Skip to main content

tv   Press Here  NBC  June 18, 2017 9:00am-9:31am PDT

9:00 am
imagine the job where you stare at videos all day looking for murder. facebook needs 33,000 people doing that. my guest henry chang knows how it works. plus, can you run a company as a democracy? and the former head of linked in to get the best job in the valley. our reporters of m.i.t. technology review and john schwartz from "usa today" on this week "press here." good morning everyone. i'm scott mcgrew.
9:01 am
facebook said it will add 3,000 new employees to monitor videos to watch for the most awful content. murder and suicide and pornography and shut the videos down before they hit your feed. we were struck by two things. one, is 3,000 additional eyes enough? millions of video flow into facebook. and what a job. how can anybody do that? henry chang spent hours looking at video like that of founder of open access epo, which may or may not be the outsourcing company facebook is using to screen the videos joined by john schwartz of "usa today." you couldn't tell me if you were. you're under nda. >> absolutely. >> are you the company?
9:02 am
[ laughter ] >> no comment. i want to play poker with you later. 3,000 just speaking generally about this with facebook adding 3,000 people to watch these. is that enough? >> absolutely not. i mean, just some figures out there facebook has videos, photos, and just photos alone of over 3 million a day. and on just per feed basis, over 4.5 billion piece of information. absolutely not. it's a good start, though. >> yeah. how fast can somebody look at a video? i suppose if you see a birthday party it's reasonable to think nothing terrible is going to happen. >> right. right. fairly quickly. i mean, our clients have a pretty elaborate platforms that allow photos, videos to scroll by fairly easily, and the folks are doing it do it all the time. >> so are you filtering the content as it's live or uploaded? the cases of facebook murder in cleveland, that was a
9:03 am
combination. i think confession live the other two that were videos were actually uploaded. could you do both? >> it depends on the client. what we've done in the past has been not live. it's things that have been uploaded and there's some sort of time gap between when it gets posted. >> you've done this. right. >> right. >> can you describe what it's like? >> as you can imagine, it's monotonous. you're looking through pictures and videos. most of the time it's not interesting, if that's a word. but every once in awhile you get the extremes. >> it's uninteresting with the occasional horror thrown in. >> it can take a toll. you can have the super humorous and crazy outtakes but the awful things that you do see. an just depends on the client's terms of service in terms behalf is being allowed and what is not. we delete or flag. >> so the agent are you looking
9:04 am
at multiple screens? give me an idea of are you looking at 16 different screens at the same time and thinking, okay, that's fine. oh, wait. >> yeah i guess it depends on what you're viewing. videos, i guess you can have a few that scroll through but you can't focus on six videos going at the same time. if it's photos you can have 150 photos on two screens and you're going through them. >> when you have a client that says i want you to screen these things for me, the client is going to describe to you what it is they don't want. because it depends on the tech companies to what is acceptable and what is not. >> right. >> do you help them define that? because it's one thing to say no nudity, because then facebook pulls the picture of the little girl in vietnam. you know, there are so many gray areas that just even writing the document that says, okay, here screener, here is what you're going to do.
9:05 am
it has to be complicated. >> it is. there are definitely gray areas. you have black and white. this is the yes and this is no. the gray areas you mark for further review and it goes to a second tier or gets kicked back to the client they may do an internal check. >> so in addition to humans, would there be a technology like an algorithm to find video image or maybe a sub head on facebook live often put a headline on what you're going to show. could that actually compliment what -- >> it actually does compliment. if you're looking at the amount of people needed to review all the photos and videos it's not possible. so you have algorithms in place that kick out a certain number of videos or pictures that need the human eye to review. it's what we do. >> how many people do you have at the company now? >> the company right now has over 1500 employees in three different locations, philippines, taiwan, and china. we do multilingual over 30
9:06 am
different languages. as you know content is everywhere around the world not just in the u.s. around 1500 now. not just doing content moderation but back office support. >> the other thing that seems to be appealing for video screeners as i was studying up on this. it seems to be where video screening seems to be done. is there a particular reason other than cost of reason? >> there is a particular reason. the u.s. as far as how philippines are alike to us rather than, for example, in india. right. culture is different. religion is different. their day-to-day activities are different. to have someone in the philippines view a picture or video they're probably going to be the closest to what america would be like. >> do you have a connection to police or -- jon mentioned the crime put out on the platforms now and last year philando ca
9:07 am
castil castille's fiancée live streamed it. >> i don't think from our end we don't have the relationship. it's kicked back to the client and that's their -- >> do your clients have it? i can see where tom is going. you see somebody saying that they're going to do something terrible. it should be more than just, well, you can't do it on platform. there ought to be a way to say, hey, guys i think something bad -- >> a notification system for law enforcement. >> right. when we notify it's real time to the clients. >> okay. there's somebody on the other end in san bruno or -- >> yeah. >> and a red light goes off that said, hey, this is something we need to look at. >> so radio and television for years it was a seven second delay. >> yeah. >> there's not one here. [ laughter ] >> so conceivably could it happen with live video over a
9:08 am
platform over a par scope on facebook live? >> i wouldn't see why not. it's really based on facebook or whatever social media what they decide to do. >> yeah. >> we're just on the other end. >> as far as you touched on it the computers can kick out, let's say a hundred videos in an hour which it says i think these may violate some sort of terms of service. which goes to a human being. >> right. >> it means that human being is looking at potentially the worst things that could be kicked out. >> potentially. >> how long can somebody do that? i mean, i realize most would be like computer. come on. that's fine. then there's a point there when we go to work for eight hours and it's nothing but people hurting each other or other terrible things. how long can somebody do that and do your employees know when they get in there that's what the job is? >> it is. they know what they're getting into. from a company perspective, we try to provide a very soothing
9:09 am
environment for them to work in. depending on the type of work we may even have counsellors on site they may speak to after seeing these videos or pictures or after their shift to kind of -- >> it's like -- instead of hearing somebody on the other line in panic -- >> the 9-1-1 operator. >> they're seeing it visually. >> right. >> henry, last question. you're on the cusp you're one of the very few companies that do this. you're on the cusp of we're going to need you guys to double down, triple down, quadruple down on what you're doing. it's a growing industry. >> yes, it is. it's growing. it's not just now. it's been growing the last few years. >> henry chang paying attention to what you're facebook on wherever. we don't know if it's facebook. >> right. can you run a company as a popularity contest? "press here" continues.
9:10 am
9:11 am
my wife tells me her water broke and between the excitement of that and dealing with this i -- >> barracuda network was able to assure the worst-case scenario didn't play out. the permanent loss of thousands of gigabytes of precious data. >> we found the files that had been encrypted and deleted them. i went straight to the barracuda backlog and restored those files.
9:12 am
welcome back. there's a device at my local car wash with a bunch of buttons on it. faces, actually. it allows customers to give real time feedback to the car wash. i would assume from time to time kids mash all the buttons. the thing to show people are happier, perhaps, on tuesdays than thursdays. the car wash might want to look at the thursday stats. that idea is spreading into offices. corporations are sending real time surveys to their employees and the data is helping managers make decisions the same day. jim barnett runs the surveys. he's founder and ceo of silicon valley-based group. you send the surveys often to employees or clients. how often? >> well, it's typically sort of monthly or quarterly.
9:13 am
but most people are focussed on getting feedback and starting conversations between their managers and employees. so it's -- think of it as continuous-basis but not constant. >> the client can ask whatever he or she wants. give me examples of things i wouldn't think of >>well, it starts with what we call our key drivers of engagement. it's things like whether you recommend a certain company as a great place to work. whether you recommend your manager to others. >> in the usa today you probably do that. we do them yearly, i would guess. >> yes. >> why do quarterly or monthly? >> historically that's what companies did. once or twice a year. ceo would get e-mailed a report. the reality is the work world we live in today is real time. people are used to getting feedback on a continuous basis. snap chat, facebook, instagram, open table, whatever applications people are using. and so our work is the same way.
9:14 am
we want to give and get feedback on a regular basis. >> do ceos have -- monthly and quarterly it's easy. >> great ceos make time for this. feedback and monitoring the health of their organization. there's nothing more important. right. we actually now have data that shows that. we have millions of employees that have been responding to our pluses for the last three years. and question see that companies where you've got high levels of engagement, companies in the top cortile their stocks outperformed other companies by 42% last year. average market cap for a company -- >> hotel is probably fairly high in retention? >> retention is a huge driver. i'm thinking silicon valley as expensive as it is to live here and the competition for the talent the way it is, you want your employees to be happy and not just beyond -- >> exactly right. it's all about business outcomes at the end of the day. in fact, we have one company
9:15 am
that installed glint and five months later they reduced regrettable attrition by -- >> i have not heard that term. >> the companies think about it people they didn't want. >> yeah. regrettable attrition something that comes up in silicon valley the retention of female employees is particularly adult. in the tech industry. companies keep saying they're going to improve and raising numbers but the numbers aren't moving. do you have any insight on that? >> well, what we do find is that companies that are able to see their data buy across multiple attributes like gender, ethnicity, date of hire, generation, et. cetera are able to take action at those specific levels. so one of the things that systems like ours does is actually sends managers an alert. and one of our companies actually got a particular alert that shows that engagement level
9:16 am
for one of their teams had dropped significantly. and it turned out it was actually a business issue. they had installed software that was allocating significantly more work to so. people on their team. >> you would never have known that. >> how big of a team do i need before this data actually tells me something? obviously a team of four and bob will never quit complaining, you know, i mean, it's not -- the end number is not big enough. >> so, you know, we see that companies even with 75 to a hundred employees sort of getting this confidential feedback from their employees. it's extremely valuable. of course, we have companies with, you know, 300,000 -- >> yeah. then you get a good statistical -- yeah. >> the consequence of, say, there's a manager or supervisor that employees are having a problem with? does that lead to action on behalf of your customers? >> it does. you can't hide from the data. one of the things our system
9:17 am
does is send you a report that shows the ranking of your managers by engagement levels of teams and management scores. when the data is staring at you in the face, you know where you've got issues and what to do about them. >> now i'm wondering for the managers will be using a system to complain about the system. do people want to be ranked like this? >> i wouldn't say necessarily ranks your managers, what it does is gives each of your managers feedback. and so one of the things it does is for each manager she or he will see you know what are the two or three things i should be focussed on now for my -- >> there is a danger of wagging the dog. right, i mean, this is not a democracy in the company. i mean, steve jobs probably never sent out, you know, surveys about how am i doing? you know, he lead from the top. right. there is that danger, you know, if you're spending more time paying attention to what your employees want it's what your customers want. >> yeah. >> so i think there's a really big difference between feedback
9:18 am
and stimulating conversations and asking the data to make decisions for you. so it's not about a polling your employees so they can tell you how to run the company. glint is about helping you monitor the health of your organization in real time. >> what is your biggest surprise that a company has found out? of course, they probably don't tell you. >> i think the biggest surprise we've seen is how consistent the data is. if you've got a long engaged team you'll have attrition on the team. so the system can do things like actually predict things like attrition, business outcomes, patient satisfaction, customer satisfaction based on the engagement levels of your team. >> jim barnett, thank you for being here. >> my pleasure. thank you. absolutely no one would know more about how to network your way to a high-paying job in silicon valley than the guy who hired people at linked in. when "press here" continues.
9:19 am
9:20 am
welcome back to "press here." the trump administration is in the middle of changing the rules on h 1 bv so more americans get involved in high-tech. high-tech said the industry can't find enough qualified americans and has to look elsewhere for talent. yet, we do know qualified american engineers complain they can't find a job. which is it? we found ourselves one of silicon valley's head hunter. former head of hr to answer those questions for us. steve is the former head of talent at linked in. i'm guessing you hired a couple of engineers in your day. >> many hundreds of them. yes. >> so, you know, we don't have to dwell, but let's talk about this for a minute. >> sure. >> is it true that a high-tech company cannot find enough qualified american workers --
9:21 am
speaking of high-tech, of course, in this day and age? >> i think the way most organizations in the valley look at this is they want to find the best tal i waent in the world. that's what the draw for silicon valley. it's not just american workers, but the best talent they can find. h 1 b have existed for decades. i think if we take a look at who the h1b turned into. people like -- >> i'm not disagreeing but it's interesting you put an interesting shade to that. or spin on that, which is it's not necessarily we can't find the americans but we can find even better more qualified people somewhere else. >> right. that's right. i think that's the goal to find the best talent in the world. because mostly your products will reach people around the world, too. you want as diverse a talented organization as possible. and every organization that i consult with today or that i was competing against when i was at linked in is really trying to
9:22 am
find super talent. and their supply is really limited. >> what happens if the h1 b -- 85,000. >> if that comes down, what happens to the companies around in silicon valley? will it be a major impact on their products? well, i've got say some of the people i'm talking to already people are making plans to find second sources, second locations. >> other than canada and mexico? >> in part. in part. all the companies -- the big-named companies facebook and microsoft have some call them holding places. i think the bigger issue is the fear of international talent, international technical talent right now is do i want to go to the u.s. for an education? because i may not be able to get an h 1v. i think we're cutting our nose off to spite our face. i think it's a risky strategy long-term. look at the recall it created by having great global talent here. >> that's the thing, too.
9:23 am
someone comes and do a good job and start their own company. every person that comes here it equates to more like 2.7 jobs. they're not only doing well here but creating more -- >> that's a great -- and the term i think of is force multiplier. you have people who are innovator and pioneers. they create more jobs for other things. and the best industry to take a look at that, for example, would be the travel industry. it used to be tons of travel agents. then all of a sudden online booking. and then trip advisor and yelp and the other sites came. there's thousands of new jobs created in the travel industry. airbnb too. travel agents their work has been impacted significantly. another view of the employees. there are stories about software engineers that come to america who were asked to train employees that took over their job at a lower salary. a lot of h 1 b go to outstoring companies. >> i'm not sure if that's
9:24 am
necessarily true. i haven't heard the stories you've heard around people training -- >> they've been widely reported. >> really? >> and regardless of how you look at it. i think if you look at the testament of half a million to million unfilled computing jobs in the next couple of years. meanwhile you have a lower number than 85,000 for h 1 b. you're creating a situation where you have to retrain americans who are here in their skill sets, which we see through coworking locations or through coding programs. is that probably what is going to happen regardless? >> yes. i think the trend is technology is changing faster than our ability to upscale the tal toent drive the necessary skills to build the technical competency that organizations need. what the discussion in silicon valley has been more around technical talent like high-tech jobs. honestly there's more nonhigh-tech jobs whose job is being transformed. >> it's like the legacy companies, too. like the banks or health care who are well established in the
9:25 am
u.s. they're really keen to cybersecurity and payment systems. they need people who have skills in those areas. i think this is -- the irony is that the trump administration on the one hand wants to create jobs. in a sense they're already here but they might be leaving because of the h 1 b. >> it could be. i think even nontechnical jobs like digital marketing. it's hard to find those people. if you go on facebook sites now or apple. each company has about 400 openings for digital marketers which suggest a huge need. >> i'm going ask you that. it i'm a young college student watching and planning to work in silicon valley. you have your finger on the pulse where the jobs are. digital marketing? it sounds like the guy who runs twitter. but not running twitter and the ceo but the twitter guy at the company. i mean, are those -- where are those jobs? i think right now the best skills that a new graduate could be focussed on is making themselves an open system of
9:26 am
constantly learning and being ag guile. >> where is the skill? i can only learn so many skills. where is my skill? >> being digitally fluent. understanding social networks and user experience, understanding so many consumers and people in the world are consuming and transacting through a mobile device. so mobile or people that understand how mobile works and influences jobs. super powerful. >> you think some of the tech companies are too picky who they want or think is qualified to work for them? one of the criticisms against facebook or google they recruited from ten colleges in the u.s. they wouldn't consider anybody hard beyond. >> that's hard to argue. i remember facing my board at linked in after we hired a thousand people and they said there's no way you hired a super stars. we said no we've been discriminating and trying to focus as well as we can. but that's the pride they have is that we are discriminating. it takes a hundred candidates for us to find one.
9:27 am
i don't think they'll change their approach to that. >> thank you. and we go to a commercial.
9:28 am
that's our somehow for this week. earlier we talked in detail about facebook's plan to better screen the live video and how hard it is. intelligently and emotionally. it if you missed it at the top the show. that interview and all of our interviews are available online. on itunes as well. i'm scott mcgrew. thank you for making us part of your sunday morning.
9:29 am
9:30 am
"comunidad del valle." i'm damian trujillo. and today, summer camps--cultural summer camps for your children, plus the "t4 texas" show, on your "comunidad del valle." male announcer: nbc bay area presents "comunidad del valle" with damian trujillo. damian: we begin today with the mission asset fund. with me here on "comunidad del valle" today is a special guest, mohan kanungo. he's here to tell us all about the special program, basically financial services for immigrant families. welcome to the show. mohan kanungo: thank you. damian: tell us about your program and who you help. mohan: so, mission asset fund is a 501(c)(3) non-profit, and we provide zero interest loans that help people build credit, and we're excited to launch a new program with the mexican consulate of san jose where we're providing free financial education to anyone who comes to receive services at the mexican consulate. damian: we do have video of the mexican consulate, and i mean, we know the folks who attend--or kind

88 Views

info Stream Only

Uploaded by TV Archive on