week taking you inside the cashierless stores that are popping up in san francisco. we talk to an award-winning physicist about quantum computing and try to get into the brain of self-driving cars. our reporters from m.i.t. technology review martin giles and marybell lopez this week on "frontier." ♪ ♪ good morning, everyone. i'm scott mcgrew. in the history of my life this is going to be an artifact. it is a roll of tums.

take it off the shelf and no cashier and the store knows who i am, what i collected and the cost of the item is charged to my account. well, i feel fine. it was small and really inexpensive and i wanted to try out the content. i'm not sure my future grandchildren will be impressed. this may not turn out to be a family heirloom because in the future most stores will probably be cashierless and it will be krishna matakuri. he runs the technology replacing checkout and martin giles of m.i.t. technology review. i was amazed. it felt a little like shoplifting. i'm sure most people have that experience, as well. how does it work? how did the machine know i had taken it off the shelf and left? >> let me answer that question first by telling you how it all started. this was about four years ago

when my wife and i just moved back from south africa after a stint there and she -- we just had our baby boy and she asked me to pick up a gallon of milk from the store and i asked to go into the grocery store and i looked in and easily five-minute wait and there was no way i was going to wait for a gallon of milk and i went to the convenience store which my wife was not happy about. the store had so many gallons of milk and here was a willing customer because of the one bottleneck, the checkout line. >> right. >> so i did some research and found out that americans spend over 100 hours a year. >> oh, you're talking to three americans who stand in line. we get it. >> exactly. how did it know? i mean, this is tiny. i picked it up off the shelf and i walked out with it and i had to scan an app. >> how did it know it even came off the shelf? >> what happens is we have cameras overhead that track

exactly where each customer is once they checked into the store. you're just checking in at the entrance and the cameras actually follow you like a dot on a map like a gps and when you grab on to something the cameras can see that your hand went into the shelf, picked up an idea and that information is actually captured with cameras and also sensors on the shelf. so that both of those agree and that's how we know that you picked that item and we add it to your virtual cart. >> what about the challenges when i think of this type of store, you have to put sensors and cameras everywhere. is this the type of thing that we expect to see in a walmart or a super target? are we going to be boundz to smaller, physical stores with fewer choices? >> even the be ally you can expect it in stores of all sizes. right now the cost of deploying these sensors is dependent on the size of the store. so the larger the store you would have a higher up-front cost and it's a lower barrier to

entry. ten to 15 years i would expect every single store to have this technology. >> no customer would say i would love to spend another five minutes in line. >> in the closet of the checkout clock. is that job well and truly on its way out? we've had self-checkout for a long while, but there are still lots of props around. is this it that technology is going to kill them? >> we're building checkout-free technology, and not person-free technology. shopping is more than just checkout. it's about discovery and finding new products. so what i expect to see is everybody who is spending hours just scanning products would actually upgrade to other kinds of jobs in the store of the future. i want into the store and i want to know how you can make the moroccan, and i don't have this ingredient, and i can't find anybody to help me with that today. so on. >> how are you going to deal with the market education problem? >> for example, someone has to

download an app and they have to go to the store and feel confident that when they put something up and put it back you're not going to charge them. >> they make it really easy for them to get their customers included. so the app is not necessary. so you could use other forms identification. >> the app is necessary currently. >> right. >> obviously, as with almost everything in retail, the gorilla here is amazon. they opened up the store on california street in san francisco. you used to be, i would point that out. i would assume you plan to sell this offer to everyone on amazon. >> absolutely. >> is amazon, you suspect they will also try to sell their technology or they will keep their stores just to themselves holding on the secret to the software. >> it's hard to say what amazon will do, but from there, it seems like they are launching their own stores and it's what

they've done so far. we are among the first companies to launch a working store in san francisco and we are inviting the experience so they don't have to wait for another company to open their technology. >> could you imagine this rolling out to the gap or 7-eleven or any store using your technology. >> that's correct. >> i guess it would mean speculate and there's no expert on casherless technology. do you think amazon will compete with you to get the gap as a customer or do you think amazon will guard this jealously because we have this technology that we don't want anyone else to see so we believe that for the foreseeable future amazon will see this as a strategic advantage for other retailers and given that they own a fairly large chain of retail with whole foods, they have a significant opportunity to take this technology to their stores first. eventually what they might do depends on how the environment

was, and what about the privacy act? because you're tracking that you're monitoring my every move. >> literally. >> literally, and you know, big tech companies haven't been the best kind of custodians of our data up-to-date and now here is a whole new range of things that are very -- and maybe i picked it up and put it back. >> you know, what will happen to my information? >> we'll have to make that the last answer, but go ahead. >> one of the things with the solution is to make it privacy friendly. the cameras don't really see your face. they can't do face recognition and for the most part, the video and aim knowledgery image is pr and it's very similar to the way online stores do today. when you click on a product they know and offline stores now have a platform that can compete with online stores with.

>> he is the ceo of zip in. i suspect we'll be talking to you a lot in the future as this continues to evolve. thanks for being with us. >> my pleasure. >> people talking about quantum computing in a way you and i will understand when "press: here" continues.

welcome back to "press: here," we'll talk about quantum computing and i thought to describe very simply what quantum computing is. one explanation involved a dead cat, but it turns out one of the best people to explain quantum computing is the prime minister of canada. a reporter famously said to the prime minister during a question and answer session, i don't expect you to understand quantum computing to which the prime minister said this. >> very simply, normal computers work by -- [ applause ] no, no, no. don't interrupt me. when you walk out of here, some of you will know far less about quantum computing, but most of you, normal computers work

there's power going through a power or not. it's one or a zero, they're binary systems. what quantum state allow for is much more complex information to be encoded into a single bit. regular computer bit is either a one or a zero, on or off. a quantum state can be much more complex than that because as we know, things can be both particel and wave at the same time and the uncertainty around quantum states allows us toen koedz more informati code more information about quantum computing. that's what's exciting about quantum computing. >> unfortunately, justin trudeau was not available to talk about quantum computing and we found someone who knows more about canada. he is here with the super cool quantum computers leads the way in the field. sadiqi has been tapped to head the laboratory effort to build a test bed for developing quantum

technologies. thanks for being with us this morning. >> how did the prime minister do? is that a pretty good explanation of product computing? >> think the prime minister is spot on. >> okay. just to bring it on home, one or zero is how we do it now. the qbit in a quantum computer can be both which means it can do all of the calculations at once. >> fair assessment? >> that's right. in fact, quantum mechanics told us that all of nature can exist in all different combinations at a different time. hence the cat that's always asleep and awake. >> a physics joke. >> right. if we take lots of qbits. in fact, even a modern number of qbits than the particles in the universe to describe the configurations. >> where are we on this? give me the timeline and we remember the jacquard loom which was the beginning of the computers. >> and that has been -- okay. and of course, leslie park. >> put me in the timeline with

quantum computing. we are in the loom there. >> we are in that era, scott and where we have the equivalent of some of these quantum computers where devices are coming together and there are some calculations they can do and we're figuring out what are other things we're good for and what's the final technology that will push us to make quantum a reality for everybody. >> and i hear people say, well, what you need to do with quantum computing. you either need to have super cool these things to colder than deep space or you need to fire lasers with the ion traps and it sounds really complicated and really difficult. is this something that's going to be a computer on scott's desk in five, ten years or is it always going to be something for very big companies. >> for all that we know now, martin, these are specialized machines to solve all those problems that your classical computers have no hope of tackle, but that being said

we've only scratched the surface about how they can enter the machines that we have around us. so the best of quantum is still yet to come. >> let me ask you a question about that, because assuming they're not going to be on my desk, we also have another challenge. if quantum computing comes out it breaks security as we know it so how do we rationalize the viewpoint that we can't be everywhere, but we need it for security. >> it is always a development on both sides of the fence and our computers are getting better and so are algorithms to protect the privacy and security in the modern internet age. >> i'm not too worried about that and i'm sure we'll keep them in check as we go along. >> is there a possibility that the nsa has a computer that we don't know about? >> they would kill for a quantum computer, right? because they can break those cryptography. is there a possibility of you being one of the world's leading experts in quantum computing that the government has something that's better than what you've got?

>> well, scott, i can't confirm or deny any information about quantum computers. >> that's a yes. >> ult are you would ha >> you would have built it, right? >> i can only tell you what's on my mind. >> is that in the lab. >> we have an effort on camp us that's an academic activity and we're moving beyond what can be done in a simple academic laboratory and it's an opportunity for scientists to get together in a facility that looks like a particle accelerator and collider. >> and this will be at lawrence liver more where you can do what? [ laughter ] >> at least it will not be stanford. [ laughter ] >> no, i think it's a tremendous opportunity for us, scott to think about where quantum technologies can go in a partnership between government, industry and in fact, academia looking at these things on a large scale and giving a home

for the next generation of scientists that are interested in this field. >> my colleagues say they're all about the bad news. oh, well. what about the good stuff? what are these computers going to do, in terms of helping us create new drugs or new materials that can sort of transform the car or the airplane? is that the real opportunity here, isn't it? >> absolutely. >> think there are lots of problems that are extremely complicated in the modern world and the utility of these quantum devices as the prime minister was telling us that in fact, we can put together a large space in just a few elements and if you have the right element to search the space you can rapidly come at a solution and that solution may be the energy structure of molecules or particular drugs designed for a particular disease or a pathway for chemical synthesis or searching large databases there are lots of things that come under the data. >> super computers can't do this. >> what's interesting about

quantum computing versus classical computing. one is ying and one is yang. they solve a certain set of problems and they're very good at that. classical computers solve a certain set of problems whereas quantum computers solve it in some sense. >> they're complimentary. >> we bfrng this as the next wave of artificial intelligence so we got a big boost when we had cloud computing come on and we can could have relatively inexpensive computing that allows us to be breakthrough models and we will see a result of that in terms of quantum computing. >> there will be a lot of give and take, marybell and after a hundred years we agree that yes, it looks like this and now the next wave is how do you put knobs on the cache. you can make it a technology. >> my cat joke has fallen short. you put a cat in a box and you

close it -- and it's got a bowl of poison in there or in your case it's asleep. i'm a cat person. the bad example is a bowl of poison or it could be a sleeping pill. you close the box and you don't know if the cat is dead, alive, asleep or awake and that's the idea of behind that you don't know until you observe something the state that it's in. i know that's a strange aside, but we talked about this cat enough that people will be interested in researching it. tell me, are we ahead of the chinese? are we ahead of everyone else on this? do we need worry about being ahead of other countries on this? the way i look at it, scott is we figured out that there's a finish line and it's worth running that raise, but all of us are at the start line. so in fact, where that finish line is if something globally, we're all thinking about and what's clear in quantum technologies is that something radical will happen and we have the material to sort of run this race and we should invest now to

decide what's the best path to get to the finish line. >> there's a discussion about a national plan and a billion dollars plus and it will go away to the midterm and is that going come back? do we need this now? we do need this now, martin and i hope it stays active in our government and the other sectors because there is a tremendous need to keep high technology alive and really to put resources entirely into the ecosystem which starts with education and it starts with our industry and in fact, it has the national labs and other entities bridging all of these things together. >> dr. irvine sadiqi, an expert in quantum computing. we appreciate you being here. >> thank you for having me. "press: here" continues.

welcome back to "press: here," autonomous cars without backup drivers will be soon be on the roads in silicon rally. the state of california just cleared the way. autonomous clouds suck up a huge amount of data, they do test drives to get real world

information about real world behavior on the road. for instance, how to tell the difference between someone about to turn and a driver who just left the blink or for the last ten miles. the cars come home and engineers download all of that data, but engineers don't share. waymo doesn't tell ford what it learned and tesla doesn't speak to toyota. tim grow is calling for car companies to come together and pool their data. this will beat up the arrival of safe, automated cars. artificial intelligence company specialized in individuals what computers see. thank you for joining us. >> thank you for having me. >> what is the motivation of waymo to be sharing with tesla or ford? what's in it for them? >> i think is a general comment on the state of a.i. and where i think some of the biggest bottlenecks are. historically technology and hardware has been the bottleneck, but today that has been fairly commoditized and it seems that everyone is fairly on comfortable ground when it comes to that. >> using the same sense.

>> oftentimes, the same underlying model when the difference is the quality of the data. maybe it would be the competitive advantage, given how much data is necessary to build good a.i. models, i think it's a greater interest for the overall world to combine the data sets and make models that are for everyone. >> i think a.i. is too important to be only a few silent companies owning the working products and this is something that all of us need to come together and work together to build a truly workable product. >> clearly, there are public safety issues with this and we've had autonomous cars go off the rails so to speak and we've had a death as a result of that. i agree with your perspective on that and a.i. is the problem with that. so how do we solve that perceived problem that a.i. is what's causing this? >> i think this is a again, one of the misconceptions that's happened over the last few years that a.i. is this black box and no one understands how it operates and we have to take the results in just the way it is.

>> a.i. is trained off of human data, and especially with self-driving cars, the basis is the neyo network and as the base implies can you build the machine to think like a human mind and recognize actions and live a life. it turns out that humans are trained and the difference between a cat and a dog and they're not necessarily listing the attributes in a hard coated manner and the solution is simple. shown in a lot of data and the best way to get high-quality data is not to use synthetic data and rather ironically to use a lot of human data. >> i find that fascinating. you have 500, 600,000 people doing this for you. they basically use your cat and dog analogy and they're telling the computer, this is a cat and a dog and it's over and over and

over. >> in the case of tom's vehicle, is it a traffic light and is it a sign and the pedestrian. >> we've got ask the cars and the a.i. and for example, i was reading recently about the studies that have been done about the trolley problem where you have five people over here and you have one person over here and the car's got to crash and it has to decide do you knock the one over or the five over. you have cultural differences in places like china or japan where the collective is very important and there's respect for the elderly, and if there are five young people and there is an elderly person, and scott and i, and we are safer there. how do we deal with this? are we going have to taylor every a.i. for every different kind of culture and situation? >> first of all, i'll answer that question with i think we're quite a ways away from tackling the problem and if the

self-driving cars have a problem recognizing stationary projects and they're optimized for movement and the cars spinning in the middle of the road can pose a risk to the driving car. >> that's a real problem and first, i'll say we're still quite aways from there. in general, there will be biases that we have introduced as humans for each of the different companies that they're operating. >> so i expect auto maifsmakers on top of a platform. >> you might top japanese cars with american a.i. and i don't want an ashtray and i want american a.i. in the car. >> and you have to decide. i think this is -- it will touch on also, and i can't bringing it out for privacy. how much do i want the car to know about me and my driving habits and how it's going to react to my driving habits. >> it will know about you and your driving habit. >> your car doesn't really know about you. >> it has a usb key stuck in the

monitor you, but not every car has that. now every car will speed up and go faster. i like the setting set here or will we have to move at the same speed. is that where we're heading to? >> do we expect autonomous vehicles to all take the road at once? i think not. it will be a slow roll out and before we have true names. one of the issues today we made great progress and we see cars driving in autopilot modes and they're not equipped to be in rain, scompleleet and snow. >> kevin growe is the ceo of a.i. thank you for being with us this morning. "press: here" will be back in just a moment. lot of money.

our consumer team tells you the one thing to check for with every gift you open this christmas. and we )ll get you ready to head out the door with weather and traffic. today in the bay. monday morning from 4:30 to 7.

that's our show for this week. my thanks to my guests and thank you for making us a part of your sunday morning.

i'mwe're also getting ready tandfor los tres reyes magosas, on your "comunidad del valle." male announcer: nbc bay area presents "comunidad del valle" with damian trujillo. damian: we begin with the special programs at the san jose public library system. nancy macias is the head of the library. well, one of the heads of the library. welcome to the show, nancy. welcome back. so, you have this thing called fine-free items. tell us what that is and how the community benefits from that. nancy macias: so, the fine-free items is an initiative that was presented by the mayor and unanimously passed by council, and it allows all children items and teen items to no longer accrue late fines.