear teams sitting out there in some secure area ready to hack Russian systems, just click it and make it really costly for any kind of Russian encroachment or effort. The Ukrainians have tremendous talent, too. Russia is much bigger, much wealthier, and a much bigger target. In terms of our security, I’m not really that worried about it. I’m worried more about just the simple abuse, because these systems aren’t really deep learning systems, these are just kind of interconnected algorithms and interconnected systems that become interconnected and people forget about how they’re even connected, really. In my case, you know, it was a ticket at a community college that I never visited that was transcribed wrong, the wrong car, that caused the ticket to come on the community college, which stopped the registration from going, which then caused the cop with the scanner to pick up on the license plate that was attached to my license that was a different car, and then stop the car. So you see how that whole chain of events happened? You know, it’s just when you have this kind of interconnection, there’s really this big machine that won’t say sorry, won’t give in at all, and there’s not even any human being involved, it’s just this thing that can keep on charging you more and more interest. It’s like payday loans gone national.
MJ: John, as those systems get more and more interconnected and, like you said, there’s really no human behind it, do you think we’re at risk of another moment, like you mentioned earlier, the moment in ‘94/’95 where the telecoms really tried to make the power grab for the internet, and fortunately they weren’t able to do that—do you think we’re at risk of another one of those moments where this time it might go the other way, potentially? I mean when you look at companies, like Facebook has all this data on us… I have a friend who’s very paranoid about Facebook and, like you said, their algorithm can pick out faces so easily. He’s tried to stay off it, but I know—
JR: It won’t matter.
MJ: Yeah, you’re on it anyway.
JR: It’s not even them. It’s once that system gets going, the big national systems will start combining the stuff on Facebook with the stuff in Google, with the stuff on everybody’s photo processing sites, you know what I mean? You can start building these kind of big, big databases. If he’s ever had a picture taken with a digital camera, he’s probably on it. There’s a repository somewhere that’s going to be picked up and then connected to this, and the system will learn it. I’m not so worried about the crackdown coming that way, or the abuse coming from that direction. What I was worried about right now that’s similar to the FCC is the FAA going, “Okay, we don’t know about drones, there’s a potential terrorist threat,” which there is, largely because we don’t have a system for actually managing a larger number of drones in the air. But we can’t have a system for managing large numbers of drones in the air until the FAA gets off its stuff and writes some simple rules, like adopts kind of an internet protocol for drone movement and identification. They’re not even ready and they’re not even capable of getting that done. So NASA’s working a little bit on that, but still, in the meantime they’ve shut it down and said, “Okay, you need to register your drone, you need to keep it within eyesight…” You can’t take your eyes off the darn thing.” You can only fly during the day, anything beyond that then you’d have to start talking about having a real pilot fly it.” The very reason that drones are popular today is because they don’t need a pilot, anyone can fly ‘em. That was the big breakthrough; autopilots became $30 things you can stick in the drone. So what that stops is the stuff that people don’t even understand that they’re missing. Like that Amazon delivery service—well, that’s actually huge. I mean, I really think that could be transformative. I mean, the small package delivery system that can come about through drones can set up great industries that we don’t even have any idea could exist—because getting across the country for a couple dollars in less than a day is pretty cool. It’s not something we even can get our heads around. And once people can start doing that in kind of an open way, in an internet kind of way, moving stuff from point A to point B in line of sight if somebody just stood the package outside their door, that would be very good for us if the US could do that first. So they’re blocking it because they’re just not able to handle the kind of vision associated with making this happen. And the same thing with the DMV. They’re getting pushback, and they’re getting worried, and they’re dragging their feet, and the big car makers or existing carmakers are saying, “Go a little slower so we can catch up,” which they’ll never be able to fully catch up even with alliances starting to put limitations on self-driving cars, starting to make it more difficult and limit what they can do and slowing down the whole industry and letting other countries get ahead, or letting the whole revolution stop before it even gets going. So that’s my worry. [laughs]
MJ: Yeah. Do you think some of that is a failure of imagination on the part of the public? I mean back in ‘94/’95, the internet had sort of a wave of publicity behind it. And I know when I talk to people about drones, for example, an awful lot of people, their reaction to drones seems to be like, “I don’t want those around me. I’ll shoot it down!” That’s sort of the gut reaction that people have, where it’s not the “Oh, that’s cool!” like the internet. I have the “That’s cool!” reaction. It’s just weird to me that the contrast, I guess, it seems like a lot fewer people have that “Wow, it’s cool!” reaction.
JR: You know what people said 20 years ago? You know what people were telling me, the average person in ‘94/’95? “Isn’t that CB radio?” [laughs]
JR: It was like, “The internet is CB radio. It’s just a waste of time. This is for stupid people who want to chat about nothing.” That’s the kind of stuff that they thought the internet was all about. You get the same kind of thing now, you know, “Oh, I don’t like the drones,” and they see drones only for the picture-taking rather than the kind of transportation delivery system they could be. I also think that people don’t see a lot of the benefits that could have happened because of the internet, so they’re a little less trusting of big revolutions, in terms of what they could actually provide us. What made the US so special over time, over all these years, and why we keep on leading the world into these new technological revolutions is that the US had a large financially stable middle class or financially prosperous middle class that’s willing to spend money on technology to improve their lives and the lives of the people around them. They trusted it. It’s called kind of an “Edenic movement” or “Edenic improvement,” you know, trying to return to Eden, from the early Christian origins of the idea. We’ve lost a lot of that. There aren’t as many prosperous households and they’re on the wane in the US. Median incomes and household incomes are squeezed and there’s less disposable than ever before, and people aren’t being treated fairly like they used to, or with the expectation of being treated fairly and with respect, and all the benefits seem to be accruing to a few. So, maybe we’re just too jaded now, too. That could be a problem, is that we’re looking at these new technologies and going, “Well, it’s not going to benefit us like it could. You know, we get some benefit, we get a lot of free services, but we’re not feeling like we’re moving forward as a result of using it.” I do, but I don’t think that’s as widely shared as it should be. Even though, you know--you ever hear Bruce Sterling talk about smartphones?
MJ: Yes, I have.
JR: “Poor people love their smartphones. It’s their most favorite and prized possession.” So, I guess I mean there is a certain amount of value that people attribute to it, but they don’t think of it necessarily in terms of internet maybe, or that technological change made that possible for them.
MJ: So, for people that are wanting to read more about how you approach these issues, or people that just want to find you online, where would they do that? Where’s the best place to do that?
JR: You know, I’m going to try to put more stuff on Global Guerrillas, and I have a Global Guerrillas Facebook site, too. I don’t mind Facebook. It’s not as good as it used to be because people are just putting a bunch of links on it now. But my Global Guerrillas site on Facebook is not so bad. And then Twitter I put a lot of little bits of thinking on there. I’ve been kind of heads-down in the military space for a year, thinking up kind of what is the equivalent of blitzkrieg warfare, kind of this combined arms blitzkrieg warfare using robotics and autonomous systems, and I think I figured it out. [laughs] But I haven’t been writing about that publicly at all. I might start doing that on Global Guerrillas, my blog. So if people are interested in that and a lot of these changes and how things are going, I’m more than happy to have them read, subscribe, get hooked up to that free email that I have on there. I haven’t really done much of it over the last couple years, but I think I’m going to do more now that I’m heads-up.
MJ: Okay. Yeah, we’ll put links to all of that in the show notes. And I would say we’d love to have you back to talk about robotic warfare. That’s a topic Matt and I are both interested in. Matt’s former military as well.
JR: Oh, cool. Are you from the Army or the Air Force, or…?
MJ: Marine Corps.
JR: Marines? [laughs] Yeah, no, some of the more accepting folks, in terms of the guys I was working with, were out of the Marine Corps. They’re kind of leading the charge. Army’s pretty good, and then the Air Force had their own kind of push that was a little bit different than I was expecting. But yeah, it’s all changing, man. It’s all changing for the better, I think. Anyway, yeah, I’d be more than happy to talk about it. Let me get some more stuff—if you see any article I pop up that supports a conversation on this, let me know and we’ll do it again.
MJ: Yeah, we’ll definitely stay in touch. Thanks so much for joining us tonight, John.
JR: Hey, my pleasure. My pleasure.
A: That’s all for this episode of Robot Overlordz. Are you interested in the future and how society is changing? We’d love to have you join our community. Visit our website to learn more and to connect with others that share that interest. You can find us at RobotOverlordz.FM. The site includes all of the show’s old episodes along with complete transcripts, links to more information about the topics and guests in each episode, and our mailing list and forums. We’d also love to hear what you think about the show. You can review us on iTunes or email us.
A: We hope to see you again in the future…
MJ: Thanks everyone for listening.