tv The Communicators CES 2018 Technology Show Part 3 CSPAN February 12, 2018 8:01pm-9:01pm EST
>> up next, look at artificial intelligence and machine learning from the consumer electronics show. the senate is debating a number of immigration proposals this week, as mart of a deal by majority leader mitch mcconnell. we'll get an update in an hour. the white house released president trump's budget today. one of his priorities is infrastructure spending. later a conversation on election security.
[inaudible conversations] >> adelyn zhou, what do you do for a living. >> i'm the i work for topbots. we focus on art terrible intelligence and machine learning. we help executives figure out what the artificial intelligence is, this technology, and, two, how to actually use it and apply it within their businesses. >> host: how do you define artificial intelligence? >> great question. we define it as using computers or technology to reach human or beyond home levels of abilities, which it's automating different processes or different parts of our lives. >> give an example. >> so many.
i mean, you can have artificial intelligence in something as simple as your music play lift or your netflix queue. it can be in your internet e-mail system and filtering out spam. that automated system is not a person they're marking things that are spam or not but a computer algorithm using technologies like learning when a.i. to do that. and you can have artificial intelligence powering self-driving cars, using vision and learning to navigate busy streets, or my favorite and one area that most people are excited about is with health care and using artificial intelligence to help doctors that are diagnosed skin diseases or people's potential of getting heart disease.
>> can that artificial intelligence teach itself, can it go beyond what humans are capable of? >> not yet. in our book, we actually spend two chapters helping people understand what a.i. is. one of the most commonly misunderstood things and in the media right now there's so much actually fake news about what artificial intelligence is doing and can do. when you think about a.i. it's a spectrum of out mission. at the really basic hear you have really simpling what we call rule-based systems. if you're decision tree, it's to say, you say a., and then i will respond with b. just automation. it's not really a.i. on the other spectrum is what we call artificial general intelligence. that's where you hear about robots being smarter than humans, having super level intelligence capability. we're not there yet. we're kind of somewhere in the
middle right now. where artificial intelligence is concluding able to create, create music, create pictures and things like that on its own as well as a.i. being able to learn. so starting to learn things but it's definitely not there and can't have a conversation with, like, you are i can have and jump from, say, talking about sports one minute and then talking about the news and then talking about philosophy the next. >> saw one of the robotics company here had a big sign that said: sorry, but we can't take over for you. was that -- do people feel uncomfortable with a.i. >> i think there's a large -- a mix of unknowns. a lot of people are not -- first, a lot of people don't know what a.i. is, and we work with a lot of the leading, like, b level executives of major companies and this is kind of like this nebulous thing of what is a.i., what is it people don't
know. people don't know what it is, it becomes hard for them and you scatter getting scared. the hollywood movie industry doesn't do is any help by coming out with movies like terminator, and we're not there yet. we're many years to go. the exact number of years, varies. if you ask any expert it's anywhere from 30 to 40, 50 years to reach the terminator level, if at all. >> who founds top boston. >> we founded topbots with two colleagues, maria and marlene. >> holiday did you fund yourselves. >> we decided to actually be a real business and fund ourselves with clients. so we invested initially our own capital, but from then we decided that the best proof of a girl business is getting repeat clients and client work, and so we're using that.
>> what kind of business? >> the best size of business for us. we have been in venture businesses but wanted to be the real business so we have clients and we have -- >> it's three women. >> three women. so kind of unusual. >> what is the reception out in silicon valley? we have heard a lot in the news about women in silicon valley. >> the three of us have been in technology for most of our careers. i think for us we're used to it. even at cs, jumping occupy the plane and coming here, 80% men. you get used to it. you form partnerships with other women and other men who are very supportive of women, and you try to do the best work and be recognized for your work. >> now, adly zhou you co-oured to other book called applied artificial intelligence. what the theme. >> applied artificial intelligence is how you actually use a.i. today in your business.
people don't actually know how to use the technology and make my business grow. so we found you have really technical books that are so technical that you and i, most people, would not be able to comprehend, and then you have books that are thinking, what will happen when he robots take over for what happens with super intelligence, for a business leader there's no handbook on how to actually apply a.i. to my business. how do i create a strategy, and how do i as a executive get buy in from the other parts of the organization to focus on something other than quarterly returns and focus on investing in a technology that will change our company. >> we're going to put you on the spot here. c-span is a media company, we come to you and say, help us use a.i. >> yes. >> what do you tell us?
>> well so many opportunities for you guys. so, first of all, we look at where do you have -- like, one, what are your core strategic goals. what are you trying to achieve witch a.i.? there's no point in just using a technology for technology if you don't have a problem you're trying to solve or purpose that you're trying to reach. and so if you're trying to, for example, reach more customers and you want to understand how do i do that? that would be a marketing and sales question. so there's a lot of opportunities in that space. so, first, you would look at -- you can use a.i. to better find the type of customers or in your case, viewers recep tonight your programming could be looking -- finding correlations among different audience bases and finding out even though this person didn't a political junkie yet, hey has potentials based on
other interests and then use a.i. -- that's one way to find the audience. another way ious can create, for example, an facebook chat bot. might use it to deliver news to different -- to your constituents so they can chat with a virtual assistant. you're like, hey, what was the latest bill being put up for vote, and this automatic intelligent agent can say that's bill has this sport and this type of probability of being voted on. these are the other things to keep in mind. that can be thought of like a companion or advice or help viewers understand what is happening in the government right now. >> how much does that advice cost? >> it can be super simple. there are, like can products you can use outside in plug and play, and if you have the technology, sometimes is i it can even be free.
or you can spend a lot of money doing a high integrated system, but i don't think cost -- it depends on your scope. like any other technology, it can be anywhere from a coup thousand dollars to millions. really depends on what you're trying to do. >> adelyn zhou, what is your biggest concern and what is the biggest drawback in your view of a.i.? >> i think there's -- my biggest concern actually right now -- you talked about it -- the lack of diversity and creating these intelligence systems. so, fundamentally, machine learning is based on using a lot of data. so they look at data and find correlations and find help figure out and to create kind of their output. if you're underlying data is faulty or you collected dat in the wrong way, you can have
biased outcomes. then if the algorithms you use are not robust and are skewed or biased in certain ways, then your outcomes can also be incorrect. an example would be in the judicial system, there are now starting to use artificial intelligence to determine whether someone should have five year or ten-year sentence and using that and giving that to the judge and the judge looks at that to help make a criminal sentence. but if the data powering the likeliness of that candidate or person to repeat the crime, the recidivism rate, is based on faulty information, it's like saying paper's ethnicity or gender or age is more likely to repeat the crime, then the output of that would be an incorrect or faulty or biased number. so a person who maybe shouldn't have been given a ten-year sentence, should only be
five-year sentence, got a ten-ee sentence instead. so depends on the people creating the algorithms and the data that this is not biased so you take into account all the different nuances of our everyday life. >> what about privacy? >> privacy. i think that's interesting because it changes across the -- internationally. in europe it's different and in china, everything goes. for us, as individuals, we do need to take into it consideration, how much private information we are putting out there. i think a lot of our data is being captured and collected, and these days data is considered the new currency within extreme learning, so with our own privacy we should be careful and guard it and also be cognizant and circumspect of different companies that are using our data and asking for the data and seeing they're using it in a potentially
justified way. >> you feel that the data protections offered today are the -- in the regulatory framework should be stronger, weak center do you have an opinion? >> i think it's hard to say. i think it's really depending on the american public and what we want. i think the thing is, even of you have prior, stronger regular layings, people never read the fine print and people are willing to give up their information for convince, and so even with higher regulation, i think it's very difficult if the public doesn't really care and that they want convenience but there should be rules and regulations in place to make sure the data is protected, that it's not -- it's being safeguarded and not being hacked and used, because our data is becoming our fingerprint and it's our individual identities and we do need to protect that.
>> do you find, especially in the -- where maybe an older generation -- do you find different attitude toward technology than with the younger generation. >> they don't necessarily understand a.i. but most people we talk, to smooth be -- might be a biased group -- they're interested and want to figure out how their business can use the technologies. i think they realize that with google and amazon and all these different companies that their company in order to stay afloat and to -- it needs to embody these technologies. the problem is that they have quarterly financial goals they need to hit and yet they're trying to invest. so they constantly have this i guess dilemma because these investments in a.i. might not happen and drive throughure
quarterly return for another quarter. >> you're a hard mba or undergrad. is that's veeroff for you? >> i've always been interested in technology and how technology impacts our lives. identity been start of startups and tech firms my entire career, so for me this is just the next phase. i love and what gets me up at night is helping me understand how these fundamental technologies can be used and applied in our life today. i'm trying to translate between the soup -- super technical work and try to make it accessible to the everyday business leader so they can use the technology and improve our businesses. >> as someone in this business, what technology do you you regularly and how do you safeguard your own privacy?
>> first, i have a sticker above my camera on my computer so if you're thinking about that, i highly recommend people to -- >> something that simple. >> yes. a lot of times when you see very easily hackers can access your camera remotely. even the pope on his ipad has a sticker that hides their camera. yes. >> one facing you. >> the one facing you, yes. that is very simple. on the privacy side there but in terms of just making our everyday lives easier, one thing that a lot of people are starting to use is speech to text. if you love texting your friends and family, it's actually so much easier to use speech to text and the technology, a.i., can do natural language transcription and it's gotten so good that sometimes you can use that to tex people instead. so that could be something simple you can try and use a.i.
in a way to do that. on the workfront, we use different processes to out made our systems, our -- when we work with clients and things like that. so we use different technologies within that. >> do you find when you work with nontechnology companies that the understanding level is lower? >> absolutely. i think that it is a gradual progression and people are getting smarter about the technologies. if you're not interacting every day with it, you're like, where do i enstart? that what's fundamental goal with writing this book, to help give people an access point, way to start to really understand people who are not necessarily swimming in technology every day to understand what it a.i. and what are the ways it can impact their lives and businesses. >> applied artificial intelligence can be read by the general reader. >> definitely.
>> adelyn -- >> for the general reader and hopefully have framework and strategies they can actually use. >> adelyn zhou, thank you for your time. >> thank you very minute. >> this is the consumer electronics show in las vegas. more from our visit coming up. [inaudible conversations] [inaudible conversations]
[inaudible conversations] >> now on "the communicators" we want to introduce you to deep pew talla who worked for a company called nvidia. >> we started 25 years ago in 1993. started as a graphics and gaming company. if you're a gamer, a pc gamer, over 200 million pc gamers now. and ten years ago we decided to do the same thing which gaming technology, god the gpu. gpu is graphic processing unit, but the difference between a gpu and a cpu which is what powers our computers, is the cpu is a
serial processor, a gpu is a -- many things. but gpu have one processor, gpu have thousands or tens of thousands or millions of thousands. so, it's running many, many, many processes at the same time. we use the gp and basically ten years ago we started expanding beyond gaming into many computer -- for example, almost all of the top 500 computers are driven by gpu technology and the last two years, gone into many different markets, self-driving cars, and you probably see a lot of the modern a.i. evolution happening. we're right in the middle of that a.i. revolution. >> what's the generic definition of artificial intelligence and what your definition? >> i think artificial
intelligence is many ways to look at it. the way i look at it is in the end, it is a tool, just like electricity. so you can use it just like electricity. hundred years ago when electricity came about it was new, but every industry uses electricity. there's no industry that doesn't use electricity. right? so artificial intelligence is going to be the same thing. it's a tool that every industry is going to use. look at the last three years, almost every industry that we know is being transformed by artificial intelligence. so give you an example. think as a consumer, all of us use iphone or android phone, doing an google or siri or alexa. all of those are using artificial intelligence. when you request a queriry, the queriry is send from your mobile device to the cloud and then the artificial intelligence is running it and the virtual
assistant is sending you back the answer. that's how most consumers start enjoying or feeling artificial intelligence, since then it's being applied to every industry. self-driving cars is a big example here. smart homes. smart cities. robotics. medical emerging where a.i. can help detect images, process images faster than a radiologist can and basically augment as radiologist to basically get an idea of, okay, look at this and get a skins of the cancer is there or a problem there or not. >> so, in your title you are vice president and general manager of autonomous machines. what is an autonomous machine. >> any device that moves -- you can thick of an autonomous machine, like robots.
robots for manufacturing, used for inspecting bridges in rural areas, a video security camera that is going to be on a police car, for example, looking out for suspects or looking out for amber alerts. so anything that potentially -- a machine that is moving and if you think about it, all of them will be infused with a.i. >> so, today so far, have i used an nvidia proticket? >> most likely have indirectly. if you're using a siri or using netflix immigration engine or facebook pictures, most likely you're using -- you wouldn't know it but when you're trying to type in -- recognize something, it's going into the cloud, and then the gpus are now in every cloud service, whether
it's microsoft or am amazon, the united states but every cloud dat center is using the gpu technology. >> you say xr rated, what does that mean. >> a gpu always pairs one with a cpu. traditionally for 30-40 years, the cpu, think of it like -- kind of what companies -- every computer, processer in a cell phone. all of those have been all of your software run. now, what happened for the past 30 years, phenomenon called -- the people call it -- every year, the cpu used to get greater than -- every two years a little over two timed faster, and compound that by 30 years, you have this explosion of a
thousand times faster performance, and we are in the early '80s to where we are now it's obvious. now, that trend did not -- cannot last forever. so, what used to be changing every year in he last five years or so, it's barely 10% and you think that 10 and 50 is not that much but when you compound over many, many years, it's going to be factor of a thousand. so gpu came in and took that role of accelerating applications and cpu would stop. to gpu is always paired with a cbu and that's accelerating application. >> how close are we to a quantum computer. >> people are working in the industry on it, research stage so remains to see what the exact time quantum computing comp it right now we think gpu computing is a new form of computing. cpu for 30 years, and gpu
computing is the next wave of computing and then in the future we see. >> so, as vice president and general manager of enough individual a, what do you do in manage people or work in -- >> having fun. but like -- first of all we have an amazing family. 11,000 engineer, the best in the world at what we do, and aberdeen working towards a -- everybody work towards a common goal. work at products, defining products, in the working with engineers and then my responsibility is to take the products to market. so work with customers all over the world, so i travel quite a bit, going customers, and essentially make the products successful, and then work on the road map for the next one. >> deppu talla for the consumer, how transformational has been cloud computing and how transform national will be 35
g. >> i think cloud computing is amazing. think about the late '90s and 2000s. and then we saw cell phones and -- they successful because of cloud computing. without cloud computing you cannot have all the apps and programs running. mapping and visual assistant and other things, the credit going to the cloud and then the cloud is able to -- you don't have the device to be a super computer all the time because you have a finite amount of energy. the cloud has access to all that devices and can do a better process, centralized computing. because of the cloud, more together they form such an unbeatable pair, and then the
next evolution of 4 g today, 35 g would mean higher bandwidth and lower late ten and i better quality. and that's par for the course. things have to get faster and better because more data is coming. the videos, what used to be hd and now devices come onboard and then more cameras in the world, whether it's cameras on our phone or cameras placed for all these traffic management and foot traffic analysis and safety and security, and then self-driving cars, one or two cameras now and will have tens of cameras in the future. so so much data coming in. no surprise you need to have lead latency and faster bandit. >> you mentioned this but so much data coming in. does it get lost?
>> well, i think if you look at today, the answer is, almost all cases, a lot of the data is lost. give you an example. think about the security and smart cameras. you look at airports and office spaces, traffic intersections. ow have all these cameras and museums and government buildings. what do you think is happening to the cameras capturing data 365/24/7. after 30 days, depending on the policy, they a race it unless something bad happens. when something bad happens humans go and look at the video, which means that almost always after the fact. so you can't really prevent, can't do dig but as you add more and more cameras, with artificial intelligence you can augment the human being. nobody wants to witch hundreds of billions of cam razz 24/7.
it's the most boring thing to do. so artificial intelligence can come in the form of trebling nothing called deep learning, the knowledge revolution now. can come in and it can basically analyze video, is there an anomaly, movement, person not supposed to be in there? what is the track pattern looking like from 5:15'm to 6:00 p.m. in rush hour. is somebody breaking the law and do all sorts of analysis and then the human can come in and finish the job. that's what augmented men 0s. so the data today is lost but the hope and the goal is that during the part of a.i., you don't have to lose all the data. you can make sense out of the data. if it doesn't make sense, don't use it. that level of data, you can make sense of it and then use it for some value. >> what is the downside to all this technology and
connectivity? >> i think in general there's always thing -- with in good technology there's always the down side. you go back to -- used for bad things and bad things. electricity used for go things and bad things. any new technology, what is the limitations, people, and what does it do for jobs, going to take away but i think in general, the being at artificial intelligence that is different than all of those is a.i. is still human. a lot of people don't understand because they don't understand it and there is a little more fear, fear of the unknown. but what we're finding in all these markets a.i. is coming into, like self-driving cars, over a million fatalities every year. if we can reduce that significantly with the self-driving car, using
artificial intelligence, that's great benefit to mankind, and salve driving car is doing it, spending time, one hour two hours a day in the car driving? surely love to drive one in one week or whatever but sometimes just want to maybe be more productive and that portion of my life -- working life is given back to me. then if you have self-driving cars, parking spots and the whole resources in the whole city is going to get better. there's always benefits to this. and then -- so, then there's policy. so, the question then is, okay, what is the downside? i think it's up to all of us to figure out, policymakers, citizens to figure out the downside and figure out what the best way to handle it, come up with the right policies and ethics. in general, i'm an optimist. i think if we put our mines
together so many benefits and he we can find a way to come around those. >> we spoke with governor rick snyder of michigan, while we were here at the show, and he was talking about tech companies setting up offices in detroit because of the -- has nvidia set up an office there? >> we worked with detroit and many makers, over 300 partners of the drive platform, from cammakers to trucking companies, to transportation companies. googles of the world -- uber-s of the world, lyfts of the world, mapping companies. think about it at the scale we're going, map thing whole world notice that difficult. you can tell, not just using gps but using cameras to map the whole world and figure out at any given time, where you are.
so using that you can actually do the self-driving very well. and we have offices in michigan, detroit area. >> are you finding that the american work force is ready for your company to be employed by you? >> i think they are. think obviously it's evolving. i believe this whole revolution has just begun. three years ago -- people talked about a.i. before, deep learning. the average person did not know much about a.i. and it's not applied, why worry too much about? but the first deployment just started, like i don't know if you use siri or google 3 or -- using it now, you clearly see there's a quality in terms of it can really -- better ans than before. and that is just the beginning. still doesn't know the context of the question that you're
asking. it could be -- you could be asking the based on your calendar or something you have to do tomorrow. it has a human -- there's that level of thinking that a.i. can help. so i think the deployment side in the work force is fine because we're in the very early stages of a.i. deployment, and as a.i. is deployed more and more in the case i'm seeing, whether it's robotics, self-driving cars, almost all cases it's augmenting the human being. it's not quite -- that's why i think the work force is ready for it, because is the work force ready for a smartphone? yeah. without the smartphone we couldn't be productive. self-driving cars, going to get more productivity and the same thing with manufacturing and things about warehouse management. how many warehouses getting so
big, so tall, you don't want humans climbing those things, for example. so those things are fantastic example where they can come in and augment the human being. i give you the example of video. if you want to -- something bad happens, right? you want to find out the video of the traffic intersection. you can spend 24 hours trying come to the frame or ask a.i. to help you get to that faster, and then finally getting you interesting piece of human. i think these are all thing that humans alone can't do one more example. deliveries. three, four, five years ago, how many package wood you get to your home? maybe one a week two a week. that's all i used to get. now guy home every day and there's half a dozen packages lying around. my wife loves to order pangams. what are the chances that a human being would be able to do all those package deliveries and
and food deliveries and don't want cars and people. not enough people to do all those of kind of things. i look at is as especially in the near term, begin that a.i. is very early stages, it's almost all cases it's augmenting the hundred, making the human being more productive. then in the future, remains to be seen, always been true with any revolution, this revolution 100 years ago was agriculture, now i think it's transformation but i think right now in the near term it's been about augmenting human beings. >> we have seen where some tech executives have limited their children's time using tech, bill gates, steve jobs. do do you do that? >> i wish i can, but they don't listen to me.
obviously just amazing that three-year-old kids -- i have a five-year-old son, started playing with phones when he was 3 and loves banging on the internet switch and that thing. and then we limit in the sense that obviously as parents we look at one hours a day or two hours a day and then get them out and do other things. generally confirm that not just kids. us, how much time are we spending on platforms, looking at all the things. i personally try to be as -- listen to an extent and sometimes you -- >> are you fighting the kids -- when they're given this at early age, are easy adapters? >> oh, for sure. it is amazing. give you a classic example. artificial intelligence is so -- the most common person -- sounds like what isn't?
this past surgical in my lab we had 12 high schooler, 14-year-old, 15-year-old, 16-year-old kids in high school, most in the bay area. they came in and did a internship building robots and they were programming artificial intelligence. ten years ago, five years ago, i wouldn't think of a high school kid, undergrad kid would be programming a.i. but it's becoming to easy and so common with kid that high schoolers picking it up. i don't now if you heard after first robotics, high school students building robots as a hobby. or 5,000 clubs worldwide, and a couple of thousands just in the united states, all over the utah. different high schools, all have robotic clubs. so building robots as hobbiy and
then with a.i. they're using the technology to do cooler things. younger kid are able to take this technology and understand the technology as it -- >> deppu talla how do you view regulation in washington. >> depends on the industry. there's obviously self-driving cars, the case of california, we have self-driving cars and they're doing fantastic. actually permission for self-driving cars to do the testing because without the right regulation hard to take technology very far. in the case of self-driving cars there's a little bit going on. in the case of super computing g and the case of using eye a.i., we work with the u.s.
government, the nation's top disputer computers are xl rated and that is coming. >> what is your background? >> my background is i have a ph.d in computer engineering. >> at ut. >> university of texas at austin. i feel like it's a thousand yearsing. it's 2001. since then i've always been working with technology. like technology, i start as an engineer, designing, architecting processes for cams a and platforms for. my previous job was i worked on the smartphone revolution, and one day ten years ago i decided i would like to be in intersections of technology and business. so i moved into the business side of things and still am very close to technology because without technology it's very hard so that's what i do now, is work on the latest technologies and i can afford to have fun
because it's a fax family of engineers. >> host: where were raid. >> born here in indiana, and been here since i was 26. >> do you think your 2001 ph.d from the university of texas is relevant today? >> absolutely. i think there's -- the way i think about it is it's relevant for what did because what i did was computer engineering, so all the processes i worked on then, now are being transformed for the next generation, even if it didn't, it's -- a ph.d is three years after -- let's say, three years, four years, five. >> after a masters. we maybe 200 years when cry joe generallic -- cryogenics come in. it's training. just because it is training a certain area, it doesn't mean you have to work in that area
for the rest of your life. you have to continuously evolve but no matter what the degree is, just a way of thinking. that's the way i thought about it. that's the way how i followed my career. didn't matter what did or wanted to do, i can always adapt to a new industry and create the beauty. as long as your learning continues, you have the skillset. >> so, a couple years ago we talked about 3d printers here and very excited. we talk about drones in the past at this show. we were very excited. this year it's a.i., vr, those things. but in three, four years, what are we going to be talking about. >> i think in my -- it's always prediction are always going to be wrong so with that caveat, i think a.i. revolution in terms of given that a.i. is not an application, it's a feeling that goetz to be impacting every
industry issue think it's -- almost certain that a.i. influencing industry, is a given. a no-brainer. now, as to what extent it's going to be depending on each industry for several reasons, technology dependent. some technologies are easier than others. how fast are the regulatoriy or policies. those are important because just because you can do some technically doesn't mean you should do it. so which industry will good where, going to be dvd or to be decided. the self-driving car thing is going to be a great thing because it's going to impact humanity so much, reduce thing number of accidents, improving efficiency, doing all this -- transportation as a whole overall is going to be a fantastic thing. in the case of robotics, going to come in and augment for
whether it's auxiliary or drones but autonomous drones are going into bridge inspection or oil and natural gas, rig inspection. humans are expensive to look at those things. robots in agriculture are going to be huge. we need them. people coming onboard in the next 40 years, have to feed them and not going to be enough farmers coming around so robots need to do that. the kind of industries that a.i. infusion is going to be -- just no-brainer. >> deppu talla of nvidia. thank you for your time. >> thank you, peter. [inaudible conversations]
>> now joining us on "the communicators" is andrew shuman, vice president for product for the microsoft corporation. mr. shuman, what are those products we can loo forward to from microsoft. >> well, thank you. i work at running the cortana team and i've deeply involved in our effort to think about how this natural system can come about and change people's lives we're in the nascent steps of how people use computers. a.i. powered experience. >> cortana is a.i. >> i would say it's a.i., yes. >> is she learning as she goes? >> cortana is i think -- if you step back and think at a.i., on one level it's big data and you can understand pattern from data but a.i. is how software can be easier to use. that's where software learns more about natural language,
human speech, gestures, physical world around us, and in that way can be much more easy to use and really meets the user where they are. they don't have to learn about the software. the software learned how to work with them. in that way cortana embodies this idea of something that you can naturally communicate with and can know you really well. it also really knows the world really well because of the work we have done with bing and understanding boston is a place, and that's an important distinction for certain people. so being able to kind of have the full knowledge of the world alongside the natural language is the capability. >> how does she compare to alexa or siri or hey google? >> i think all of us probably compare in one interesting way in that it's very early. these devices and systems are pretty simple to use and they're awesome. when my hands are full in the kitchen it's awesome to set the timer that way or to get music playing that way. but what i believe, and i
believe is in unique thing about cortana, we can think but how we're better connected to thing that we be in about you. if you're using office at work, for example, you can use cortana at hope to do succeedling and manage you day, get in the morning and core khanna when die have to be at work or need to dial into a meeting and leverage core tan newscast. cortana knows you and more about you and your busy day and the projects you're working on and can connect you back to things you're working hard on, perhaps on your desk top is powerful. one small feature that shows this is we allow you to do reminders and they can be con tech to -- contextual. it i get to work and i'm in front of my pc with the power of microsoft word and office and the remind irpops up in context
and know. so being that ubiquitous assistant who has your back. >> even though microsoft is less than 30 years old or about 30 years old, we think of it as an old line company anymore. how do you think of microsoft? what is microsoft's vision? >> well, i think microsoft certainly has been around a while but has also continually transformed into different areas. i worked in the office group and now the natural language and bing group and cortana. you think about the sea chengs and how to think about how they impact everybody's ability to get more done and be aware of what is going on around them. think that constant reinvention is faction. i think it's brought a change in the culture by thinking more about how to go after emerging, growing areases of interest such
as a.i. and those areas get me up in the morning and get people excited. cortana has a lot of young people, too, on the team so a good mix, i think, of old hands and young hands. it's great. >> what is going to be the impact of 5g on microsoft products? >> great question. i think 5g is so interesting because it means we can have all of these devices be connected without any setup, without any work, and so the idea that all a of my device can be connected to the internet and available for us to do interesting cloud services and processing on them is fascinating. on an industrial side it's this stream of data that helps enterprises be able to run their companies better and know more about what is going on in their systems and then on a home level it's just being able to kind of manage your busy day. like, where are my kids? what time are they getting home? those things enabled by having sensors that are crequeed and
updating. that aspect of 5g is interesting. >> you talk about the devices as if, oh, yeah, i'll just remind cortana to tell us. what's the technology that goes into those? >> say that again. >> we talk about these devices very casually and we kind of expect them to respond to us, but how much technology is put into them? >> well, i mean, it's an enormous amount of work. to get natural language processing going has been a long, long journey. microsoft first invested in natural language and translation back in the early '90s when we founded microsoft research and we're building still on that knowledge and information. one aspect that we leverage a lot is the fact we have built the bing search engine which crawls the web of content on the internet and that gives augusts great understand offering natural language and natural language processing becomes a key engine that really has taken billions of dollars to build and
make -- that's i think the core part of the technology is being able to understand human language and be able to use that in a very natural way. >> with all the data that is being collected by microsoft, what happens assassinate where does it go? >> well, microsoft -- as one would expect, microsoft has a lot of very important enterprise customer and we're build in an era we have to treat data as the customer's data. it start width having very good systems by which people can control and manage their data. if you're an enterprise customer or an end user consumer. so that data is --s as a sure is one over this moe trusted provided are wassed a hear to local laws and have data in other countries and a lot of places where the data can't be looked at by microsoft employees and only by people that have
worked on the data. we work hard on keeping the data as an incredibly important asset for people to have all the right protections they expect through their own stuff. >> how long have you been with microsoft. >> 25 years. >> how didout yet? >> how did it get to microsoft? >> yes. >> he rainy enjoyed programming when i was a kid. id a had an apple ii and got hired to work on microsoft project called rn and systemy and that became the outlook project then i worked on outlook, and wrote the calendar in outlook, and now years later i'm working on artificial intelligence. >> andrew shoeman is vice president for products from the microsoft corporation and has been our guest on he "the communicators." [inaudible conversations]