Cloudflare TV

Kids & Technology: How Do We Keep Them Safe & Sane?

Originally aired on 

Best of: Internet Summit 2018

Session 1

Kids & Technology: How Do We Keep Them Safe & Sane?

  • Jill Murphy - VP & Editor-in-Chief Common Sense Media
  • Sara DeWitt - VP PBS KIDS Digital
  • Moderator: Alex Dyner - Head of Special Projects, Cloudflare

Session 2

Kubernetes & the Serverless Future

  • Kelsey Hightowner - Staff Developer Advocate, Google Cloud Platform
  • Moderator: John Graham-Cumming - CTO, Cloudflare
Internet Summit

Transcript (Beta)

supported by Cloudflare Cloudflare Tv I'm glad the room just cleared.

OK, so our next session, switching gears a little bit, is about kids and the Internet.

And our next panel is the perfect panel to speak about this.

With me are Sarah DeWitt, who is vice president of PBS Kids Digital, and Jill Murphy, who's the editor-in-chief of Common Sense Media, which provides content about the suitability of media and tech content for children.

So this is a big topic. It's a hot topic in a previous panel.

Julius Genachowski mentioned that one of the sort of backlashes against large technology companies today, part of that is driven also by the effect of technology on children.

So Sarah, we talk a lot about here about free and open Internet.

This is, I think, pretty clearly an area where some regulation, at least by parents, is required.

So how should parents think about what kinds of content their kids should have access to?

There's so many ways to look at this. One thing that we should all keep in mind is that the idea that tech is bad for kids, the anti-screen time movement, Jill and I were just talking about this back there, it's actually a very upper middle class issue.

We do a lot of our testing with kids from Title I schools, do free and reduced lunch.

Those parents are often desperate to get more technology into their kids' lives.

They feel like their kids are falling behind.

So this is a really interesting debate that's going on all the time.

But back to your question of how do you choose the best content for kids. One thing to keep in mind is that kids look at technology as play.

This is a game. This is something fun to do.

And so we are always saying to parents, you've got to think about what your kids like to play and also what you're hoping they can learn from the experience.

And so you want to look for something that really kind of meets kids where they are.

If they're really excited about dinosaurs, then you want to go find a really good dinosaur app for them to play with and read the reviews and do that kind of thing.

And you want to look at who's publishing this content. And you also want to think about how technology fits into your child's life.

You want to actually think about what times of day are the times that kids are going to play and make it part of their schedule.

Like, you're going to be able to play with the iPad on Tuesday nights while I'm making dinner.

And then afterwards, we're going to talk about what you did.

And I want you to show me what you played with or show me how you won it.

So what we say to parents a lot is the thing with raising kids to be kind of healthy with the technology they're using is not to just hand them a device whenever you need them to be quiet, but to really think about, what are the times of day?

And what are the times in the family's life when it's going to make the most sense for your child to be able to play with technology?

You talked a little bit earlier as well about, when we were backstage, about sort of combining online and offline and how kids absorb information differently and maybe better when you have the combined experience.

That was interesting, I thought.

There is some really great research out there. One study, very recently, that came out from Texas Tech was looking at the combination of screen time and conversation.

And they were looking at a PBS Kids show. This wasn't a study we commissioned, but that's one of our shows.

Mr. Rogers' Neighborhood, sorry, Daniel Tiger's Neighborhood, the spinoff of Mr.

Rogers. And what they learned was that kids who watched the show were gaining skills with empathy.

It was actually deepening their empathy.

But the gains were much greater if a parent talked to the kid about what they watched after they watched it.

And so they also tested just parents talking to their kids about empathy.

And that also had some gains, but it wasn't as great as the combination of the two.

Something about letting kids kind of get into that world and then having the characters as context to talk about actually increased the benefits of the technology.

So if you go away with anything today, it's talk to the kids about what they watch and play.

That is like a huge piece.

Jill, tell us a little bit more about Common Sense Media. I think most folks in the audience probably know about PBS and PBS Kids.

Common Sense, it's a great organization that's been around for I think 10 or 15 years.

Almost 15 years.

I'm sure everyone here is very familiar with it and is signing up for our newsletter as we speak.

We're a nonprofit organization. And so I oversee all of our ratings and reviews.

So we review everything from books, games, movies, TV shows, apps.

And we do it from a parent's perspective. We do it from a child development perspective.

We rate everything from two to 18. So we're looking at everything from Elmo to The Hangover.

And we know how much media is aging down dramatically.

So we know that a kid as young as eight is interested in the latest Marvel movie that might be out.

So we're really conscious of the type of media that kids are absorbing and what they're ready for and at what age.

And then we do a ton of parenting advice as well.

So we take a lot of that content and we put it into best lists.

We put it into articles like 50 books every kid should read before they're 12.

Just trying to make it really easy for parents and kids to make a selection with all the great content that's out there.

And are there certain trends you're seeing recently?

There's, I think both on the, when you're saying both on the kind of very positive side, I think often, especially recently, technology has a bit of a negative implication with kids.

But are you seeing sort of trends that are very positive?

And then also on the flip side, it'd be interesting to hear both sides of that.

We sit very much in the middle of this whole quote unquote debate over, it's too much content or it's too much screen time, yet we're all handing our phones and tablets over to our kids.

One positive thing that has come, I'm constantly saying kids are winning in this game when it comes to TV.

There is so much great TV out there right now on all these different streaming platforms.

If anything, there's too much and it's really hard to make a decision. So that's been one really positive spin.

I think that, and that runs from the very young ages.

Preschool content is really strong right now. PBS, of course, all the way up to kind of between space.

So that's one really great thing that we've seen. Reading has been back on the rise.

So that's been another really positive spin. So again, we cover all different types of media.

When it comes to technology, specifically social media, this is, we always kind of say zero to eight is when the parents are more in the driver's seat.

We have a little bit more control over the decision-making.

Nine and up is like damage control. It's when parents come to us and they're like, what do I do about the cell phone?

What's the right age to get my kid a cell phone?

My kid's on social media. Are they too young for it? Their friends are on it.

I don't know if I want my kid to be on it. It's really just, we're all making these decisions on the fly.

And so we're kind of constantly encouraging parents to take a beat, take a minute when it comes to making those decisions.

Don't make them too quickly. Oftentimes we hand over a phone because we bought a new one and we're gonna give our old one to our kid and that's their first phone.

And the next thing you know, they're exposed to a ton of content that you weren't ready for.

So kind of easing into some of this has been something that we're really focusing on.

And a report you guys published recently about the effects of social media, I think it was on teenagers or it was on teens.

I thought it was fascinating as well because I thought the report would come back and say, oh, it's horrible, it's terrible.

And I had a pretty balanced conclusion that I was surprised by.

I think most of what we have been finding is fairly positive actually. Or I would say more of a balanced approach.

Kids are, you know, it's the parents that are concerned.

We're the ones that are freaking out. The kids are kind of feeling like, I know that these are my friends online but they're also my friends at school.

They do not make a distinction between, you know, that's my social media life and this is my offline life.

Like that does not exist. Bullying doesn't only happen online and not on the playground.

So this is just, I think that's something that parents are kind of coming around to that this is very seamless for teens but they're also pretty sophisticated about the way that they navigate social media and kind of starting to self-manage their own habits a little bit, multitasking and that sort of thing.

That's not all kids and not all teens. But I would also say that one thing that has come up again and again in the last few years of our research is parent role modeling, which, you know, I'm so guilty of all of that but being a role model and trying to show your kids the way that you want them to behave with their media, whether it's putting the device down at the dinner table or putting it away, you know, not jumping for a notification, that has a lot of impact on teens especially and younger kids.

Yeah, it's interesting. Sarah, we were talking about, go ahead.

I was just gonna add to this. Do you, have you followed this idea now of narrating what you're doing on your phone?

So if you think about back to when we were kids, if you were invited to a birthday party, you got an invitation in the mail or your kid gave it to you or you got it at school and so you had this physical thing.

You knew you were going to a birthday party. You might go with your parent to buy the gift.

You were there when your mom called RSVP. You all got out the map together to look at where you were going and then you go to the birthday party and now it all can happen by the parent on the phone and the kid doesn't see any of it, doesn't know that like, oh, now my kid has ordered, now my mom is ordering the gift online and now she's looking at where we're gonna go and so the idea of actually saying to the child, I'm looking at my email because you just received an invitation to a birthday party.

What do you, to actually include kids in what's happening on that device, I think is a really great idea.

Yeah, and also, I just wanted to circle back to what you had said before about talking to your kids.

I think parents often, this is kind of our mantra at Common Sense when it comes to advice for parents.

It's always talk to your kids, talk to your kids, but what do you say?

Parents can fall apart when it comes, it's showtime. Now I gotta engage with my kid about what they just watched or read or played.

So we do, and this is my shameless Common Sense media plug, but we do have conversation starters within all of our ratings and reviews because it does feel like a little bit daunting to take on this task and it is much easier to hand over a device and just, oh, I'll keep them busy.

But I think there's very few things that our kids go out and do that we don't come home and ask them about.

So kind of using that same framing of they just engage with something online.

They're watching some other kid unbox something on YouTube or they're watching some other kid show a room tour, which just drives me crazy.

My kids love those. But talking to them about, well, why do you like that?

Those are essentially commercials at this point, in my opinion.

So how do you try and start to help them understand what they just watched, why they're watching it?

Is it real, is it fake? There's advertising in that.

Those kids might be getting paid. They're all teachable moments and you don't have to take all of it on at once, but using a little bit of it as leverage for a conversation can be really valuable.

You think about the dinner table when you say, what did you do at school today?

Well, what did you do in math to actually say, what happened in that show today?

Can you tell me the story? Yeah, especially for little kids.

And by the way, kids love to talk about media. They love to tell you about a book.

Actually, it's almost annoying. They'll just go on and on about a show and whatever it was, at least my kids, for sure.

A mom last week I met on the edge of the soccer field told me, she's a learning specialist, and she told me that after her kids watched TV, she was very clear that she doesn't like her kids doing screen time.

She goes, but when I let them watch Wild Kratts, I then have paper and crayons out for them to draw what they watched.

And I was like, okay, super mom, that's great.

Therapy. So the next Saturday I was like, okay, I'm gonna try this.

They were watching a show, and when they were done, I was like, guys, here, come to the table.

I've got crayons and paper out. Can you draw, what was your favorite moment today?

And they totally did it. And I don't think it would work all the time.

Yes. And you still hate that mom for telling you. But it was a good idea.

That's a good idea. You mentioned this before, when you were speaking, in terms of the kinds of content that the kids are getting as well.

And we're talking about a distortion between the incentives that, especially folks that work at for-profit organizations have in creating content and creating apps.

They're motivated by, sometimes, by different metrics that are necessarily tied to what's best for kids.

Yes. How do you sort of, how do we work through that, I suppose?

This is a major issue. The majority of game content for young kids is from the app stores.

It's in Google Play, it's in the App Store, it's in Amazon. And the success metrics for those stores are, like, what drives revenue is time spent, how long you're in the app, how many ads you see, how many in-app purchases you download.

And this is trickling down into the kids' space. So it is incredibly hard for a kid's content product to break through without it being huge, without it being successful on this revenue side.

And that revenue is, those success metrics are not actually good for kids.

Like, this is a real, it's a real problem to be designing something that is just sucking kids in for long periods of time.

And if you watch different kinds of games for kids, you can see, like, at the end of every game level, it immediately hooks you into the next one.

I mean, I know grown-up games do this as well.

But it's a problem for kids who don't have those executive functioning skills yet to put something down.

So I really believe this is an area that needs to be disrupted.

We need some models to come up that are actually focused on how do we use this technology to help kids learn and to help them gain some pro-social knowledge to understand how to put it down.

And it's possible to do those things with the technology.

But right now, that's not being rewarded. There are people trying, but they just can't break through.

Yeah, I think these are the challenges that then we hear about from parents, which are, my kid is having a temper tantrum after I tell them to turn off the iPad.

I might have submitted that one.

But I think those are the, it's a constant battle, and the autoplay, and it just, it doesn't stop.

And I think, yeah, using the technology in a way that kind of sets up those boundaries, and with purpose, not just for the sake of, like, oh, we think it's the right thing to do, but with the real teachable moments within the platform would be hugely valuable.

And I think it eliminates, and we'll start to eliminate this generation of parents just, we're hovering over our kids, and, like, we're setting a timer.

And one of the biggest increases, you asked me about, trends that has come our way is this need for parental controls.

I mean, every parent is just asking us, what's the best parental control?

I just have to shut off the Internet.

I have to shut it down. I actually think they're doing it for their spouses.

But ultimately, I would be. But ultimately, they are, you know, they just, we're on our kids.

And what are we teaching them? We're not really helping them learn and get to that executive function.

Have you seen good tools or good technology that helps people do this?

We don't promote third-party parental controls.

It's not really something that I feel like is, or that we really feel like is the go-to.

You know, we're really much more, our philosophy is really about talking to your kid, engaging with your kid, media balance, play for a half hour online, play for a half hour outside.

So that really is our approach. But I will say there are a lot of companies out there that are creating the parental controls directly in the platforms that you're already using.

So you don't always have to get a third-party product.

You know, your tablet likely has parental controls. You know, iOS just launched screen time.

We've got a lot already, and YouTube's got parental controls.

Netflix has parental controls. They're not easily, they're not that easy to find all the time, but they're there.

Yeah, it does fall back to the parents ultimately, right?

It's not something you're gonna start regulating. Yeah, and I think it'll be interesting to see what happens with iOS 12, like with the screen time stuff, if that changes anything.

But the experience that I've had in the kids' media production side, talking to other developers and the things we do, and the things we've done as well, is parents will tell us over and over and over and over again that they want these controls.

They'll even tell us exactly what they want it to do, and then they don't use them.

And this, it just, so many folks have said this.

They like the idea of it, but then ultimately it's a little too complicated to do, it's a little too hard, and they forget to do it.

And I will also add, some people, you know, they need that.

It depends what's going on with your kid.

I mean, it's not just for fun that I'm a parent and want to see every image that my kid is texting.

There may be some other issues where you feel like I need more insight.

So great that those products are available, but I wouldn't say that they are the first place to go necessarily.

And you'd mention as well that people in the technology industry clamp down the most, or let their kids watch the least amount of screen time and kind of interact with apps the least, which is ironic, given that they're the people that's in place to regulate it, right?

Yeah, maybe some people in this room. I can't tell you how often I hear that from people at tech companies who are like, oh, that's great what you're doing with PBS Kids, I don't let my kids use any screens.

I'm like, if you're in the industry that's actually trying to build new technology, I want you to be thinking about how your kids might use it and not in a just, I'm afraid of it way.

I mean, the truth is, kids are in a world where they're seeing the technology around them all the time.

They're increasingly seeing it at school. They're over at friends' houses and they sometimes get to play on their friend's tablet.

I think it's much more important for the parent to be kind of setting up, this is how we use it in our house and these are the ways that we use it and we don't bring it to the dinner table.

We charge it outside our bedroom. Like, all of us do that, right? We don't have it in the room with us when we're sleeping.

But things like that. Trying to kind of set up some healthy habits at home early as opposed to just letting kids kind of navigate their own way because they are going to find all the technology.

Another topic that was interesting you talked about, we'll talk about AI later on in other panels as well, but sort of conversational UX, right?

And kind of think about and see where technology is growing all the good things that it can do for learning and for kids.

What are good examples of that that you've seen out there?

Oh, right now it's really early. I'm not sure I've seen anything that I think is really fantastic.

I think everybody's just kind of messing with it. I do think that there's an incredible positive potential here.

So again, I work with kids who are between the ages of two and eight.

When I started working on content for kids, we were dealing with the keyboard and the mouse.

Like, that was how the kids were navigating.

I mean, if you can think about a three-year-old, they don't know the difference between left and right, like clicking was a big problem.

They don't know how to read.

They don't know how to write. I mean, typing was impossible.

And the technology just keeps getting easier for them to use. The touchscreen was like a revolution in the kids' learning space, for sure.

And voice agents now, we have an opportunity for kids to interact and learn content just talking and asking questions.

That's pretty huge. It means they could navigate into more information about that dinosaur or something really easily.

Now, it's early days.

Alexa and Google aren't asking the right kind of follow-up questions yet to kind of get in there.

And parents are also pretty terrified about this. I mean, we should be thinking really hard about the privacy here.

But I think there's a great opportunity for us to begin to help kids on the learning journey by allowing them to talk and listen.

And Jill, I know privacy is a sort of big topic for Common Sense Media, especially recently.

We do, we focus our privacy work right now on the education space.

There's so much, there are so many apps going into schools directly.

A lot of technology being adopted by schools. And student data privacy is certainly a concern.

Data privacy at home is also a concern. But as we were talking, I don't know that it's reached that critical mass yet where parents are clear what the real issue is or what's being taken or what's being reused.

So everyone's kind of just like, ugh. But we're still moving forward with adopting some of the technology in our homes.

So we're gonna be looking at it a little bit more in next year for our consumer parenting side.

But I would say, I agree.

I mean, parent, or privacy rather, is certainly an issue that needs to be considered.

And it creates challenges when you're building new platforms, when you are looking at privacy and wanna create something.

I think we're all expecting things to function a certain way.

We want it quick, we wanna purchase it, we wanna swipe for it, we want the answer very quickly.

But there may be a little bit of a trade-off when you wanna adhere to certain privacy regulations and policies with the technology that you're going to be getting.

So I think that's a little bit of why parents or consumers at large are kind of like, ugh.

I think we're all very willing to hand over our privacy in more ways than we realize for the ease and quick use of some of these very cool apps.

Yeah, yeah, interesting. So we're starting to wind down here.

I wanted to ask in terms of, if there's one thing that folks in the audience, I'm sure there are a lot of parents in the audience, can take away both from this talk and just from your general knowledge about how best to promote technology and screens with your kids, what would that be?

One thing. Or two things, or five things.

Do you have anything off the top of your head? Well, I already said talk to your kids about what they're doing.

I think it sets up a really good habit very early that media and technology is something that doesn't have to happen in a black box.

It is something we talk about, just like we talk about our school day.

And as kids get older, that's an opportunity to also get into critical thinking questions.

Like, what do you think would have made that game better? Or did you like the way that character was treated?

Being able to really get kids into bigger conversations.

I think the other thing that's really important, so I think I've got two things.

The other thing is that there is no one-size-fits-all solution.

I am asked all the time, I'm sure Common Sense is too, what is the right amount of time for kids to have with screens?

And there's no way to answer that, because you have to know your own kid and how well they are able to regulate and handle media.

So I have two kids. One of them would watch and play all day if I let him have the opportunity.

So I have really specific limits. My other one really is cool just playing for 30 minutes and then wants to put it down and go do something else.

And so we have to kind of approach it a little bit differently. And you have to really know your kid to be thinking about what is gonna make the most sense for them.

I would just agree with both of those things. I think we were just talking about, everybody wants a number associated with, everyone wants a prescription of what is okay for my kid, because we've got our Fitbit with our 10 ,000 steps and we've got a number that we're going towards for everything.

And the AAP, the American Academy of Pediatrics used to say zero screen time for kids under two.

Now, even they in the last year have come back and said, 18 months.

And by the way, quality is really what matters here. So I agree, knowing your kid is so vitally important and being comfortable having conversations with the parents of your kids' friends is also so important, because it'll kind of get you to that place of, oh, okay, we're all dealing with this together.

And that's very personal.

And in your own little community, it's not something that you wanna, you don't necessarily, we struggle a lot with social media and getting a presence online for ourselves.

And oftentimes we're like, let's have people share their cyberbullying story.

I'm like, nobody wants to share their cyberbullying story publicly, that's insane.

So these are personal decisions, personal by child, personal by family.

And so really being aware of what your kid is okay with. And again, not just time, but what are their fears?

What are their concerns? What do you want them to learn about?

There's a ton of opportunity with all of this technology, with media in general.

And it's also just such a great way to get your kids talking, as I mentioned, annoying as that might be, but really engaging with them about their media choices.

You can use that as such a great skill building opportunity and always media literacy, I personally feel is just at the core of so much of what kids are gonna be dealing with moving forward.

So helping them to start from a very early age, really as young as four and five, understanding commercials and advertising, what's targeting them and really trying to crack open the messages that are in front of them in a pretty simple way.

Is a great way to start that critical thinking.

So we can't punt our parenting responsibilities to an iPad, is what I'm hearing.

Well, don't feel too guilty if you do it sometimes.

I'm sure we all do it. So we have time maybe for one or two questions from the audience for our panelists.

Go ahead.

Hi, thanks for the panel. You guys talked about critical thinking and the new media.

How do you think about kids growing up in this post truth age where fake news is everywhere and kids can basically find an answer to any questions they have and not necessarily the right answer?

How do you think about that? Thank you.

For us, for media literacy, we really incorporate digital literacy and news literacy into that.

And we actually do curriculum in schools. We're in over half the public schools in the country and we have incorporated news literacy into our curriculum as well because we do feel that's hugely important.

But, you know, we used to, our old parenting advice was turn off the news when your kids are in the room.

Can't do that anymore. They're getting the notifications on their phone or they're hearing about something on the playground because some other kid got a notification on their phone.

So really being ready to have conversations. I mean, unfortunately, it does fall back a lot on the parents to be ready to have conversations with their kids and get a good sense of what they are exposed to.

And of course, protecting the younger kids.

But that train has left the station as far as notifications are concerned.

So I think arming yourself a little bit and being comfortable.

And I also think culturally in this country, we're very fearful of the news and the news is fearful.

But kids much younger than this, kids, young kids in other countries are much more familiar with having conversations about war.

Whereas we are very, you know, try and keep our kids very protected.

So opening up that dialogue early on is important. Hi.

Whoa, that's really loud. First of all, I just wanted to thank you both have actively used both Common Sense Media and PBS Kids, have preschoolers and a kindergartner.

My question was an extension of the parental controls. I do wanna give my children choice because it's hard for them to not have any screen time.

But do you have recommendations for how to curate?

I wanna provide them basically a safe playground where they can choose content because once they have Internet connection, they can get to everything.

And so it's really hard as a parent to constantly track what that content is.

So any advice you have on using systems or creating a curated space for them to consume content?

Are you specifically asking about YouTube?

That is one of the places. But in general, even if you go to Amazon apps, on the television, you go to one screen, the next show comes up and they can scroll through everything.

And so you don't necessarily know. And I'd love to be able to be in front of the television when they're there.

But if I be doing something else, you can't always be monitoring everything.

Can only watch so many episodes of Paw Patrol.

But Daniel Tiger, on the other hand, I go watch all day.

There are a lot of parental controls, as I mentioned, but you can also like on YouTube subscribing to channels is a really good way to go because then you can kind of curate this walled garden experience.

YouTube Kids is available specifically for that walled garden experience.

You can go in and block channels.

They've given you a lot more control as the user to go in and block things that you're not interested.

Hall videos or toy videos, if you're not interested in that, but you wanna focus them in a different direction.

So those are two examples.

There are other platforms out there. There are other apps that actually have created this walled garden for you.

You can come to Common Sense Media to find those.

And those are just a few things off the top of my head. So we are, we're out of time.

I just wanted to thank Jill and Sarah for a great discussion.

Thank you. Thank you. Thank you. Thank you. Thank you.

I can't do that.

You can't cross your legs? No, I just do this. Well, okay. I'm not sure if that's an important thing.

All right. Now I'm nervous about my body language.

Kelsey Hightower is a staff developer advocate on the Google Cloud Platform.

Kelsey, I have no idea what a staff developer advocate is. What is it? Me either.

So the job title is developer advocate. And if you think about what we classify as a developer, which is anyone that will pay us money for Google Cloud services, right?

So anyone that's creating things, the goal is to advocate for that user.

So not sell our products to people, but understand how they use them, how they should be able to use them, and actually make our products do that, right?

So you advocate for, be the voices of the customer.

What's Google Cloud? Google Cloud?

I mean, I wanna get to something, which is there's something going on called cloud.

It's a rather fuzzy subject. Yeah, so. What is it? There's this whole public cloud movement, right?

So this idea that, like we have freeways, this public infrastructure.

Computing gets to a certain scale where you won't necessarily be able to do it in your basement or maybe in your own data center.

So in Google's cloud case, there's a lot of technology Google has used to build things like YouTube, Google Search, the things that make Google successful.

So a lot of those tools are offered by the hour in Google Cloud.

So when you think about the cloud, it's a very broad definition of terms.

Some people use it as a virtual data center, run my virtual machines instead of in my own data center, I'll run them in Google Cloud.

But I think what makes the Google Cloud unique is we have a different approach to certain problems.

And it's not necessary that we can ship that software to you that you would run in your data center, but we can put an API on top of it, and then you can consume it in very similar ways that Google would.

But let's dial back a bit.

There are other people doing this, right? So I can rent computing power, it's what you're saying, from Google, but I can also do it from Amazon.

So many people might not think of Amazon.

We don't use that word. Well, you may not use that word, but I'm allowed to use that word.

There is another company that sells books that will also rent you computing power.

So, and in fact, there's another one which is in the sort of Seattle area as well.

We can say Microsoft. We can say Microsoft?

Okay, Microsoft then. So there are at least three examples of companies that are selling computing power to people.

And I think a lot of people may not realize that they can now, you can buy arbitrary amounts of computing power.

Yeah, and they're all unique in their own ways.

I think Amazon does a good job by bringing their retail expertise and offering a platform that helps make them successful.

You see this with Microsoft, where you've taken the mail server out of the enterprise's four walls, and now you host it as something like Office 360.

And the whole idea here is to make it very easy to adopt these products without necessarily having to become an expert on how to install it, how to upgrade, how to secure it, or hiring a whole IT team to do it.

So you kind of see Microsoft's expertise being offered by the hour, whether it's Windows Server or any of their unique products.

And in Google's case, this kind of, we have this new term like cloud native.

What does it mean to be able to compute anywhere in the world that you want to do, and also have that response?

If you ever go to and it was down, you would probably think the world is about to end, right?

Like Google cannot be down, right?

So in order to be able to do that, there's a high investment cost on infrastructure, the network cables.

We have undersea, underwater sea cables because we like to have this closed network for performance.

All of that is available by the hour.

So that's Google's unique spin on the cloud. And prior to this cloud thing, anyone who wanted a lot of computing power, they actually buy physical machines, right?

They go and buy them, install them somewhere, run them, all this kind of stuff.

So how many people have you put out of work by doing this? Because they don't want to buy.

I don't think the cloud can put people out of work.

If you think about maybe, there's always this transition in technology that we can imagine that maybe some people won't be able to do the boring parts of buying a server.

I've worked in a data center. I've cut my finger on the racks or putting these things into a server.

But we end up doing those, we boost the number of people who can write their own software.

We magnify the number of people who can actually run globally, right?

A lot of small businesses couldn't imagine building 50 data centers, right?

There's not enough affordable power for that to happen. So now that same company's able to focus on their product.

I think in the cloud world, Netflix is a great example of leveraging cloud to become a major player in the video streaming space.

So I think you end up saying, maybe you won't have thousands of people racking servers in a data center, but you will have way more people using things like machine learning or any idea that they have without investing in infrastructure first.

Which is interesting, but programming is really hard, right?

In the session I did with Darren and Sidney, I had this quote from Wilkes in the 50s about how he suddenly realized he was gonna be debugging for the rest of his life.

I mean, it doesn't seem to be getting any easier, does it? In some cases, it's been getting easier in terms of libraries.

If you wanna do any security or cryptography, more than likely you won't write that from scratch.

You're not gonna go and learn all the mathematics behind that.

You're gonna import a library.

But what is difficult is that there's almost hundreds of languages to choose from.

There's so many ways of going about this that for most people, the patterns aren't quite clear.

So most people get an idea that they wanna do, and somehow you see yourself building the same thing over and over.

How many times are you gonna build a user login form?

How many times are you going to build a email thing that sends people notifications?

So the complexity from that regard is keeping some people from being able to build ideas and ship them.

So what we wanna do with this whole kind of cloud thing is move into a world where we can eliminate some of those.

So you see frameworks where I think Ruby on Rails was a good example. You wanna build a website or an API, you can run a few commands and it would bootstrap most of the boilerplate for you.

But as we catch up with the tools, new things come out like mobile or machine learning.

And then we kind of reset the clock again for the number of people that will be able to do that day one.

So what we're starting to do is see people come out with these products that have the user in mind.

How broad can this be adopted on day one without going through the whole cycle and learning curve?

And the user here you're talking about is somebody who's writing some amount of code to put these things together.

Because there's another thing going on which is this maker movement with all these hardware devices.

And it almost seems like it's easier to do hardware than software sometimes.

Like hey, if I solder these two things together, this LED will light up or I can connect this GPS thing.

Why has that happened and software has remained so very difficult?

Yeah, like in the physical world, there's a lot of rigor and discipline required to build something safely.

I was on a city tour in London yesterday.

And you see some of these buildings are 900 years old right next to a building that's still in progress of being built.

And some of these buildings will actually outlast some of the newer buildings.

They will come and go. So the rigor and discipline needed to build physical things.

But you also see some advancements there like a 3D printer.

Like there are people printing parts at home, replacement parts for something that's broken that's plastic.

You can do that in the hardware space.

But somewhere this didn't happen for the software space. You still need to be an expert to write software, right?

You still need to know how to edit text.

You still need to learn a programming language and all the syntax. So we hope that we can actually get to a point where most people can just take an idea and represent that in a way that's a little bit more scalable.

And there'll be custom sauce on top, right?

There's gonna be unique things you have to do. But we need to get to the point where people can print these ideas without becoming a software developer first.

Haven't we tried that? Like visual programming languages. I mean, every level of high level language we keep saying it's gonna get, you're gonna be able to drag and drop stuff and Yahoo pipes and all this kind of stuff.

We're failing, aren't we?

Yeah, and I think that's because the technology moves so fast.

There will never be a checkpoint where we can drag and drop and become Netflix overnight.

That doesn't make sense. Because the bar's gonna continue to rise.

So the level of stability you need to get to in order for drag and drop to work, that means we have to stagnate.

We're in no position to stagnate, right? Things are still moving.

It's still early. We look at the entire computing landscape. So I think what you're gonna end up seeing though is the first step is, we're thinking about this world called serverless.

And it's kind of this new term, but it's very confusing because there's still servers underneath.

We're not gonna get rid of the actual computers.

But what we wanna do though is, when you have an idea, we don't want the first thing for you to think about is the server, right?

We want you to think about writing just enough code to get the business use case done and then deploy it somewhere.

And all of this, where it should run, how it should run, the security aspects, there's been so many best practices over time that we've checkpointed into libraries, we've checkpointed them into frameworks, and now we're moving into a world where they'd be just come platforms.

So given a platform, maybe it's not quite drag and drop, but you can go from writing 2000 lines of code to writing 50 lines of code and leveraging some of the same power and getting some of the platform pieces for free.

How important is open source to this whole getting people more involved, building this serverless world?

Yeah, so open source is kind of this unique thing where software is pretty easy to produce or once it's produced, it's easy to share it.

And for most of us, the ideas in software, you kind of need to see it because you may need to manipulate it for your own needs.

So you kind of need the source code, right?

It's like getting the raw materials for something that you assemble in your home.

But the other byproduct of open source is this idea of sharing, right? So there are people that sell software for money and in order to make people pay for stuff, it has to be hard for them to get it for free.

But in the open source world, we saw this transition from white papers, ideas, to actual implementations that you could just use as part of your own software.

And now that we have this kind of world of sharing, every company that has expertise can now share their unique space in computing.

So if you're an insurance company, maybe you share your algorithms for determining if a car is totaled or not, where a drone can fly over.

And you can make that available as a library. Once we start to have this level of software, then the game changes, right?

And it becomes a platform play. So we're not necessarily selling the bits of software, but they end up becoming all of these Lego bricks that if you have imagination, you can pluck from the open source world and build whatever you want.

So that's kind of where we are. The open source world represents the new Lego bricks that you can use to assemble something unique that we haven't imagined already.

Well, there's still some secret Lego bricks, right?

Because I can't download the whole of Google and run it at home, right?

You won't give me the whole thing, if I ask. No, we won't, because we need you to pay.

It's how I got here. Fair enough, fair enough. People do need to pay.

So we talked a little bit earlier when we were chatting before about event -driven programming.

This is a very, this is a different way of thinking about programming.

Just take us through what that means. So the event-driven programming, so serverless is this new term.

But if you break down serverless, the things that allow us to write less code is because we now have, we can now react to things versus listening for and waiting for a user to show up, not quite sure what we want to get.

And in a world where you're not sure what you want to get, you have so much logic.

If the user wants to do this, then execute this bit of code.

If the user wants to do this, execute this bit of code. And in the serverless world, backed by event-driven programming, what we're going to do is just have your snippet of code at the very lowest level of the function.

And what we're going to do is map events to it.

And an event would be, there's a file that has been uploaded to this particular piece of storage.

We're going to invoke this piece of code in response.

And you think about that world and you say, well, where has that worked already?

And you think about your car. Your car is very sophisticated. And the people that build sensors for like your tire pressure, right?

Does my tire need more air or not?

And thinking about that world is, there's a piece of hardware with a very small bit of code that submits an event.

Your tire pressure is not where it needs to be.

But it's too complicated to know who should respond to that event. So that pressure's only job is to alert the system to say the tire pressure is at this particular number.

Now, maybe the computer in your car decides to light up something on your dashboard to give you an indication that there's a step that you need to do.

For every other element of your car, the door is open, the trunk is open, the windshield wiper comes on automatically.

These are all event-driven systems.

And it's very easy to come up with new events and tie them together without coordinating with everyone that it takes to build this car.

And that's where it becomes super powerful.

And then you get a coordination problem, right? So one of the things I'm skeptical about, I'm gonna slightly argue with you, is that sounds great.

There's all these events coming in. And then we have something like, well, we need to make a decision if the door is open and the seatbelt is plugged in and there's a child in the back.

So suddenly you get into this fundamental decision-making part, whether there's events or whether you're doing it.

And that seems to me the really hard part about programming is what do programmers do?

They hold this idea in their head of these sequence of things going on or multiple things happening at the same time.

And there are some people who are kind of special who like that.

And there are some who don't. How do we help people understand that part of the essence of programming, which is that ability to close your eyes and imagine all this stuff happening at the same time?

Yeah, so I guess when we write traditional software, we think of this as a state machine, right?

If this happens, then you progress to this state. If you're in this state or there's an error, then this is how you handle it if you're doing a good job of handling your errors.

But the problem is that's so much boilerplate in order to represent that.

In the serverless world, we're only designing functions. You can imagine designing this state machine visually.

So here's the bit of logic to handle if the car seat is open.

Here's the bit of logic to handle if the front door is open.

Now, if these two things aren't true, then maybe the car doesn't start. And it's much easier to draw that if you know that this thing only does these.

And then over here, it only has to check for one piece of logic.

Are these two things true?

And I think that's where we wanna get to with these platforms where everyone, the number of people who can express these complex relationships.

You see this in like the shipping industry.

If you wanna ship a package from the US to Australia, think about the route it takes.

And the only thing we get is this label. And you scan the label when you drop it off.

And I can use that tracking number. And along the way, the person who puts it on the airplane may have no relationship with the person who puts it on the ship, has no relationship with the person who puts it on the car.

But from my viewpoint, all of these events, all of these interactions create events where I can track that package.

And they can ensure that it gets there on time.

So society runs this way. We're all kind of event -driven.

If someone taps you on the shoulder, that event invokes you to turn around and give them attention.

And it's probably the only way to really scale a truly complex system where no one person can hold all the interactions in the head.

So now we're talking about describing logic and response to relationships.

Let's just go back and dig a little bit deeper into serverless.

So the book retailer, Microsoft and Google, will rent you computing power.

But I'm too stupid to configure that computing power.

It's too complicated. So this idea of a serverless is that I don't even need to know about the underlying hardware or the operating system or how it's running.

I write some code and it runs somewhere. It happens. Yeah, so if you were thinking about the early days of cloud, we're going to virtualize your data center and make it very convenient for you to utilize it.

If you have 100 servers in your data center, maybe with the click of a button, I can give you 100 servers on my cloud platform.

But this is a very server-centric world. And that means we need people who understand the server.

So if your whole goal as a company was to build software to give people directions around the city, but before you can do that, you have to configure the server.

You have to make sure that it's secure, what operating system.

Should you put it in this particular region in the UK or should it be in the US?

All of these decisions have nothing to do with the customer and the thing that you're building.

And also, compute isn't infinite.

So no matter how much money you have, you can't just go to every cloud provider and say, give me a million of these servers.

Not enough power.

And most people don't utilize them very efficiently either, like your car. Your car probably sits in your driveway way more than you actually use it.

There's a cost to that.

And we get traffic jams. So the bandwidth of how many of these current servers that a cloud provider can provide is actually not infinite.

So what we have to do is you have to transition people off into a new world where we say, how much compute power do you need for the task at hand?

And let us choose the best server.

So that way, if one of the servers were to die, then we can easily move that piece of logic to another server.

So you don't have this tight coupling with your app and this server.

And it just makes everything easier to deal with and we can innovate faster.

I can change the server without your permission much faster if you don't wedge yourself to this particular architecture or this particular machine type.

So you're saying that the underlying cloud provider might take my computer and move it to somewhere else without me having any idea that happened.

And I can give you a better price.

So maybe as a provider, I know the status of my particular fleet.

When you go to ship something, they give you a price based on what's cheaper for them.

And given that, we can actually see the price of compute go way down.

What we want to do, you shouldn't have to have a million dollars to be a global scale company using technology.

Maybe we can get it to the point where it's cheaper than what you spend on coffee per month.

But the only way to do that, we need to decouple you from these things that are wasteful or too expensive to maintain at their current capacity utilizations.

Haven't we been here a long time ago with, you know, IBM would sell you a giant mainframe and they would run multiple jobs on it and you could submit and it would all magically happen within the mainframe.

This is not new, is it? Yeah, it's unfortunate that the mainframe has been given a dirty word in most enterprises.

It's like, oh, it's legacy technology.

It isn't fit for the future. And the thing is, it did a job very well.

If you think about the properties of a mainframe, you had this system that most people didn't know anything about infrastructure or hardware.

You put data in, results come out.

Rinse and repeat, and you can have everyone in the business world use the system without having an operations team bigger than the business team, which is not the case for most companies today, right?

IT is bigger than that.

And then things change, you know, networking, mainframes were pretty expensive, the demands on power, where you put them.

So we kind of went through this kind of backwards world of everyone has their own piece of boxes.

They're not very efficient.

They're not very redundant. Now we're moving back to this cloud world where it feels like this mainframe world, there's this time-sharing machine where you put your data in, data comes back out.

So I think we're definitely coming full cycle, but the mainframe didn't evolve fast enough, right?

When the Internet comes around, the mainframe wasn't the right device.

And other forms of computing, like mobile, right?

You're not gonna put a mainframe in your pocket even though you have one now.

It just wasn't the time back then. All right, so I put in the title of this session, Kubernetes, what is Kubernetes?

So Kubernetes is a very low-level detail, but it represents this checkpoint, and maybe some of the evolution since the mainframe days.

I guess since I've been doing this, you know, you have the Internet, and then there's standard protocols in the Internet, TCP, IP, but we don't really talk about those things.

We just say the Internet.

And then you have the big vendors, IBMs, HPs of the world giving us these operating systems, right?

This is your classic Unix operating systems. But then we see the open-source world come out around this time to say, we can make that free by collaborating on this layer of the software.

So this is the operating system. If you're running an Android phone, Linux is running that particular device.

If you have a server, Linux is probably powering that.

So now you start to see the democratization of the operating system.

Now Kubernetes represents this other checkpoint.

So a lot of the larger businesses like Internet-scale companies like Google to Facebooks and Amazons of the world, we had to do this by taking these cheap pieces of machines that fail all the time, unlike the mainframe, but make them work, and you get into this world of distributed systems.

We can go on and on about the complexities of manage little devices that fail all the time over multiple continents.

There's a lot of research in this area around distributed systems.

But Kubernetes comes along and takes all of that knowledge and turns it into a software package that you can use to install on those machines that you currently have to give you that ability to think at that high level of extraction.

So Kubernetes is just basically this checkpoint over time where it's no longer the secret sauce of the Internet-scale providers.

It's something you can actually download, experiment on your laptop, and install on your own server.

So now that we're at this checkpoint, it also represents the first time that we're saying that the machine isn't the abstraction.

It's your application. So Kubernetes runs containers.

And when you think about packaging your app as a container, you don't think about the machine.

You think about the app, and everything is required to run it.

And then you give it to Kubernetes, and it makes the decision on what server to run on.

So this is kind of that precursor to getting to a world where we can decouple the machine from what the user wants to do, and that's what Kubernetes' role is.

And we went through a period where there were virtual machines, but they were still kind of machines, really.

You're still worrying about the operating system, and there was virtualized hardware.

But what you're talking about is, I build this app, and I just throw it into this Kubernetes thing, and maybe Kubernetes decides to run three of them because the workload is higher, or it moves them around because a machine failed.

But that's all transparent to me as the developer.

Yeah, so I always describe Kubernetes as the thing that the best system administrator would do.

So if you didn't have Kubernetes, and you think about your IT staff, and you ask them to run this application for you because they're your abstraction to your infrastructure, your IT staff, and if you look at how they make a decision on what server should run this application, they do this act of what we call scheduling, right?

This machine is best fit to run this software because it has enough memory or enough CPU, or we're not going to deploy it in this region because it doesn't meet the regulations.

So most of the time, we've regulated this to humans to make this scheduling decision, and they use this advanced technology, a spreadsheet.

This is where they keep all of this stuff, right?

They label this, and it's very primitive to watch people work in this fashion.

So what we've taken is a lot of those algorithms, and we created schedulers.

And now what we need people to do is articulate what are your requirements, and you give it to something like Kubernetes, and it makes those same decisions in, of course, fractions of a second.

So now you kind of level up the number of people that can do this.

And something fails, or it fixes it for you. And this is all super easy to use, right?

Kubernetes is dead simple. It's not. No, no, no, no.

We often, in our industry, confuse simplicity. So the thing is, I think there's probably a total set of complexity that you can't get rid of when it comes to computing.

It has to live somewhere. If you have one machine, then that one machine is super simple.

You don't need any of this distributed systems business.

You can just focus on that one machine, but then all the complexity is in your head.

What happens if the machine goes down? You have to hold all of that in your mind.

Or you can take it, and you put it in a software like Kubernetes, but then it becomes pretty complex in order to handle the case of this machine died, but only choose this one if it matches the criteria.

And all the components that make even that system itself that's trying to make your system highly available also be highly available.

Once you get to this level, then it's very hard to operate this platform.

It's like building your own car. When I had my first car, I was able to change the spark plugs.

I even swapped out the engine in my mom's driveway. But if you look at a new car and you pop the hood, you're like, ah, no, they don't even sell the tools for this.

But the car is so much more complex, in many cases, way more reliable and safer than the cars from 20 years ago.

So that's kind of the trade -off.

As technology gets more advanced, it's gonna be very hard for the average person to fix it themselves.

All right, listen, time has flown by. I thought we were gonna be, we could sit for another hour, I think, and talk about this stuff.

Let's take questions from the audience.

Awesome. Thanks very much. We'll get some microphones running around.

This hand went up fast, so this is gonna be a great question.

Yeah, this is a question that was very burning from the beginning. Question about the complexity.

I think oftentimes we see when a new technology comes along that it swings full cycle to very hesitant banks who are like, oh, no, this is too dangerous.

And then suddenly traction comes, which I feel is right now Kubernetes, and it even overswings as, okay, let's do everything, even things that you should do with Kubernetes.

And I would be interested to see what do you think should not be done with Kubernetes.

For a long time they said state or would put in Kubernetes.

Now I see people doing that as well. So where do you think is the boundary between what Kubernetes can do and what it should not do?

Yeah, what he described was a hype cycle.

Anytime new technology comes out, no one understands it, but they start associating problems that it can solve.

And we saw this with Kubernetes, right?

It actually came out of the gate, and it solved that distributed systems run my application problem.

And when you look at it, and I see this with all technology, people try to take the old world and move it forward and attach it to the new technology.

So you mentioned things like state, right? And state is pretty hard.

So state is this idea that if I have a website and someone uploads some photos, I want those photos to be available for every user that comes to see those photos.

And you have to store them somewhere. Now, maybe for the last 40 years, we've been using this idea of there's only a single server, right, this whole POSIX interface, there's only one server.

And if you land on that one server, then it's pretty straightforward.

But what happens when there's 10,000 servers?

How do you distribute all of those images? That's not a problem that Kubernetes can solve.

But since Kubernetes solves the other distributed problem of keep this application running globally, then people look at the state problem and say, why not treat the state problem like the compute problem?

And they have different consequences.

When you lose someone's bank data and their balance is zero, that's a problem, right?

If the application isn't there, you just try again.

Users know there's like a quick hotfix for refresh. Ah, it's down for money, just try it again.

So I think people look at Kubernetes as it will also solve the data problem in a magical way.

And I'm not convinced of that because in distributed systems, we actually have different ways of storing data.

But most people aren't necessarily willing to adopt them.

We have object stores now. So instead of writing to or assuming that you can write to the local machine that you're on, you have to believe that there's going to be some other endpoint that's designed for global storage and failures, the same thing Kubernetes is good at, compute.

And you can give this object store, like people that use Google Drive where you upload a document that you share with everyone.

You have no idea what servers those are on.

But the underlying storage mechanism just says upload it here, and you can share URL with someone else.

And when they hit that URL, no matter where they are in the world, we replicate that data and make sure that they get the data in a timely manner.

You can't do that with the old design. The old design has none of those semantics.

So for me, I don't know if Kubernetes can solve all of your problems, okay?

So what we started to do now is Kubernetes is starting to stabilize and draw on the line in the sand.

So you're starting to see new platforms being built on top of Kubernetes.

In this talk, we talked about serverless. So Kubernetes runs your application in these container images, but you still write all the code.

We can have people adopt serverless platforms by saying Kubernetes take care of the distributed systems problem, but we'll build new frameworks on top.

But that will never be the responsibility of Kubernetes.

It will just be the foundation for people that want to build these new platforms, and then we'll evolve again.


Thanks so much for being on stage with me. We're between this lot and coffee, so I think we better get off stage before we get lynched.

Thank you so much for being here and telling us about these things.

Round of applause. ♪