Skip to main content

tv   The Communicators Mark Mills  CSPAN  August 27, 2018 8:00am-8:37am EDT

8:00 am
you're watching book tv on c-span two. television for serious readers. .. .. and i'm also a partner at a tech venture fund where we
8:01 am
invest in software companies for the oil and gas industry. >> host: how did you get there? >> guest: where? >> host: where you are now. >> guest: northwest university engineering school, the dean of a good friend, i didn't go to that school but i know a lot of professors. it's fantastic institution. i learn a lot hangout out with this goes. the manhattan institute has the ball. i began writing when i was in high school, some writing contest when as a young writer in fiction and nonfiction. always like to write so i have a pathological compulsion to write stuff. but the manhattan institute the ball because i write. they invited me to join them as a scholar some years ago. the venture work i write about tech as well. i was a practicing scientist, early in my career. i used to do useful work, built
8:02 am
things. semiconductors and missile guidance and fiber optics, medications systems. my early career was in building stuff and it went into nuclear weapons and strategic defense policy, reagan white house sides pop office. that led to consulting and investing again. one thing led to another. i wrote a newsletter. for while was very successful. migrated from that into doing some technology venture. long explanation because it was no simple vector like going to business school and getting a job. i learned business the hard way by failing to make things work. >> host: you written a book called "work in the age of robots." are you invested in robots first? >> guest: yes, but i did find robots, virtual robots, google and voice and amazon's alexa are
8:03 am
simply robots, artificial intelligence engines that respond, do things semi-autonomously. we do invest in both kinds of robots but in the oil and gas business, i'm not an investor in the hardware side of robots, cb three oh. yes if you want a dog in the race in that sense but i have been a longtime and scholar at investor in robots. now they're getting interesting because they're starting to become useful. that's both exciting and for a lot of people frightening because they think as most people know, the general trope out there is that robots are both virtual and real are about to take weight most work for most people and we're all going to be put out to past year like farms were in the 19th and 20th century. >> host: in your book you say
8:04 am
that currently there are about 2 million industrial robots in the world, and he is are they? >> guest: in the world. in automobile plants. the largest use of robots and, of course, these robots are using one or two arms only. you're not walking, talking. our automated welding machines, automated handling machines. the tool industry that dominates mechanical robot so far, car manufacturing and some electronic manufacturing but that comprises of about 15% of the manufacturing universe. most manufacturers don't have much automation yet here reason is most robots are not flexible enough. they can do tasks that are highly repetitive, easy to , assembly-line type. >> guest: exactly. a lot of manufactures don't like that and increasingly even mass production is becoming as
8:05 am
customization. it doesn't lend itself to this highly repetitive after robots are not so good at that get. they are starting to get better at that. the most exciting robot are called -- attempted by northwet university professors which sounds like what is, a collaborating, cooperating robot works with you. this allows you to do more and robot that could do on its own. some are racist. in the sense alexa is a cobots. alexa or siri can do things but it needs you. it's a collaboration. that class of robots is most -- the da vinci surgical robot used for prostrate surgery and liver surgeries, noninvasive micro intrusions. it doesn't operate by itself. the surgeon has his hands in and
8:06 am
sees what's going on up close but the precision that is affected by the robot is an application of the human. it doesn't replace them. there's no clear path, by the way, to easily replace the human. it's a bit like,, it's a bit like all a car and artificial portal and airplane and artificial bird. both serve a similar function but they amplify that function in a way that a horse doesn't by itself. robots are the same thing. they're not good at replacing people in more than we walk anymore. a lot of us walk a lot less but we wouldn't walk to china. we wouldn't sale to china. so the amplification factor the car for the shipper airplane made is what robots in fact, do over all. that's the key to the difference of the dystopian view.
8:07 am
add productivity, and wealth and add jobs. >> host: you also say in your book that, and use the car analogy, you say it is basically 1890 when it comes to development of robots. >> guest: i tried to think about the state would really at. you have to use -- analogize it in some fashion. the virtual robot, alexa and syria are not good enough yet. some very funny youtube videos. you should google it up as they say. very funny. they are good but not good enough. they are pretty good. in automotive terms artificial intelligence that's in the cloud, the virtual stuff are kind of like model t's. they are pretty good,
8:08 am
commercial, viable, useful but model t's are open to the air. they didn't have it in close. they're pretty uncomfortable and in 1918, a forecast of future ae the car and one of the forecasters that we would be fully enclosed and weatherproof. obviously model t's were not so great but it started the car age. that's what we are with artificial intelligence, the beginning. when it comes to c-3po, the general purpose what's known as an anthropomorphic robot or general purpose robot, we are essentially in the 1890s where the very first cars that were invented by diesel and by mercedes-benz and by several american firms, they were making very clunky cars. they were cars. they cost two or $300 -- thousand dollars. they sold like 500 a year. then a few years later they sold a few thousand. then you get may be a million of
8:09 am
them but it's not like hundreds of millions. in the robot one when it comes to the general purpose robot it's like 1899. we haven't got a model to get. you can't buy a model t. you can't name a robot except in science fiction that could walk around to do anything except controlled by person in a laboratory. those are great and youtube videos but you can't buy one. they don't exist a as a product. it's difficult to make those things. >> host: where are these robots we have today being manufactured? >> guest: the united states, germany and japan are three liters in robotics in general. in artificial intelligence the virtual robots, i have to say where the leader. we dominate the world. look at products like siri and google but similar products to help plan trips, these are typically use, silicon valley and other domains and the united states. the physical robots, c-3po again, we are pretty good at it
8:10 am
but germany and japan are comparably good in some respects, better. japan is a true leader. they have already begun to test a commercial healthcare robot which will solve problems. if people are or older have to be picked up and moved here people do that. it takes two people which is very unproductive way to use nursing talent to pick up a person. sometimes more to do it carefully. it's more productive if a robot that was compliant and had just one task that could help a nurse or the doctor take you or me up and from garney to betty, those are commercialized, very specific purpose but if you look at that is to replace nurses to amplify the nurse or doctor. it reduces healthcare cost because the nurses doing other things more valuable than being brute force. >> host: 2 million more robots
8:11 am
coming online in the next decade, industrial robots. try to probably 20 million more service robots are coming online in the next few years. literally. these so far do very specific in their services like the healthcare robot. they vacuum. the most popular robots sold by volume is like the vacuum pick the second most common robot sold in volume today is the robo lawnmower which is becoming popular. mostly among men, unsurprisingly. maybe you could say the vacuum is popular by women. the third robot which is a robot, autonomous drones, are now coming up in significant class a robot because if i deliver a package, by the sediment amazon in the city, that's not happening yet, but in rural areas and in africa that are now commercial deliveries of
8:12 am
medicines by robots. they fly. they are not walking deliver the medicine to an outpost and fly back and get another supply. far more effective than try to have a present of a truck. much more effective than a robot walking obviously. those are powerful amplifiers. that market is growing faster than industrial robots because it's easy. i cannot the path i described come for easier to engineer an industrial environment where manufacturing the complex expensive things. they're both starting to grow rapidly. if you're an investor, for example, this is an exciting place to be because yes, depending on which class you're picking turbine circa 1918. if you buy fort stockton might not have been googled by 2008 but judd and a 90 year run, not bad. artificial intelligence what it is so popular right now is because they have correctly figured out that is going to go a lot, it's early days. it would get a lot better. it would get a lot less sort of
8:13 am
annoying. right now i think most of us would agree the utility, the value, sometimes it works well but often it is noise. they don't quite work as well as they want yet. as powerful as computers are there still not good enough. still not fast enough. that trajectory still happening, still getting much more powerful computers year-by-year. we are clearly in sight of computing what i call useful. useful computing is i don't have to figure how to install stuff. i've had to figure how to ask a question. it's a natural and the computer in the cloud can naturally talk to me and asked me did you mean, is that which meant? like a person would. isn't that scary works for some people that's scary. i think millennials it is probably not. maybe for my generation, people are nervous about it but as i think as they get effective and helpless in health and economy i think people will adopt. if you don't about them there will be no market.
8:14 am
>> host: what is exactly the role of ai when it comes to robots? >> guest: so artificial intelligence robots are basically the same thing in our mind we have artificial dividing line. a washing machine is technically a robot. an automatic washer -- >> host: not including that in -- trip to the kind of talking about our were a welding or machining tasks can be done autonomously, a small drone aircraft can take a package and deliver to someplace and bring it back. or where i could say to my phone or two of alexa, i'm thinking about flying to las vegas next week, what are the flights look like? i don't want to tell you the day because i want the robot, this s is a robot smart enough to know to remind me mark, you were supposed to have lunch on saturday with so-and-so, you got a trip planned with her children this date. what da they do you really wanto apply? that doesn't exist yet.
8:15 am
that would be useful. >> host: you talk about a concept called ambient computing, which is what? >> guest: site don't take credit for creating the phrase. it goes back 50 years to a professor at mit, and he imagined that 50 years later, this is a pretty impressive guy thinking half-century of that computing will get cheap enough and effective enough that it would basically disappear into the background of our general life and be in the air. by that. that doesn't everything becomes automatic. it means it's easy to use as electricity. you don't think about electricity except when it doesn't work. you don't think about how it is delivered. it's just easy, always available. electrification is an ambient feature of our modern society. computing is done and get future yet. it's still too awkward. we moving towards a bit computing whether it be in a
8:16 am
constructive sort of useful way the background. >> host: we do have quite a bit of ambient computing, don't we today? >> guest: its ambient , let's go online, look at her phone tried to i would it is readily available computing but i still have to take something out of my pocket. i still had to go to a desk. i still have to have specifically addressing that. i have to ask it in the precise way because it ask siri in the wrong place which of course makes a lot of youtube fun, that's not useful. the ambience is a feature when it becomes just intuitive and natural. the computing and that's to me. i don't have to adapt to it. >> host: does murphy's law apply to computing, to robots? >> guest: murphy always wins. murphy's law always applies of course. there's always consequences. some really good. i would see history shows mostly
8:17 am
undetected outcomes are good but there are unintended outcomes which are not good which we have to guard against both behaviorally and sometimes legally, laws and regulations. sometimes as parents that we deal with our children having access, when there were no cars we did worry about kids being hurt and injured driving the car when you learn to drive. we have a culture that is unique in all it all of human history e last century. you teach children how to drive a one-time thing that can kill people. this is pretty scary. if you talk to people 300 utica, they wouldn't let 16-year-olds drive rent and something that was a time and to kill things. that require a cultural adaptation. we will go through a similar class of cultural adaptation to this new class of debating. before it happens, the dystopias are often murphy will have some things that will not be so pleasant and that's what will trigger both the regulatory legal and cultural adaptations.
8:18 am
how do we deal with these new tools? >> host: has the regulatory regime affected robots yet? >> guest: it has because the first thing that the faa did was banned drones from being used on -- they recertify them and give you some boundaries. you can apply drone out of eyesight without special license. the regulators, he'd use the word but reared their ugly head to speak. they have because you're to be sure things are done safely. when every new title to the comes along, regulations make sense onto basis. one, you want to have common standards, that form of regulation whether it's through common platform. a perfect example be if every electrical appliances a different kind of plug it would be hard to plug things in. cell phones at different difficult for long time. now they're only two standards. that regulatory environment is a
8:19 am
natural. intercedes -- the other, the safety one and the social ones are more challenging. sometimes regulators are behind the curve. sometimes the overreact. sometimes they don't react at all. i would say the overreacted on drones. drones of it difficult to make effective for the companies people are thinking about. and prematurely i think they went too far. under reacted in self driving cars because they're basically the technology is premature, way too dangerous and we learned that the hard way with some fatal accident. you will see regulators tighten the noose, i hope they don't tighten it too much. without long way from really effective self driving cars but to get to the have to be safe and that will require regulations. for think tank that is a free-market think tank, 74 against regulation treatment to against the wrong kind of regulations are over regulating. in those words there's a lot of
8:20 am
work and debate but those words have civilian. >> host: the title of your book is "work in the age of robots." for every new robot the comes online is that a job lost? >> guest: the short answer is no but there's a short answer yes in this way. always been the case we've try to advance machines to take labor at a task. we've been doing it forever. we would all be living on 40 acres and a mule. you got to take labor out of things to lower the cost. what that means is when you take labor out of something if you didn't own that task and just that tepco the person that did the labor -- because there's more of those tasks done you end up with either no decrease or ask an increase. that doesn't some people don't get displays.
8:21 am
the essence of the moral challenge of automation, of industrialization has always been, not that you don't get more jobs and automation. you always get more. it's a tiny percentage who are disgraced, they are literally fairly displays. they have a moral and not illegal, a moral and social obligation to minimize that to help people who are displaced. that's always been the case. a minority that are displaced, most people benefit. majority get huge benefit. so those benefit somehow went to pick up how to help people. there are ways to do that. in fact, the way to do it, we now have the tools that are taking the jobs away other tools that can help people. social media, automation allow the kind of retraining and align it with jobs and tasks they could have done te ten or 20 yes ago. if you lose a job, a platform
8:22 am
can make it easy to find a job with the company. very hard tinges go. much easier now. we training online much easier and cheaper than it was 20 years ago. the first time in human history the tools that are taking work away from some people will make it easier to retrain and get new work for the same people. >> host: you take an alternative view of the need for stem jobs. >> guest: this notion of the wood has to be a coder. this rush for kids and seems to code, learn code. you will not get a job. this is nutty. first of all it's a statistical fact that are more farmers than there are coders in america. only 1% of the population actually farms. of all jobs in america only 8% are in s.t.e.m. fields. 8%. not ten. there are not 10% jobs in america intact general.
8:23 am
besides technology, engineering and math. we don't need everybody to be in today. we need people to understand. i worked as an engineer. i think it's great to have great engineers and computer science. we don't need a lot of them. we have this whole revolution started with not even 8% of the workforce. if you have a child that loves science and math, great, cardigan. don't discourage of the kids thinking they can't get a job if they're not a coder or indigent. that's nutty. hithat doesn't mean you think would in a lousy job of teaching science and injuring and math to nonscientists. we have to step up. we are in terrible shape.
8:24 am
i think the future pension issues like the future farming. engineering are so good to make things to do the task better that we will a few a few inches the future, not more. >> host: what's going to be defect in your view of 5g on this robot revolution? >> guest: so 5g is an insiders, 5g is the next generation of high-speed wireless connectivity to your cell phone, which should be online at the end of year in certain locations? >> guest: most people haven't even gotten the 4g it which allows you, if your 4g or lte can stream video. i.t. will let you get high death streaming video. it's like fiber optic without wires. it's very important for robotics everything to do with automation. the most difficult thing with
8:25 am
respect to a robot is it needs more computing power than it can carry. it's going to have to have computing power about it, parks and buildings underground and connected wirelessly to that to get that information to be updating. the brain will be in the cloud. very, very high-speed to make that work. only 5g makes that possible. without 5g there's no self driving cars. there's no c-3po without 5g. >> host: so at what point is the basic structure of the internet, in your view? is acquainted to change or is a changing? >> guest: it's going to have to change but it's changing out. for george and i, we share common good of the state of the net where it's going. we are sort of in, i don't know, i would say 3.2. 1.0 was internet dialer. internet 2. 2.0 is what we have. three today which is what pretty
8:26 am
high-speed, pretty nice, you can shop online. three is the cloud we are today. which gives you alexis, siri. internet 3.0 connect to the next generation of computing and it can't do it for right of reasons. not the least is to george's point and it's been my for use, not least because of the security challenges that we have to solve. the security challenge having secured communication securing my personal information, but they give permission, my transactions, controlling all my data come all these things people now care about. that requires a new architecture. the new architecture, it would be like saying trains, it would be hard to take a train to the shopping mall. the car added a new infrastructure. we still have trains. many training miles today as 15, 100 scott scott stop moving around because they move a lot
8:27 am
of stuff by train. they move people by car. as the economy goes the internet we have today will be just as big and bigger but we will later on internet 4.0 which will be different, puts to the edges, highly secure, magically secure. all these transaction. imagine i have biological data about myself which one will a -- be able to have soon because i can eat a computer a day with my vitamin pill. i want that data from a doctor when you go to the hospital. i don't want them to know when getting sick. i don't want to know. how do i secure that? we don't know, we can't do that because the data, the volumes are too big. security is an on feature. if i tell you this it will make sense. it seems like weightless and visible compared to physical security. at the want to secure money to put in something they can have it a safe.
8:28 am
locks and walls. i have to build physical infrastructure to get sicker to come gates fences. thexact same thing in cyberspac. it involves hardware and power and overhead. every time i have hard work our innovative to communication systems you don't see it but it is real. it slows everything down. just like any security does. no one wants a communication network to go slower so it will have to get ten times faster get, your 5g. ten times easyjet and then i can stick inside that faster, easier system of burton because now it's not a question of burton of security. security costs hardware and cost time. always. there's nothing magical about cyberspace. the same features as physical space but you just can't see. the reason you don't have it is because it's so dark part of it if it were easy we would've done it. the reason of coders talk about and write apps, they pay no attention to security. the truth is they don't.
8:29 am
the reason is because it's hard. it takes up, so far overhead which costs money. things that are supposed to be fast becomes so. things that are supposed be easy become difficult. anybody will tell you to want to do, go and make your current system more secure to two-factor authentication what does that mean? it meets all of us have expensive. i tried to log into my account online and the robot on the other side which is a virtual robot says to me i'm going to text to you at this phone number, right? and you have to put that number and a second channel outside of computer channel, cell phone channel, i can do a three factor authentication to i can say to my banking transactions. you can do this today. know about a money can delete my account over, economic, $1000. unless i am -- [inaudible] third factor authentication or get the phone one of three people. my wife, me, my dog.
8:30 am
that can be done today. think of how much boom, pay. i want three factor authentication real-time. technically doable, difficult to do with the faster networks, faster computing. that the architecture george's writing about in his new book. life after google which i'm promoting george, he's a great writer, a great thinker but he's right. that's what we are going. that's what will happen. it's that because it has to happen. they can happen and everybody wants it. it's not like the government have to say i want more security. i want more security that is easy. that's what this does. >> host: so mark mills, this is a thin book tragic it is. the publisher has a series that they do to push ideas out where
8:31 am
their proposition is if you write it, people read it. as all writers know, writing a short about is in some respects hard because you're trying, you're forced to compression ideas, can't have a lot of sloppy extra information or extra words. it makes your case, make it briefly. it was a good challenge. i like the challenge. i hope i met the challenge of making the case that robots are coming both virtual and real, and they are way overdue and we need them because we enough productivity deficit in america. and they are great thing. one of the most important and biggest economic revolutions since the beginning of industrial age. it's not industry 4.0 which is a nomenclature. it's more than that. the industrial revolution was about transportation, about culture, acids were built.
8:32 am
how we could entertain ourselves. ourselves. all those features changed when hundred 50 years ago. we are changing them again for the first time in 1000 years on what you include a survey saying tech firms are thought to be ethical. >> guest: it's interesting, the surveys were surprisingly and encouraging. so not offices are ethical. businesses are like people. most businesses are, in fact, ethical especially in societies. but people seem to think that tech firms are particularly attend to be ethical firms. i think it might be true. visages, i don't know. if they do, and this is out in my book, their moral responsibility is not to accelerate this revolution that is coming. they will do the anyway. their moral and ethical responsibility is to deal with
8:33 am
the murphy's law effects can be edited to consequences people who are displaced who are not the majority, the minority. i don't want the government doing that. i want those companies to think about what damage they are causing. we should not be doing what karl marx what about the beginning of industrial pollution, we propelled the economy. no one denies the economy grew magnificently there was a lot of damage because of the businesses were not in tune to the side effects. not just environmental side effects. social side effects. i think tech firms probably are more attuned to that. which means they need to step up, do something about. which is not lip service. you have to think about what's happening to my point, that if your peace of artificial intelligence is displacing worker, what process is limited?
8:34 am
spreadsheets limited huge accounting officers. those people got just that their people today whose jobs ap displaced by ai that will get jobs. so create them. a social media platform help people get trained and get jobs. they can make it something that is a feature, not a bug, your publisher is encounter and encounter publishes conservative books. is this a political issue you're writing about? >> guest: it involves policy and politics so yes. i'd like to think it's not a political issue in the sense of left-right, conservative liberal. liberal. there some alignment on this i think. you do see some of the parties taking position you would call political on the technology but so far, interesting, i'm not, i don't think we can avoid things becoming political so far this issue is apolitical. a political part of it is what do you do about it. so you have some people who say
8:35 am
it peoples work is being displaced by automation, we, the liberal side of it i will say we should give them a universal basic income. we should give them an income. if we displaced them with the robot we should give them an income. i disagree with a profoundly. the conservative side was a if we displaced somebody's job we should give them an income. we should figure a out the markt functions to do something to make sure that person has help. we should still have a safety net which is what an point insurance is. no one thinks you should get rid of those things bu but i don't t to add new features. the reason why is i know we don't need them. in 1961 president jack kennedy gave a speech to congress and is at the most important issue of our day i is the loss of work in detroit from automation. 1961. he was right. detroit was expanding its production that time and laying people off simultaneously because that's when automation began on the assembly line. he talked about universal basic
8:36 am
income in 1961. 1961. so did president johnson after that. they could just they didn't get to put in place but we did create things like unemployment insurance and those kinds of programs. >> host: mark mills is the author. the book is called "work in the age of robots." thank you. >> guest: thank you. >> c-span, where history unfolded daily. in 1979 c-span was created as a public service by america's cable-television companies and today we continue to bring you unfiltered coverage of congress, the white house, the supreme court and public policy events in washington, d.c. and around the country. c-span's brought to you by your cable or satellite provider. >> c-span2 is in canada with the canadian house of commons citizenship and immigration committee held a series of

34 Views

info Stream Only

Uploaded by TV Archive on