tv Press Here NBC October 11, 2015 9:00am-9:31am PDT
9:00 am
9:01 am
>> good morning, i'm scott mcgrew, each year at this time, the faa reminds pilots they may not fly over san francisco over three straight days to clear the way for fleet week and the blue angels. it is an old fashioned term and notice that air people would be more appropriate. but bottom line, the city waterfront is a no fly zone. you may not fly drones during fleet week. the federal government has been working hard on drone rules, whether they can be used commercially and some of the best -- drones flying one second
9:02 am
for obama. not only were you a close advisor to the president, you were a law professor, presideof obama. >> yes. >> we are talking about the buzzy drones civilians use, not predator drones in afghanistan, that kind of thing. >> exactly, small drones, 55 pounds or less. you were one of the very first people to recognize that --
9:03 am
and athat point the faa had been paying attention on working on safety issues but at that point the whole of federal government really started to pay attention and talk about the privacy issues. i don't know if you followed the issue in north dakota where a legislator said we don't want to have our local police have guns on drones so we'll put up this legislation and what happened is now there can be tasers on drones. in the state, the police
9:04 am
department can engage drones with tasers. is this something that's going to be replicated elsewhere? >> my understanding is i did hear about this legislation and my understanding it was somewhat of an unintended consequence. i can imagine perhaps an active shooter situations there could be times when it could be appropriate for law enforcement to use an armed drone with less than lethal force but that's certainly not what law enforcement is planning or to use drones for at this point. >> i'm really interested in sort of where the commercial development is. you can announce you're going to have something that books delivered by drones, marijuana delivered by drones and pizza, whatever it is and the press goes crazy but it's not really happening in scale yet, right? is it going to be a while before it happens in scale? >> right now whether it's legal
9:05 am
or authorized depends on the intent of the flight. if it's recreational or for fun, i can fly under a set of guidelines if it's commercial, i have to get a special license, 333 exemption, the most common of the licenses that the faa is issuing. there's been huge demand for everything you've described but almost 2,000 have been approved by the faa and faa recently said they are trying to approve another 100 every week. >> even when you get the commercial license, you're bound by a lot of rules. almost $2 million, a company out of new york city, they were flying legally under the faa license but apparently violated a number of rules at the same time? >> i just looked at what has been alleged against the company but a lot of that activity predated from when they got their 333 exemption. >> they were breaking the rules
9:07 am
9:08 am
way. how do you control that? is that hopeless? is that a dangerous mess? >> we've seen no drone zone example, that's very positive press in terms of educating folks that when they buy a drone they are not just buying a toy. you need to be responsible and understand the rules and understand what the rules are where you're flying and most important when you're a hobbyist, you don't qualify. >> what i would like to do, we publish this on the internet as well. stick around, our television viewers, we're going to talk about the creator of old school. if we can ask you questions for home drone operators, can i do this? would that be all right with you? >> sure. >> you'll be able to see that.
9:10 am
9:11 am
can compete in a global technology. this is the basic plan behind alt schools, sometimes inside a storefront popping up around san francisco and silicon valley. this would be crazy expensive but money is not a problem. max vandilla a former googler and his school got a million people from people like mark zucker berg and the widow of steve jobs. what do you know about education? you're an engineer, aren't you? >> so, i have a technical background. we started alt school two and a half years ago it was in part r partnership with child educators who spend decades how to personalize the school experience for children and how you reorient and what is a very different world today that you're preparing kids for. >> you surrounded yourself with experts. what's your interest in education? why of all of the things the
9:12 am
next googler can do to save the planet was it education? >> my mother is a teacher and sister is a teacher now in altschool. it was nowhere around me but i had a 2-year-old daughter back in late 2012 when i was thinking what next. and i wouldn't be doing this if i wasn't a parent. my daughter will start at alt school next year when she can at 4 years 9 months and my son leo will. >> we're doing it in a way where ultimately this can scale and the experience i would want for my own kids and i'm jealous of them for getting to have is something that gets better the more people get that experience versus the model we have now where it worst more kids get a certain experience. >> i would be interested in your take on tech in the classrooms. looks like you use technology. there's been a lot of bumps along the way.
9:13 am
there's a whole debacle about the i pads in the l.a. classroom where the content wasn't loading and it was a big misfire. most classrooms i've seen they are never state of the art and almost at the point of being unusable. should we kind of toss the idea of computers in the classroom out? >> i don't equate computers in the classroom with using technology to support a different kind of education. i would say the most important technology is invisible when you go into the classroom. it does what technology does best, which is to reduce the cost of complexity and allow you to be more flexible. you would think of large scale systems like the internet or air traffic control and they can deal with an amount of complexity that would be impossible without using computers in a fundamental way. that flexible translates into a more personalized experience for students and more autonomy and more flexibility versus wants
9:14 am
and needs versus another. it's not about replace ag world process -- >> you have computers in your classrooms but you're using classrooms to study how kids are learning and churn out lesson plans as well, right? >> that's right. so about a third of the time is less -- a third of the time the student might have a screen but only about half of that time are they learning off the screen versus using it to photograph or interact with anyone with something like xip referring to their own playlist of activities they are responsible for that week. you can use technology to kind of represent what is a child doing and provide suggestions and provide information but ultimately that should be an input into what's a learning process that takes place off screen as much as possible. lesson screen learning the better provided you're getting what you really want from technology, which is a level of transparency and flexibility. >> given the probably the early
9:15 am
adopters of this, i want to guess you're up to your eyeballs in metrics versus brand x. what are those numbers telling you? >> well, first off, we're looking at satisfaction. we really believe that if you don't have a school experience through the kids enjoying being in school. you've kind of lost in the first step. we measure satisfaction from students and from parents and educators and employees in the company every month at a very detailed level. that's the first kind of data. it has nothing to do with like what is the child doing this moment that i think is incredibly actionable and something that i would push other educators on to say kind of what is your satisfaction. how is it trending? what are you doing to make it go in the right direction? on the other side we look at standard space progress of students both through the constant assessment that teaching team is able to do in reflection of what the student has done today and in terms of more periodic kind of standard
9:16 am
based assessment we can reconcile with what would be done elsewhere. >> you have investments and grants and parents paying you $20 thousa2 20,000 a year. but what about public schools? we have this incredible schools in silicon valley. you want to use the ideas about using computers to measure student satisfaction and determine where they are in the lesson plan into public schools too, right? >> we see altschool as a three-phase journey. let's create an ecosystem that is different and makes it easier for people to start and run great schools that we would love to send our kids to where we're the only inhabitant much that eco system. higher priced schools in metro area. you have to walk before you run. we're starting with this complex ambitious goal to enable all
9:17 am
children to achieve their fonl potential in a way to take those first steps and it's easier. we have seven schools right now -- >> you can make mistakes in private school as well, that didn't work. >> you have flexibility. we have seven schools in two and a half years and working on our first public school if that was the route where we began. but ultimately that goes to the second phase, you can enable other people to start and run great schools including charter schools and public schools using the same platteform and tapping into the network that we've built for ourselves initially. and the thing that's exceptional about altschool, we have 45 educators right now and 45 engineers. there's not another school on the planet with a 1-1 ratio with that scale. when we control every aspect of the experience for ourselves initially, learn very, very quickly so that we can then move on to a model that has relevance for other people to start schools and ultimately in the
9:18 am
third phase for the existing large school systems where the majority will be educated for the foreseeable future. >> what do you think as a parent and educator of this really huge trend of trying to teach young kids like coding. there's a lot of coding schools everywhere around here even and big push by a lot of people to get more cs classes. is it appropriate at a certain level but younger level they should be learning logic or something else? >> so there's a lot of studies that show that kind of getting your 5-year-old to start coding is counter productive and tends to be 10 to 13 where actually people that started that early have a benefit prior to that there's no benefit and some statistical detriment. so yes, absolutely, you want to think about what are the foundations of programming and what are essential logic skills
9:19 am
but more importantly, what are basic character skills, perseverance, curiosity, what are kind of cognitive foundations that you build on later. you want to work on those. it's almost like night answer to technology, best training looks nothing like programming for a very long time. once you start to look at middle school, i think for most kids, programming should be a means to an end. there's going to be a subset of kids that are interested but the essentials to think about programming as a way to do other things you're interested in, whether it's create music or art or learn about history. that's where education really meets the ship. >> the founder of altschool, thank you for being with us. >> thank you for having me. >> the world will soon be run by algar algarithms
9:21 am
welcome back. recently psychologists at the university of minnesota published a study that proved machines are better than people at hiring. it's better to let a computer, an algorithm decide whether you should hire that candidate. your hp will pick a better applicant than your hr. this is actually happening. linkedin runs an algorithm called people you may want to hire. tough luck if you're not on the
9:22 am
list. this is affecting you right now. algorithms and machine learning maybe the most important development since computers and nobody knows it better than this man. it would not an exago race to say he is one of the brightest minds in machine learning, a sloan fellow, fullbright scholar and author of "the master algorithm." thanks for being with us. we're talking about a mass algorithm like unified theory in physics? >> this is exactly what it is. the master algorithm is the grand unified that brings together the different types of learning into one learning element that can learn anything from data in the same way that the standard model in physics describes the universe. >> an ultimate smart learning computer, artificial -- are artificial intelligence and big data and machine learning kind of all the same thing? >> they are related but they are different. think of it this way, big data
9:23 am
is the fuel and machine learning is the rocket and we want to have intelligent robots and what not, machine learning is how they get there and the learning runs on the data. >> well put. >> are you worried about the eventual outputs of ai, elon musk has put out a warning about that. >> there's a lot of talk about the terminator scenario. the truth is among ai experts, nobody takes that too seriously. the reason is in hollywood movies the ais and robots are humans in disguise. the real ones are nothing like human. being intelligent and human are different things. we tend to conflict the two because we've never seen them separate. part of what people should understand it can be infinitely intelligent without being ofny danger because it doesn't have its own goals or doesn't have
9:24 am
personality or consciousness. it's intelligent but not human. >> what if somebody bochs the code review. there are a lot of smart people at apple and microsoft and bugs in the koid. they are not coding something smarter than us. does that worry you at all? >> it does worry me but in the following sense, what happens is that chances are the result will not be smarter but stupider. the computer will make awful mistakes it wouldn't if it it was smarter. a lot of bad things that happen in the world, including disasters like power outages and rockets clashing. but the dur is cure is to make system more intelligence. >> where it can detect? >> exactly. when amazon decides this is the sort of thing you like this and you like that. i don't get to see the inside machineations of that. i mean, we are depending on algorithms to have been honestly written to say the algorithm says this is what we should do to save this man's life or what
9:25 am
not? >> you should be able to see the machineations of the algorithm. it's very exciting that you get this personalization from companies like amazon and what not. but really you should be able to take to the algorithm and see what it's doing in a way you aren't now. amazon has its own goals that aren't necessarily aligned with yours. >> if you showed me the algorithm, i would say, oh, yeah, i would have no idea what i was looking at. >> surprisingly what they do is -- these find people who are similar to human taste, like if you like the same music she did and just gave it five stars -- >> not unreasonable. >> it gets more interesting when regulation, public interest becomes involved, something like red lining, we're not actually looking at rays we're looking at
9:26 am
geograph geography. that's not kosher. there could be lots of workarounds where you don't understand what the correlation is. should that be discoverable by all users or by competitors? >> this is a great example. this is something that's being discussed in machine learning today. how did we design the algorithms so they don't discriminate. the learning algorithm doesn't know gender and race. and for example, a lot of people get credit today that they would not if they weren't being graded by an algorithm. there are places where they can pin the data. you can design into the algorithm goals one of its explicit goals to be fair in terms of gender and race. this can be done and it is starting to be done. >> you teach at the university of washington and must get the sense of future of the world is probably sitting in your
9:27 am
9:28 am
9:30 am
i'm damian trujillo, and today facebook comes to our studios. they're called latin@ right here on your "comunidad del valle." male announcer: nbc bay area presents "comunidad del valle" with damian trujillo. damian: we begin today with fleet week, which has been in the bay area for all this week, this past week. garth langley is public affairs officer with fleet week, and with us also is miguel torres, a member of the united states marine corp. welcome to the show. garth langley: thanks you so much for having us. damian: we have some amazing video from years past of fleet week. talk about what we're seeing here and what can we expect to see this year. garth: well, you know, we're bringing back the 35th annual san francisco fleet week, a long-time san francisco maritime tradition, of course brought to you in 1981 by then-mayor diane feinstein. you know, in the last decade or so, we really re-shaped fleet week. you know, it's not only a time to celebrate our armed forces,
390 Views
IN COLLECTIONS
KNTV (NBC) Television Archive Television Archive News Search ServiceUploaded by TV Archive on