The day after the keynote, Daring Fireball's John Gruber took the stage at Mezzanine in San Francisco with not one, but two special guests from Apple: SVP Worldwide Marketing Phil Schiller, and SVP Software Engineering Craig Federighi. Below is a full transcript of their remarks, and check out the full Talk Show audio and video on Daring Fireball!
KAFASIS
Ladies and gentlemen: Welcome to Mezzanine! Won't you please take your seats and silence your cell phones.
John Gruber
[chuckles]
KAFASIS
Daring Fireball Productions, in association with The Daring Fireball Company LLC, is delighted to welcome you to a Daring Fireball presentation of The Talk Show — Live from WWDC 2016! And now: Won't you please welcome your host, Johnnn Gruuuuber!
[applause]
John Gruber
Thank you Paul! I actually needed that reminder; my cell phone was not on silent. Thank you, Mike.
I will start by thanking our sponsors. This is the fifth year I've been doing the show, fourth time here at Mezzanine, and I think that Mailchimp has been sponsoring our bar all along, and if not that, then at least as far back as I remember. So, just in case it isn't clear, the bar is on the house, it's on Mailchimp. Mailchimp, if you guys don't know, they are — do email newsletters, like my friend Ben Thompson over at Stratechery, those go out through Mailchimp. They also have a bunch of new features, stuff that integrates with with online stores, and integration with just about any online store platform that you might be familiar with. And then you can get your customers to get email when products that they're interested in are available, or... whatever.
[audience laughs]
Great company! If you need to send email, go to Mailchimp.com. And, please, let's hear it for them for the open bar!
[applause]
Also back with us for the fourth consecutive year as a sponsor of the show is Microsoft. And at four years, it's not even like a "Whoa, that's weird, Microsoft sponsoring..." Nah! It's like awesome. And it makes total sense.
They have this website. It's going to give you so much more information than I have time to give you now. Anydevanyapp.com.
That's the message that they're trying to give: That any developer, whether you're working on mobile or the web, for any type of app — if you need cloud services, it's now called the Azure app service. If you need that kind of stuff, go check it out — their website has so much information.
Here's a funny thing: They had the same website last year, but instead of anydevanyapp.com, I said anyappanydev.com.
[laughter]
And we are, in fact, streaming this live. And, y'know, the show went on, and in the meantime, I gave out the wrong URL for a pretty pricey sponsorship! [laughs] And what happened was there was some kid in Australia who was watching the live stream who quick, like, jumped on and registered the domain.
[laughs]
True story! This is an absolute true story. If you guys see Matt Hansing, he's here representing Microsoft; he's about this [gestures] tall, him and [Craig] Hockenberry are gonna have a fight after the show's over. You can ask him, he'll vouch for this.
So they got in contact with him, were like — oh, man, that's Microsoft now — "We better get this domain," and it was already gone. And they contacted the kid, and they were like "Oh, man, this kid is gonna, y'know, he's really gonna let us have it." And he was like, "Well, one of those Xboxes would be nice!"
[huge laughter from the audience]
So they sent the kid like, a box with an Xbox and all the cool stuff that you could possibly imagine that goes with an Xbox, and they got the domain. So I think it's safe that you can just go check out the information from Microsoft. Go to Anyapp... or anydev... dot com. [laughs] No, anydevanyapp.com! Microsoft, great sponsor.
And then last but not least, we have one more sponsor, this one's new. And surprisingly, this is the thing, because we think Microsoft, how are you going to go bigger than that. But this is actually one of the few — I mean, I'm guessing maybe three or four corporations in the world with a larger market cap than Microsoft.
It's Meh.com.
[surprised laughter.]
Meh.com is the store that I would run if I were going to run, like, an online store. And let me be clear, I have absolutely zero interest in running an store. It seems like a [laughs] seems like a terrible job. And a lot of hard work, and I don't like either of those things.
So, yeah, I'm not going to run a store. But if I did, it would be like Meh. And here's the way Meh works. They have one product a day. That's it. You don't even know what it is. You have to, like, go there at midnight and find out what they're selling today. One thing, daily deal, usually at like, an unbelievable price. I've said this before: I'm half-worried that they're, like, stealing these things — and I don't know if me endorsing it like this makes me complicit in a crime, because when you're selling, like, a $120 stereo for $14... usually it's like that scene in Goodfellas, where they're selling cigarettes out of the back of the truck.
[laughter]
But what they really do, the other thing they do, is they concentrate on making everything real funny, the descriptions of the products are real funny, they have funny videos every day, and I really do get the feeling that they'd be happy if you just go there and check them out every day and you never buy anything.
That's like the gimmick, or the thing, it's like: Here's the product. Buy? Or meh? And you can just, like, type MEH, and then they're like, well, that guy didn't like that. So my thanks to them.
On the guests for Talk Show 2016
John Gruber
So, last year was a little different than in previous years, because we had an actual special guest. What happened was, the back story on it, is it was a week before WWDC, and I still hadn't asked anybody to be on the show. And I was putting it off, because I kind of had it in my head that I kind of wanted to see if I could get Phil [Schiller, worldwide marketing at Apple]. And I put it off, because I didn't want to hear no. And it was like, a week before, and I was all "Well, this is ridiculous. I'll just ask."
And so, I sent an email to Steve Dowling. And I said: "Look, this is probably ridiculous, and so just feel free to say no. But: I do this show every year, and I think it would be really cool, I think it would work really well if Phil Schiller came on, and the day after the keynote, and we could talk about it, and nerd out, and go into detail that you can't get into in a keynote."
And he wrote back, and all he said was "Not ridiculous. Let's talk tomorrow." And next thing you know, a week later, Phil Schiller was screwing around, not coming out behind the [laughs] curtain, and making me wonder whether, like, maybe he went to the bathroom? Maybe we miscommunicated on what the cues were going to be. And it was GREAT! I mean, I don't know how many people were here last year?
[big cheers]
It really was great. It was the best time I've had on stage in my life, and then I watched the video, and I didn't even really die watching myself. I was like, "Oh, this is actually pretty good!"
And it ended, and it was a big surprise, we kept it under wraps, everybody seemed pleasantly surprised and it just made it all the more fun. And then, the show's over, and I go backstage, and people are like, "Wow, that was great, I can't believe it, that was amazing, that was amazing." And I start meeting people, and it was about three minutes — three or four minutes after the end of the show — when the first person came up and said: "Well, you're really going to have a hard time topping that next year!"
[laughter]
Annnd, I thought "Wow! That did... not occur to me, because this... this week has been a blur, like, I really just asked a week ago, and then we set this up, and I've been thinking up questions, uh... and you're right!"
And there's only so far up I can go, you know, there's only so many different ways that we could go up. And so, one of these years, it is absolutely going to be the case that it is not as good a guest as the year before. I mean, one of these times, it really is going to be John Moltz coming out.
[laughter and some awwws]
And that'll be great! And we'll have a good show. I mean, there might be more people leaving to go to the open bar mid-show — which you can do by the way, please! Really, run up a good tab, we're good here.
But! This is not that year. This year, I think it's a little better.
So, this year, how do you top Phil Schiller? Here's how. Ladies and gentlemen, please give a warm welcome to... Phil Schiller...
[bemused groans from the audience]
And! Craig Federighi.
[Huge applause and cheers as Federighi and Schiller enter, shake hands, sit on the couch.]
Craig Federighi
Wow!
[Audience hoots and hollers. Federighi laughs.]
John Gruber
True story. Got a text message about 45 minutes ago, an hour ago: "Do you guys have any food here?"
[Federighi chuckles.]
No. We don't. We have lots of booze, but... So when they get here, here's what Craig... Craig has a boxed lunch from WWDC.
[laughter and whistles]
And that's what he's eating! And friend of the show and announcer Paul Kafasis asked him, "Is that a WWDC boxed lunch?"
And the answer is:
Craig Federighi
It aged well.
[chuckles]
John Gruber
It was an old WWDC boxed lunch.
Craig Federighi
[laughing] Yeah.
John Gruber
So, let it be said — Apple does eat their own dogfood.
[Laughter, groans from the audience.]
Craig Federighi
True that.
Phil Schiller
I can absolutely validate that for over 20 years of doing surveys from WWDC, every single year, the number one complaint is the food. And so, we resigned ourselves to the fact that if that's the worst thing that comes out of WWDC, all is good.
[Craig laughs.]
It's tradition, so...
On the keynote
John Gruber
So I always start the show, I started it the same way last year: I always ask the guest "How do you think the keynote went yesterday?"
Craig Federighi
Good audience. [laughter] Great crowd.
Phil Schiller
Great presenters.
[wry laughs]
Craig Federighi
[gesturing to Schiller] We were missing one!
John Gruber
Once again, Phil Schiller was not on stage at the keynote. This is becoming a new tradition.
Phil Schiller
I was teasing with Tim that between Craig and Jeff Williams, and now Boz[oma Saint John, head of Global Consumer Marketing for Apple Music and iTunes], I don't meet the minimum height requirement to present.
[laughter]
John Gruber
But, Craig, I honestly thought — I spoke to you yesterday, briefly, and I told you I thought you did an amazing job. I mean, how many people thought Craig did —
[big cheers]
Because you — it's not just that you're up there and you're covering stuff, but you covered, like, three hours of stuff in 90 minutes, or however long you were on stage.
Craig Federighi
[pauses] Yeah.
[laughter]
It was a lot, yeah. I mean, the team did a tremendous amount of work, and we try to pack it all in.
John Gruber
Well, the article I saw on the Ringer did a — I don't know if you saw this — I'm not going to go into detail on the article, but the headline was "Apple's Craig Federighi is perfect."
[laughter]
Phil Schiller
I read that article, and I can only confirm that it's half-true.
[laughter]
On overriding themes
John Gruber
So, I didn't think about it yesterday, but today it occurred to me that there sort of was a recurring, overriding theme in the announcements yesterday. Which, in broad strokes, was that you guys have opened up a lot of stuff to third-party developers that was previously reserved for Apple's first-party code.
Craig Federighi
Yeah.
John Gruber
Quick list, and I might even miss some. CallKit, so that VOIP apps can get the same Lock screen privileges for incoming calls as the Phone app and FaceTime, which took years.
Craig Federighi
Yes.
John Gruber
Messages, so that WhatsApp can, you can specify a contact. When I text Craig, default by going to WhatsApp —
Craig Federighi
That's right.
John Gruber
— instead of iMessage?
Craig Federighi
That's right.
John Gruber
Siri API, iMessage apps...
Craig Federighi
Yeah.
John Gruber
Maps extensions.
Craig Federighi
Yep.
John Gruber
And even non-Mac App Store apps can now use CloudKit, and a bunch of other iCloud stuff.
Craig Federighi
Yep! That is true.
[big whoops for that]
John Gruber
Is that a coincidence? Or is that a strategic part of the plan for this year?
Craig Federighi
Well, with iOS 8, we started that with extensions, you know, opening up the Share Sheet, for instance. For awhile there, it was, if we didn't build it, it couldn't be in the Share Sheet. And so we had to build a Twitter interface ourselves, and a Facebook interface, and as of iOS 8, we started having extensions for extending the system with sharing, widgets... And so we built a lot of the technology with XPC services, if you folks know what those are, and autoprocess UI, and all the building blocks to make this possible.
And this year, we really felt like, uh, giving the developers more and more opportunities to let users do what they want to do across all these experiences. It was, y'know, a way that we could make the platform better for all of our users, so... yeah, it all came together nicely. With Siri, as well.
John Gruber
And a big part of it, it seems to me, as the platforms (plural) evolve — because it's definitely — especially iOS and Mac — what it means to have an app is more than [what it is] on a Mac: Okay, you launch an app, and a window opens up, and you are in this window, and it's yours as the developer. Where, on iOS, it's a little simpler. It's like, you get the screen.
But now, to be an app that's really taking advantage of the best and newest that the platform has to offer, you need to be inside other apps. Widgets inside iMessage.
Craig Federighi
Yup. I think that just makes sense for mobile. I mean, if you have an app, and the right place to interact is on the notification on the Lock screen, and you don't want the user to have to unlock the phone and launch your app in order to get something done — or invoking your app with Siri is going to be the quickest path to getting something done — we want to make that possible.
And so, I think that's what you're seeing here, as well as, what you say, inside of Maps. If you want to book a ride, or you want to get a restaurant, or any of those things, it's going to just be a quicker and smoother flow if you're integrated into the place where the user started instead of requiring switching around.
And so this is opening all that up, and I think developers are going to do a tremendous number of things with it that we didn't even envision. It should be an exciting year.
Phil Schiller
It's also just an evolution of the success of the app model, right? I mean, apps took off, been wildly successful, with this amazing software process, and then you want to have apps in your Maps, you want to have apps in your Siri situations, and you want to have apps in your messaging... and so we like apps, we like them everywhere, we want to use them in many places, so to me it's an evolution of what's going on with apps in general.
On XPC, bundles, and new technology
John Gruber
And you [Craig] mentioned XPC. And I know this is a fairly, fairly nerdy crowd. But I do think it's a years-long shift, where... in my layman's terminology, XPC is Inter-process communication, and it's a way for different processes that can be sandboxed and all of the privacy and "hey you, this process can't diddle with the data of this process without having it in a shared location" — that they can still communicate with each other in a rich way.
Compared to the old days, the Mac OS has always been extensible, and whether you want to go back to the Classic Mac OS with INITs), or the Nextstep days with...
Craig Federighi
Bundles, yeah.
John Gruber
Bundles, and input managers, and...
Craig Federighi
Mmmhmm!
John Gruber
Remember in the early days of Mac OS X, when we had the haxies, and the input managers....
Craig Federighi
Oh yeah.
[laughter]
John Gruber
And that was — in laymen's terms, the fundamental difference is those were ways to extend apps officially or unofficially, where the extension code was running within the process.
Craig Federighi
Yeah, and from a stability point of view, and a privacy point of view, really bad news. So, we started years and years ago, with Mach messaging, and on that, we built XPC as a form of remote procedure call, or an asynchronous messaging, structured messaging thing. But we then created what we internally called XPC Containers — which are really what you now think of as XPC services, which are the ability to package a whole bunch of code, and let the system manage launching that code, tearing that code down when it needed to, but exposing services in that way.
And that turned out to be really important — even internally within the OS! We were using it for quite awhile within the OS, before it was exposed as a mechanism for third-parties because it allowed us to set different security boundaries around different — this is really getting nerdy, but —
[cheers]
John Gruber
Nah, this is good!
Craig Federighi
— But around, uh, [laughs], because if you're going to go load some image format, even, or run a doc, run a Spotlight converter or something, that's going to run over all your documents, you want to make sure that if that thing crashes, it doesn't crash the overall process or Spotlight index or app, you don't want it to have any more access than anything but the one thing it's supposed to have to do the job.
So this was all part of our security and sandboxing architecture, but then, with iOS 8, we saw the opportunity to combine that with, essentially, remote views, the ability to say that the UI that you see on screen that looks like it's all from one app is actually composed from the main app, but also one or more XPC services serving UI into that, and we manage all that. And that gives you this single experience, but where all the stability boundaries, and the security boundaries are in place. And that's enabled us to take this extensibilty model from something that was really hacksy-prone in the Nextstep, and well, nit —
John Gruber
Yeah, nit was...
Craig Federighi
The old days. And make it much more stable. And so that's been, now, a building block for all these things that we're doing. And iOS 10 was just really stepping on the gas on the places where we could do that that made the biggest difference in user experience.
On the removal of stock (and Stocks) apps
John Gruber
One of the most surprising changes, and again, I think that this is in the spirit of openness, or flexibility on Apple's part, and relinquishing control that previously wasn't relinquished. And it surprised me, that you can now remove a whole bunch of the default apps on iOS from your home screen.
Craig Federighi
Though you would not want to.
[big laughs]
You have the freedom — just knowing you have the power that you'll never use, it's...
John Gruber
It's one of my favorite features on the What's New site. I love the page, because it even goes out of its way to say "Y'know, because of all the compression that we use, and the techniques that we use, and the shared frameworks, they only take up 150MB.
[laughter]
Craig Federighi
Yeah! Well, okay, so, this is true. This is true. We should be really, really clear on exactly what this feature is and what it's not. Because it's not everything you might think it is.
So what it is is, you are removing... when you remove an app, you're removing it from the home screen, you're removing all the user's data associated from it, you're moving all of the hooks it has into other system services. Like, Siri will no longer try to use that when you talk and so forth.
We're not actually deleting the application binary, and the reason is really pretty two-fold. One, they're small, but more significantly, the whole iOS security architecture around the system update is this one signed binary, where we can verify the integrity of that with every update.
John Gruber
Okay.
Craig Federighi
That there's no mixing and matching going on between all of these different pieces. And so, if you go and say, well, I don't like... what's an app that someone would really... I'm going to get myself in trouble here. Okay.
Phil Schiller
Hmm...
Craig Federighi
[fake smile] I can't think of one! I...
Phil Schiller
Stocks.
Craig Federighi
Stocks?
Phil Schiller
Stocks. Some people don't follow the stock market.
Craig Federighi
Fair enough. Some people do not follow the stock market, or there's not one in their country...
Phil Schiller
Which is good for them, yes.
Craig Federighi
And so they might remove that app. And when you do, it's hidden, and any user data and preferences and so forth associated with it is gone. If you want to get it back, we were thinking, well, how do we let you restore this. And we thought, "Well, people are naturally, when they want to go get it back, they're going to go to the App Store and search for it. And so, you go to the App Store and search for it, and it'll show up, and you'll say Get, and it will reappear [on your home screen].
John Gruber
'Cos that's how they know to install apps.
[laughter]
Craig Federighi
The download will be remarkably fast.
Phil Schiller
Exactly.
Craig Federighi
The compression technology... good stuff.
Phil Schiller
It's been... and it has lead some to mistakenly report that we're moving these apps out of the system bundle and into the store for downloading, and that's not really the case; we're just making that the easy mechanism for restoring, seeing it from the store side. But it's really still part of the system.
Craig Federighi
Good to set the record straight here.
John Gruber
That's interesting. Because that means there won't be, like, an update to Mail that comes through the App Store, it's just like it used to be: It'll be part of the system update.
Craig Federighi
That is correct.
On pre-announcing App Store changes
John Gruber
Well, speaking of the App Store, this last week...
[Craig leans precariously backward to reveal Phil, to laughter from the crowd]
A week ago...
Phil Schiller
There was a reason I sat on [the far] side! I just thought these two were going to totally nerd out, and I'm just gonna let them have fun. And I... have no problem with that.
John Gruber
A week ago, there were a bunch of changes, improvements to the App Store. And in a certain sense, one of them did not get mentioned in the keynote. But review times for apps submitted to the App Store are waaaaay faster than they used to be!
[huge applause and cheers]
Craig Federighi
We thought, this is one of those cases where we can address a problem before it starts to boil over.
[laughter]
Just in anticipation of potential future.
John Gruber
For the audience at the keynote, though, to not even mention and just take that applause is amazing, because you know that it's coming. And developers are pretty happy about that.
Phil Schiller
It would have been an easy way to get applause, but we didn't stoop to that trick.
[laughter]
So, yeah! It was exactly — people have all these awesome conspiracy theories, and they're fun to read, but it was exactly what we said, which was that we were working on the keynote, we actually thought about having a whole developer section to talk about the App Store and the Keynote, and looking at keeping it, we really wanted to get [the keynote] done in just under two hours if we could. And you couldn't really talk about that, and the subscription stuff, and the ad search stuff, and all that, in three minutes.
You really needed, probably, about fifteen minutes to explain, and it just wasn't worth losing fifteen minutes of product time to talk about that when if we could, instead, just talk it to people ahead of time.
And so we decided to do something we've never done before, which is before the keynote, explain some of this. However, it was kind of tough to do, because here we're talking to you and a few others, and saying "Here are things we're doing for the App Store," knowing that still had to come, a few days later, apps working with Siri, and apps working with Messages, and these are huge impacts on developers. And a new store for Message apps, we're going to come out with. So we couldn't really tell the whole picture of all the things we were doing.
So we told sort of half of it, and waited for the rest.
On App Store search ads
John Gruber
Well, part of it that goes together. So, one of the improvements last week was search ads. And...
Phil Schiller
I noticed — I don't know if any of you [gestures to audience] noticed before we came out, there was an ad that showed up first, as John, you did your ads before we started this session.
[laughter]
It was really nice, thank you!
[clapping]
And I found two of the three were relevant to what we were discussing!
[Craig cracking up]
I won't further — for the benefit of your advertisers, I won't mention which one I didn't find relevant to my interests, but...
[Gruber laughs]
John Gruber
I was going to be nice! I was going to say how there's a tie-in that you couldn't mention a week ago, where the idea of the search ads is that it improves discoverability. And there's a discoverability aspect with the iMessage apps, where if I send you a widget through an iMessage app —
Phil Schiller
That's right.
John Gruber
And you don't have it yet, there's a very subtle, y'know, I forget what exactly it says.
Phil Schiller
Yep.
Craig Federighi
Get.
Phil Schiller
Yeah. Two — couple of the very interesting things that the team did in working on these message apps is: #1, that if I send you something, if I send you a sticker, if I send you a JibJab, you get to receive it and experience it without having to download the app. And so, you can do that on a lot of these things. Where some other service, you're always being hit with a "Download this in order to see what someone is sending you!"
So the team really wanted to have a great experience for the receiver — you don't have to do that. However, there is attribution there, and you can choose to get it. If you're like, "Wow, those JibJabs are really cool, I want to download them too, and share them with friends." Hopefully that'll become a nice viral marketing, in addition to other ways for users to discover apps in messages.
John Gruber
On search ads: Make the case — when we talked last week, you did. On the phone call, I thought "Yeah, that makes sense." And then I went away, and looked at my notes, and I was like, I'm not sure I get it.
[laughter]
Make the case on this part, on this particular part, that the system that you guys have designed can, and should be, to the benefit of smaller indie developers and it's not going to be dominated by the biggest companies —
Phil Schiller
Mmhmm.
John Gruber
That, with the, y'know, budgets that are more than everybody here combined.
Phil Schiller
So, the two sort of priorties we set on the team as they were working on it was, if we're going to do this, we have to do it in a way that, number one, protects user privacy. There are many ways that companies do it where they're not protecting privacy and we need to understand that. And secondly, how do you do it in a way that gives advantages to small and indie developers, because it's easy to imagine a system that didn't do that.
And so, we set out to think of all the things we could do to make that possible. And there's a long list of things. And I won't go through all of them to bore you all, but there are many things.
Things like:
- First of all, there's no minimum bid. So we don't set a bar, if you have a very small amount of money, you can just do what you can with a small amount of money.
- The fact that we're going to work really hard to try to make relevance the top priority, over bid, for why something gets shown. That the users are the ultimate deciders of what gets shown, based on their clicks, they're a big input to what is relevant to the search result.
- The fact that we're going to work hard to try to police and improve the whole metadata system if we find, as it easily could be abused to hurt [small] developers.
- The fact that — and this has been a hotly-debated thing — the fact that you can do conquesting. You can use someone else's brand in your ad words that you want to use. As we thought about it, that is more likely to benefit the small developer than the big developer. Because the big developer isn't going to pick on a lot of small developer terms, but a small developer can try to latch on to a big developer's name. If I want to search for Angry Birds and your game, you can. Right? And so we think that that can help them.
- The fact that there's no exclusivity. So a large developer cannot say, "And I want to be the top bid, and I'm going to spend everything I can to buy out this term." There will be no exclusivity, there's going to be a rotation there, and as that rotation appears, the relevance will help drive it further.
We're trying everything we can, and I think one of the best things is, right now, once we're in beta throughout the summer, the downloads the users get from the ads are real downloads to benefit the developer, but we're not charging [for ads] during the beta time. So there's a chance for everybody to get in and try it out, help us learn from it, and drive real downloads and real business without any marketing spend.
So we're trying to think of things we can do, and we'll think of more. We'll take feedback and see what's happening, and where it works and doesn't work, and where it feels like they're getting stomped on, and we'll try to do all that we can to make it better.
[cheers and claps]
On App Store subscriptions
John Gruber
And the last bit of news with the App Store changes, y'know, big third of it, was an expansion of the categories that are allowed for subscriptions. I don't know if you noticed, but there was a little bit of confusion last week about the difference between apps from all categories versus "all apps."
Phil Schiller
Let me just, let me explain that. So, our intention is exactly as we talked about. Which is, we're opening up the subscription model to all categories, so what kind of an app you make doesn't directly have an impact on whether you can have a subscription model or not.
There are, we wanted to open up subscriptions to all developers of all apps. That is the hope. However, there are a couple of little "gotcha"s where we have to be careful. And so, that's why there's some caution here.
Number one: If you want to create a professional app, and you're going to to maintain it, and do updates, and you want to have an ongoing revenue stream, that's of course an intention of this.
[clapping]
Yeah, let's clap on that! But do users really want, and I'm sorry to pick on this category if somebody makes this app, because I'm sure there's examples where you would want it, but do you want a flashlight app to now be an app you have to pay for forever with a subscription model? Users probably don't want that.
And so, we have to be sensitive, first of all, to: Is there some minimum functionality where users now get pissed off, and say "Everything's turned to subscription, I don't want to buy stuff anymore, this is not okay," and now that's a drag on business on the App Store and therefore, we all lose. So we feel a responsibility. And I read your thing that says, "Hey, why don't you just let the market choose—"
John Gruber
Right.
Phil Schiller
Well, what if the market screws itself up and it does badly? And then we all lose. So we have to be able a little bit sensitive to not do something we think could backfire and hurt all of us. So we want to be careful about minimum functionality, so there will be some guideline around that.
Which we already have a guideline on minimum functionality for anything, you can't just wrap a website and call it an app. But there will be a little bit more minimum functionality for subscription.
John Gruber
I think the guidelines include, a longstanding guideline is that the App Store has plenty of fart apps already.
Phil Schiller
That is absolutely one of the rules.
[chuckles]
And then, there is a secondary issue. And we're working through this. There are certain states and governments where there are laws about creating a subscription revenue stream without a clear promise to the user of what they're paying for down the road.
And so, our legal team's been working with us on this, on trying to make sure we put in place in the store the right way for developers to make clear their intention to deliver value for that customer, or else they'll be breaking the law by asking for a subscription with no intention of delivering value down the road.
So we want to be careful of those things. So those are the kinds of reasons we have caveats on it, but the intention, I think, is what we all want.
John Gruber
Alright.
[cheering]
On the Mac App Store
John Gruber
The Mac App Store...
[chuckles from audience]
I'm not going to say it's been treated as the ugly stepchild, but maybe the slightly less attractive stepchild? And a couple of examples: TestFlight beta testing was in the iOS App Store. Craig [Hockenberry], is it in the Mac App Store yet?
Hockenberry
[from the audience] No, I don't think so.
John Gruber
I don't think so. Alright. So no TestFlight...
Phil Schiller
[dryly] Hi, Craig, how are you doing?
[laughter]
John Gruber
Video reviews. I know it seems like that really works, like there's just, instead of static screenshots to show your app on iOS, you can have a video that shows it in animation, and then lots of times, for developers who are doing the cinematic experience of really making the app feel great, the video can do so much more than a static screenshot.
Craig Federighi
Mmmm.
John Gruber
And all of the news last week applies to all of the App Stores.
Craig Federighi
Yes.
John Gruber
So that in of itself is a change, a change in the way that the App Store is distributing new features.
Craig Federighi
Yep.
Phil Schiller
So, we love all of our kids — I'm sure all of you do as well — equally.
[laughter]
And so, we love the Mac App Store, we want it to do well, we want to support the developers in it, we care a lot about it. We use it ourselves, it's a very important store for ourselves. We've moved all our software distribution into it, and are very happy with that. So, we're one happy software developer that's using it.
[scattered laughter]
And we still think, in the long view of all of this, it matters a great deal. We think it matters for privacy, we think it matters for security, we think it matters for quality on the store. We've all seen examples of apps that have been hijacked on servers, where people download stuff that have viruses injected in them, and we don't want any part of any of that, all of us.
So we think it's still an important solution, and we're dedicated to it.
There are things through the years in the Mac App Store that haven't been fully implemented because they didn't make as much sense on the Mac as they did on iOS, or the engineering effort was really high for a benefit that wasn't seen as as big, or whatever. Example: So, TestFlight. For the engineering involved there, people have felt that there are a lot of opportunities on the Mac from a website to download apps for test, and to distribute beta software, so the need wasn't as great. Right? It was a clear need on iOS, not clear on Mac.
So that's why some decisions were made and trade-offs, there. But, as you say, as I've been working more with the App Store team since December, I've really pushed the team to please make sure everything makes sense across all the stores as much as possible, and maybe there'll be some exception to that that we have to make, but we don't want to. We want to try to do everything the same on all the stores as much as possible, including the Mac App Store.
[applause]
On iPad app pricing and lack of pro apps for iOS
John Gruber
So, one thing the Mac App Store has been good for, and the Mac software ecosystem in general is good for, is that it seems to support higher prices of apps, for truly professional apps, deeper apps. And there's a consensus — or maybe not, consensus is the wrong word, maybe you'll disagree.
But there's a lot of people who think that one of the things that's holding back the iPad — especially now that it's the iPad Pro — from replacing a MacBook for someone who might want to, is that it lacks the same depth of deep apps for work that the Mac has. And the reason is that the pricing pressure is more like iPhone-style, couple of bucks, as opposed to Mac-style, where $50, $80, $100 software has long been the norm.
Phil Schiller
I think you see two things happening at the same time. Number one, the iPad's capabilities are growing as a PC replacement product for some people. I know some people have made some statements about that, I don't know who.
[laughter]
And so, we're trying to make it more and more powerful, making larger screens, keyboards, the more powerful processors, and all that's happening to drive it into a more capable product.
At the same time, you've started to see more professional applications begin to make their way onto it. And so, I think we're seeing changes there. We're seeing... certainly, apps that have a similar version on your iPhone that you want on your iPad will have similar pricing. But other apps that may be coming over from the Mac, or PC, are bringing on pricing models that are more like that.
And so you're going to see this duality with iPad, that there's a little of both happening. And we see an increase of the more professional apps happening. And we see stuff in flight with developers we're working on that's really impressive desktop-quality software, more and more coming to iPad.
John Gruber
Yeah, it's definitely not the hardware. 'Cos the iPad Pro stands toe to toe with the MacBooks on any technical measure you can give it. I mean, beautiful displays, powerful CPUs, and stuff like that. So it's not holding it back.
Craig Federighi
And I do think if you really look at some of the professional apps that are on the iPad, it's... I mean, some of them are really first-class. I think the iPad Pro's going to accelerate that, and we absolutely want to find any way possible to make deep investment by developers on the platform possible. Because, I think, we'll all win when that happens.
On passwords and macOS's auto-unlock
John Gruber
Alright. New topic. Privacy and security. I remember a couple of years ago, maybe more — I don't know how many years. But I was at WWDC, and I somehow wound up in a session on security. I don't even know why I was there. But it was interesting. I think I was talking to somebody, and he was like "I gotta go into this thing on security," and I was like, "Well, I'll go with you," and I went in and listened.
And at the end, it was when they were still doing Q&As, and I remember this very vividly. Somebody asked the question of somebody who was on the engineering team in charge of security, gave a rant about how passwords are terrible, and people pick bad passwords because they're easy to remember, and passwords that are hard to remember, or hard to crack, or hard to guess, are unusable, or less usable. "Have you guys given any thought to what's next beyond passwords?"
And there was this pause, and the speaker...
[Gruber intimates looking down toward the mic.]
"Yes."
[laughter]
And it was like, well that's an interesting — that's a very interesting and truthful answer. And we've seen, I think, in the intervenining years, some of the things that might have been circulating. Touch ID...
Craig Federighi
Yeah.
John Gruber
And now, one of my favorite features you guys announced yesterday, can't wait to use it, is...
Craig Federighi
Auto-unlock?
John Gruber
Auto-unlock.
Craig Federighi
Yeah! Yeah.
[big cheers]
John Gruber
So can you talk about how that came to be?
Craig Federighi
Which part of it? I mean, caring about security? Or, uh...
John Gruber
Well, no. With Auto-unlock in particular, the details of how — what are you guys doing to make Auto-unlock truly secure? That it's not, y'know, that I'm not over here opening Phil's MacBook because he's in the room.
Craig Federighi
Yeah. Yeah. Well, of course, this — it's a continuation of the work we did with Continuity to develop really low-power BTLE-based discovery protocol, so your devices could discover each other continuously with acceptable overhead from a battery point of view. And also, all the authentication mechanisms we put in place as far as having your devices know that they're your devices.
So that's kind of a foundation. The big challenge with Auto-unlock is you don't want a kind of, a relay attack, where Phil is actually, well far away from his office, and someone basically has a Bluetooth listener that's going to forward a signal to you, 'cos you're now by his Mac, and this Mac is having a conversation with Phil's watch over a very long distance.
John Gruber
Right.
Craig Federighi
And so, we're actually able to do time-of-flight calculation using peer-to-peer Wi-Fi, where we literally can measure how long (at the speed of light) it's taking for the signal to travel from your watch to your Mac and back!
[applause]
That's a very fast stopwatch! And so, because of that, if you interpose any kind of relay, it will introduce a delay that immediately would tell us, there's hijinks afoot, so.
John Gruber
Yeah, make sure they type in their password.
Craig Federighi
[laughs] Yes. Absolutely. And so, that piece is critical. But I think, y'know, on the bigger picture, Touch ID is one way that we've helped with passwords, but actually, on iOS, the Secure Enclave and that whole architecture, the fact that your device is not encrypted just with your passcode, which, honestly, whether it's four digits or six digits, it's short enough that if a brute force attack were possible, it would be — you readily could break into something.
But instead, it's entangled with a hardware key that only the Secure Enclave runs, and the Secure Enclave will only do its unwrapping when running Apple-signed software, and will only let you try ten times. And so, fundamentally, yeah.
[applause]
That was the first, yeah, very important step to saying you could have a practical-length passcode with really industrial-strength security. And so, we keep pushing on this.
On deep learning and photos
John Gruber
Continuing in the privacy vein, it's a good segue into Siri, "Deep Learning," AI, these sort of features that you guys... it was a big part of the presentation yesterday. Because a big part of your on-stage message about it was the emphasis on the way that the systems are designed to protect users' privacy, and the technical implications of that.
Craig Federighi
Yeah!
John Gruber
So one of my questions: When does deep learning happen? So, like, I'm on the phone and I'm taking a couple of pictures of the event and stuff like that — when does the...
Craig Federighi
That analysis occur? Yeah, so, if you upgrade your device to iOS 10, and you have your Photos library there with your 10,000 photos, or 100,000 photos on it, the analysis of that kind of backlog will occur when you're plugged in on AC overnight. 'Cos this is a considerable amount of computation that's going to occur that we would not have happen in your pocket.
But when you're out taking a fresh picture, at that point, we will instantaneously form the analysis on that hot photo as it's going into your photo library. We can do it that fast. It is, like scene classification, I mentioned yesterday — was it yesterday?
Phil Schiller
Mmhmm.
Craig Federighi
Yeah. [laughs] Long time ago. That it is, like about eleven billion calculations that have to occur to do that thing — "That's a horse! That's a mountain!" — but the GPUs on iOS devices these days really cook, so we can get through that essentially instantaneously with the photos.
John Gruber
And on the privacy part, my understanding — and correct me if I'm wrong — but my understanding from what I've learned is if you've got iCloud Photo Library, and I take a couple pictures with my iPhone, the photos will sync to the cloud, and then they will go to my iPad and my Mac, but the deep learning analysis doesn't go with them. Each machine performs its own processing on its own time when it's plugged in and appropriately. Is that true?
Craig Federighi
That is true right now. So each device does its own processessing. In the future, we could share the results of, like, the first one who does the work, just share, just to make that work go along for the ride. But today, it's gonna be each device doing it independently.
When you think about what's going to happen if we release iOS and OS X on separate days, everyone's iPhones will race to do all this work on their library first, and then the Mac will, it'll be fine at that point. So we wouldn't have saved the iPhones from doing the work if we'd had the Macs share their work, or share the work of the Mac.
Phil Schiller
And just to add on that view of someday they may not all have to do it, it's a view where we're not, Apple will never actually know that analysis ourselves. We won't see that data.
Craig Federighi
Right.
Phil Schiller
It's a way to do it, but we're out of the loop.
Craig Federighi
Yeah. To be clear, the photos themselves are, the architecture sets are encrypted in the cloud, and the metadata — any metadata about the photos that you create or that we create through deep learning classification is encrypted in a way that Apple's not reading it.
[applause]
On differential privacy
John Gruber
I wanna, I want to get... [laughs] I want to get nerdy on this differential privacy thing.
Craig Federighi
Yeah!
John Gruber
'Cos it's a phrase, it's like an official thing, I've learned a little bit more, it's not just a phrase you guys made up, it's like a...
[laughter]
Craig Federighi
[chuckling] It wouldn't have been the phrase we would have made up.
John Gruber
Right. [laughs]
Phil Schiller
We would have done a better name if that's what we did.
[more laughter]
John Gruber
But like, in the State of the Union yesterday, I mean there's real math behind it. This is not just a name that is applied to policies. This is —
Craig Federighi
That's correct.
John Gruber
A branch of statistical analysis —
Craig Federighi
Yes.
John Gruber
That, it... talk to me about it. Give me a little, I know you touched on it in the keynote. Give us like a slightly juicier layman's overview of differential privacy.
Craig Federighi
Sure. Yeah, of course the idea is that if we wanted to know what word, y'know, a new word that everyone was, that lots of people were typing, that we didn't know so that we would stop marking it as a spelling error. Or maybe we'd even suggest it on the keyboard.
Phil Schiller
Like "Meh." Or something.
[laughter]
Craig Federighi
Yeah, like now it's just, it's trending, it's hot, we want all our customers to be able to know that word, but we don't want to know you and Phil in particular are typing it. We want to have no way to have any knowledge of that.
You can imagine if what we're essentially assembling is a picture of little pieces of data, y'know, of the forest, but all we're getting is a little piece. And when we get that little piece, even each device will statistically, much of the time, even lie about its little piece. Right?
But those lies will all cancel out with enough data —
John Gruber
Right.
Craig Federighi
— and the picture will suddenly resolve, with enough data points, will resolve itself. And so, and yet, literally, if we were trying to learn a word, we would send one bit — we'd send a position and a single — we'd hash the word, we'd send a single bit from the hash, we'd say at position 23, Phil saw a 1. But Phil's phone would flip a coin and actually say, "Actually, I'm going to lie about that. I'm going to say zero even though I saw a one."
And that's the data that goes to Apple. And Apple, with enough of that data, can build a composite picture and say, "Holy smokes, we have a word here. And this many people roughly are seeing it." And that's typically what you want to know. You want to know what's happening at large, but we have no desire to know what, specifically, who is doing what.
John Gruber
Well, it's typically what you want to know. It's not typically [laughs] what your competitor would want to know.
Phil Schiller
And part of the reason that this is so important to get into is because, the theory that you can just anonymize the data and send it up, and all's good, and it's a bunch of crap —
[laughter and big hollers of approval]
Because I can send all this data, and say "Well, I don't know who you are, oh, but I happen to know the same location you go to every night, and I know the same place you go to work every day, I've got all this data, I just don't know your name, or ID. Boy, it's really hard to reverse-engineer that anonymous data!"
Right? So what you need to do is create a system that goes beyond anonymizing to really make it impossible to reconfigure who that user is.
[loud applause]
John Gruber
So the way I have it written down here is that if it works as you're describing it, it means it's not just that Apple doesn't use that information to reverse the anonymity, it's that mathematically you can't.
Craig Federighi
We can't.
John Gruber
It's — the design of the system is such that it's not even possible if new executives come in in a few years and maybe they would like to, y'know.
[Craig exhales, laughter]
Craig Federighi
[mumbles]
John Gruber
Well, companies change!
Craig Federighi
No, no! That's absolutely true. And the point of view, honestly, the point of view that someone says, "Hey, I know we know a ton about you, but don't worry! We're nice guys! And it's all good!"
Well, okay, maybe you're nice guys, ten years from now, who's running this thing? Or, what if someone breaks into your computers? Are they nice guys?
Phil Schiller
Right.
Craig Federighi
Right? So you just don't want to have any central source that has that kind of knowledge, because in the fullness of time, anything is possible.
And so, differential privacy is, I mean, there are mathematical proofs that will show that you cannot, with any confidence, determine anything about any of the people contributing to the data set. And we think that's important.
On Google and Facebook and cloud data-gathering
John Gruber
Alright. Speaking of companies that do collect some information about people, Google and Facebook —
[laughter]
— they're two competitors that, and I know a lot of times, when you talk about these companies, you might talk [vaguely] about search engines, and you might talk about social networks —
[more laughter]
Because you're gentlemen! But I will name names, and I'm going to just point out that Google and Facebook are both actively pursuing a lot of the same goals. I mean, just the image analysis, "That's a mountain, that's a horse" — those companies are showing similar things.
Craig Federighi
Yeah.
John Gruber
You guys are showing similar things. But it really is, I don't want to abuse the metaphor, but it's a 180-degree different tactic, where they're doing it with cloud servers, and doing the computing in the cloud on data that they've aggregated there, and your method is to do it, distribute it on the actual devices.
Critics are saying — and not me, I'm not saying this, I'm like, let's see —
[laughter]
No, I mean, I'm totally like "Okay, I saw your keynote, I'm looking forward to trying it, and let's see if it works for me!" I don't know. I think it might!
Craig Federighi
I think it will!
[laughter]
John Gruber
But critics are already saying, and they've obviously, since the keynote was just yesterday, I've seen it in a couple of articles that your strategy is doomed to keep Apple behind them, because the Google and Facebook way is the only way that works. And I'm not quite sure where that comes from, because...
Craig Federighi
Their PR department. I mean...
[huge laughter, clapping, Schiller makes a wry unintelligible comment]
[chuckles] of a prominent search engine or social network provider of... that we don't know about.
John Gruber
I think that part of it, in my mind, is maybe that there's an assumption on the part of some people in the press that a server farm has this massive amount of computational power, and that a puny little phone can't compete. But it's not like there's one person's iPhone who's trying to do the image analysis for all the photos on iCloud...
Craig Federighi
Like, there are a billion phones to throw at this problem.
John Gruber
Right, a billion active devices. So like, the billion active Apple devices that are out there in the aggregate have an enormous amount of CPU power.
Craig Federighi
That's right, that's right. The other thing is, there's this idea that, well, if you don't have the data, how would you ever learn? Well, turns out, if you want to get pictures of mountains, you don't need to get it out of people's personal photo libraries.
[laughter, clapping]
Like, we found out that we could find some pictures of some mountains!
[huge applause]
We did some tough detective work, and we found 'em.
[laughter]
So... [breaks down laughing] that's pretty good.
On Siri
John Gruber
So, moving on. Siri. Siri now has an API, and it's six categories. I don't know if I wrote any of them down. But it's like ride-sharing...
Craig Federighi
Messaging, photo search...
Phil Schiller
Voice calls.
Audience member
Payments!
John Gruber
Payments. And one more.
Craig Federighi
Sending money, yeah. No, we did payments. Alright. I can say it a few different ways, we can get past six.
[laughter]
John Gruber
Well, there's six distinct categories.
Audience member
Workouts!
Craig Federighi
Workouts!
Phil Schiller
Workouts!
John Gruber
There we go. Thank you. This is why I should have a live audience at all my shows.
Craig Federighi
This is crowdsourcing, right here, but it's totally anonymous.
Phil Schiller
And we don't know who said it, because...
[huge laughter and clapping]
John Gruber
So why restrict Siri to those six specific categories?
Craig Federighi
Yeah. It comes down to modeling the domains well. In order to understand what someone is saying — people are going to speak to Siri in a whole bunch of different ways, and even in a whole bunch of different languages. And when they say, when they want to say "Send a message to Phil" saying that "I'm going to be late for the interview," then we... I could have said that in dozens of other ways. I could have said "WeChat Phil that I'm going to be late for the interview," WeChat Phil using WeChat, telling him that, et cetera.
And I could say something like "WeChat Phil," and then I'd need to know, like, okay, well, what do you want to say to him? And Siri knows all of this because Siri understands the domain of messaging well. It understands all the vocabulary, it understands what the verbs are, what the objects are, and can collect them and can do so in a dialog.
And so, we want to make sure that when you're talking to your assistant, your assistant is consistently intelligent about understanding you and how flexible you can be in talking to it. To do that, we had to develop those domains.
And so, these are the domains that we've developed in a way that developers can plug in. We'll do more and more of that over time, and of course we'll search for more and more flexible ways to enable developers to do that time, but we want to make sure that what we do is preserve the intelligence of your assistant.
It would have been really super easy for us to say, "Hey, just tell us a trigger word, or the name of your app, and we'll hand you a string."
John Gruber
Right.
Craig Federighi
And good luck. And so you'd say something to Siri, and most of the time, you'd get back the app doing something crazy, and the user would say, "What in the heck, Siri doesn't understand me, I don't understand this."
In this case, we're able to be consistent about Siri's ability to understand you. And so, we'll make models more and more powerful, and we'll do more of them for more domains, but we start with a baseline and have a quality experience around what we cover.
Phil Schiller
And I think that this is an insight into how we... it's not right or wrong, how we approach things differently than some other companies do. We've all been seeing stories for awhile saying, "Hey, Apple, some other companies are doing some assistants, and they're allowing these other apps to be, these bots to hand off, and do things for them. You're not, you're behind."
Where, when we have thought about doing it for awhile, and we've thought about it since the very beginning of Siri, which is, we needed a solution to — how to keep Siri from being smart at one thing and stupid at another? That would be an inconsistent experience, and not what we want. We need Siri to be equally smart at all the things we do.
And as it gets expanded, that intelligence needs to extend, and so the team has been working hard at that, where others shoved in [trigger words] quickly to do things that don't translate that intelligence to third-party apps. And so, to do that means that you have to, with intention, add categories and domains. The hope is to add more and more so that users can ask anything they want over time, use any of their apps that they love, and it all works. It just takes time building domains. So we'd rather take the time to do it right than rush out just because it gets a good story to say you have something.
Craig Federighi
Yeah.
[applause]
On iMessage
John Gruber
One of the things I've — like, in the last year or so, maybe more, but I've noticed it, and I bang this drum a couple times a month on Daring Fireball, is why the industry as a whole doesn't seem to count iMessage as a messaging platform. And, y'know, the number that always gets thrown around is monthly active users, and WhatsApp has so many monthly active users, and so they're worth so many billions of dollars, and... iMessage has... it must have, it has to be right up there in terms of monthly active users, daily active users, hourly active users.
Craig Federighi
Yeah.
John Gruber
Users sending iMessages during the Talk Show...
[chuckles]
Is that frustrating?
Craig Federighi
I... I don't know. I mean. It's okay? [laughs]
Phil Schiller
No, because customers...
Craig Federighi
I mean, really...
Phil Schiller
Customers don't read those things. It's all inside the beltway kind of, like, who feels prouder that they made a list.
[laughter]
It doesn't matter to users.
Craig Federighi
Yeah. Messages is the most-used app on iOS, period. So, it's used a lot. And certainly, we saw that every time we'd add a couple new emoji, it would be the biggest thing. We work all year on, like, a new file system or something...
[laughter, cheers]
Craig Federighi
And it turns out the rest of the world outside this room was more excited about the two new emoji! So, we figured, y'know, if there's one place we can make a tremendous difference in how people experience iOS fundamentally, they're spending a lot of time in Messages.
And so, we put a ton of creative energy into it, and ultimately, through opening up to developers, I think the collective energy that will go into making Messages great is going to be phenomenal.
John Gruber
In the keynote, I was sitting in the middle of the floor. Halfway back, halfway in the center, just right in the middle. Really, it was a great place to hear the reactions. The biggest reaction I thought of the entire keynote was when you announced that emoji were going to 3x. [Laughter]
I'm not exaggerating. It was like a real, visceral buzz. Here's a crowd of people, you know, developers who are more technically minded, and are here to hear about technical details, and this thing that is really just, you know, just more fun, got this really powerful reaction.
Craig Federighi
Well, next year we're going to 4x. [Laughter]
Phil Schiller
See, this is why we don't let you out. [Laughter] Now we're going to be held to that and next year, when we don't, they'll be like "You said 4x! Apple, you're late, you're late!" and then it will be, "Finally, 4x!" [Laughter]
John Gruber
Coming down the home stretch—
Phil Schiller
There's a few competitors right now: "4x, let's beat 'em to 4x!" [Laughter]
John Gruber
The Onion story about the Schick CEO says "Screw this, we're going to 5 blades!" and, like 3-years later, she came out with a five-blade razor.
Craig Federighi
Yes!
John Gruber
Can't underestimate what people will stoop to.
Any of the other iMessage stuff that, I mean, clearly it's a lot of work, a lot of it is very fun, some of it is — you know, the developer integration — really turns it into a platform. It's not just a thing that people can text with any more, it's a thing that people in this crowd can write software for. Is there anything that stands out that maybe didn't get enough love in the keynote?
Craig Federighi
Well, we didn't talk about the way in which I think these apps can spread, kind of in a really great way virally, we didn't talk about that at all. I think that's going to be really powerful for developers and is going to make it worth developers' while to put some energy into it. We made them really easy to create, so if artists — we think there'll be a community of artists that will build sticker packs that are just really fun and they don't need to write any code to do it. So we think that will be really important.
Also, I think, the way that they are distributed — it's not just about the extension. The extension can be a part of your app and so, there's some cases where, you want a model where, the extension is sort of in cooperation with your larger app experience. I mean, one simple example would be like if you have your sports app, your sports app knows what your favorite teams are. Well, your extension in messages, that lets you share those clips, is going to know that as well. So, there's a connection there. We have one where something that people like to do a lot is share music. You hear something, you think, "wow, this is great" and you want to tell your friend about this awesome song. Well, if you go to the Apple Music extension, it knows what's now playing in your music, and it knows what you're listening to for your songs, so that's just one tap to share.
So, there'll be interesting integrations where the Message extension is sort of the tip of the iceberg for an experience you have in your app as well.
On watchOS
John Gruber
Moving on to watchOS. What you guys do year after year is make iterative improvements. You add features. You take what was slow and make it fast. You take what was ugly and you make it look better. But the performance improvement on app launch on watchOS 3 —
Craig Federighi
It's dramatic.
John Gruber
— Does not look like one year over year. It's crazy. And I really did, in the keynote, think "I gotta see this". And then, when I got a hands-on with a Watch running watchOS 3, it's for real.
Anybody in the audience, have you guys upgraded? [Applause]
Craig Federighi
It's for real.
John Gruber
How is that possible? [Laughter]
Craig Federighi
A couple of things. We actually had some RAM to spare.
John Gruber
Really?
Craig Federighi
Yeah! Turns out that if people are running… if they have their favorite ten apps, we can keep them mostly running. We can keep them mostly resident. We can halt them so they're not burning CPU, but we can keep them mostly resident. Which means you're not doing all the work that goes into launching an app when you take them live.
The other thing is, it turns out that when we first were coming out with watchOS we were being really conservative about understanding how people were going to be using the Watch and trying to make sure we could hit our goal of very solid all-day battery life. So that you could use it all day and charge it at night.
And we found that we actually really overshot the goal, which was an area of just massive focus and paranoia through the release. We needed to make sure we squeezed every little bit of juice out of the thing. So, realizing we had this budget, we said, look, we actually have enough to do background updates. We'd overshot enough that we could keep apps both in memory but also keep them up to date throughout the day. So, when you look at them, they're already there. It's not like, launch and then wait and have them get the information. It's, they already have the information.
So, those were two really vital techniques. The other thing is, as you build something new and different as the watch, you finish, and you live on it, and you figure out what's really the essence of this thing, and appreciate which problems are the most important to solve, we realized the watch is all about glanceability. It's useful to the extent that, okay, I can solve my task, I'm done. If I'm up here and I'm waiting and I'm fiddling around, my arm's getting tired, this is no fun anymore, I'm going to do this a different way. And with that as our obsession for the last year. We've taken all of those tasks and said you've got to be able to finish the task, end-to-end, in two seconds. Right? And that means the launch had better be instant part because now we need to let the user think and do something in two seconds and get it done. With that focus, you find a way. We chipped away.
John Gruber
What really struck me, once I got a hands-on with it and could really see it, was just how much the design changes to the navigation of the user experience are exactly coinciding with the engineering improvements to make it faster.
So, the fact that glances are no longer a separate thing is because the apps themselves in the dock can serve as glances because they're getting the background updates because you made the changes that let them stay resident in memory.
Craig Federighi
Yeah, it's nice when it all comes together.
John Gruber
It really is.
Phil Schiller
The other thing I'll add is that, once you start to use the new watchOS, in addition to having the apps come across faster and you can get access to them quickly, watch faces in a sense become apps themselves. In a sense that you change the ones you use, rearrange them, and change the complications. For example, I would normally keep the activity rings on my watch face, but now I can choose to make that the next [watch face] and swipe over to them and swipe back, because I use the activity watch face vs. needing the rings. And then I can have a different watch face for some other time of the day for when I need other actions and access to apps. So that becomes a much quicker and more useful way to expand the things you do with it. It's a really profound change throughout the interaction model.
Craig Federighi
Yeah, and really different complications too. So if you're going to be more in your workout mode, you would swipe over, and the complications that are relevant to that, and therefore the launchers that are relevant to that, are essentially right there. So you kinda go, here's what — I'm at work, and I'm going to be this way, I'm out with the family I'm going to go this way, and you have all the activities that are relevant to that, it's like you have almost a custom doc or custom launcher based on your watch face. So that's another element where I feel like it's really come together in a nice way. [Applause]
On Swift
John Gruber
Just wrapping up, coming down to the home stretch. Swift. Now, you were on my show a few months ago when Swift went open source. We had a good time.
Craig Federighi
Lot of dynamism.
John Gruber
And we talked about Swift use within Apple, and why you guys can't yet write the OS in Swift but that engineers are using it to write unit tests and things like that. It's getting used. But I saw in the announcement that the new Swift Playground app is itself written in Swift.
Craig Federighi
Yeah, well actually in OS X, like most of the Dock—
John Gruber
macOS?
Craig Federighi
—and most of Mission Control. Yeah, well, oh god. [Laughter]
Phil Schiller
Another dollar. [Laughter]
Craig Federighi
In our Sunday rehears—
Phil Schiller
Gee, when was that name first hinted at? [Laughter]
John Gruber
I don't know.
Phil Schiller
Last year, right here. I thought I was being so transparent, too.
John Gruber
I saw right through it.
Phil Schiller
I know. You were very polite.
Craig Federighi
In our runthrough for the show, when I say, "Oh, and we're changing the name to macOS" — and this is on Sunday — and my next slide is to say something about how we have these great new features in macOS. I literally go, "And so our new release is macOS. And so we have some great features in OS X." And I … it's tough. We spent a lot — 15 years — it's a long time and I think we all feel great about the new name.
Anyway, in macOS the Dock is substantially converted, and Mission Control, all those areas are using Swift a lot. So it's starting to spread a lot internally, there are some barriers, but I think this year the most important thing, and I think Chris Lattner really covered it in the State of the Union, is getting the source stability. And so we decided this year we're going to put that over all the priorities. Take what we've learned, y'know, when we first shipped Swift a couple of years ago, the objective was: Let's make sure that it's familiar from an API point of view. Y'know, minimize the kind of transition of, "Hold on, I've gotta learn all new method names for all the classes I already know?" And so we really bias toward that. Now people are so familiar with Swift, the priority is: Let's make sure those APIs are all very native to Swift in their feel. And so we've done all the hard work to update all the APIs, all the naming conventions, and take some major APIs like Core Graphics and libdispatch and make them just awesome for Swift. [Applause] Yeah, it's important stuff. But, what that means is we've achieved that level of source stability, so next year it won't be like, "Oh boy," y'know, as a developer. So that's the important thing.
ABI stability which means literally the Swift binary you built could be linked against the future libraries. That's one that we made a lot of progress, haven't gotten all the way there. But that's far more of an issue for us internally than it is for developers. It's important for developers, but I think the source stability one was the right priority and I feel really good about the progress the team made on that.
On topics from WWDC 2016 that deserve a little more attention
John Gruber
Last question: How about one thing that you guys announced yesterday, whether it was in the keynote or not, maybe something that missed the keynote, but one thing that you think deserves a little extra attention. And I'll let you guys think about it. I'll go first and your correct answer is probably new file system.
Craig Federighi
Is that right? [Laughter, Cheers]
John Gruber
I'm gonna say Universal Clipboard, because I've always wanted this. And for me it's links, it's like I'm on my phone, it's like, "Oh, I want to post this to Daring Fireball," but I'm in my office so why would I do it on a phone? I'll go sit down at my iMac and do it with a real keyboard, but how do I get this from here to there? Man, what I want to do is just copy it and go over there and hit CMD + V. And, again, the thinking through that you guys did of how to do this in a way that isn't going to surprise people in a bad way — there's like a two-minute timeout, so if I copy something on my phone right now and tomorrow I paste it in my Mac, I'm not getting that because it's really, like, a ways of detecting what's in your—
Craig Federighi
Well and even the communication is, like the other continuity features, peer-to-peer. So it's not like you're sending everything you copy up the cloud all of a sudden just so it can get down to the other device. It really is about [two devices right here] copy and paste, which I think is absolutely what people want. And it has the right privacy and performance characteristics. And, as you say, it gets rid of the surprises. And it just turns out to be the most … once you have it, it's the most natural way in the world to do these kinds of things. So I think the team did awesome work there. Yeah, I think that's great.
Do I have to say new file system again? [Cheers] No, I think the new file system is great. And by the way, I mean the prospect of … this is one you have to get right, let's say. [Laughter] And so we have an awesome file system team who really knew which problems we needed to solve for a world of flash storage and has done a super-solid job. And we're being conservative about how we're rolling it out as a developer preview, so people can kick the tires on it this year, but we look forward to making it part of the products going forward and I think it's going to great. And obviously we didn't talk about it, 'cause we don't talk about peer developer preview material there. But I think it terms of something that is important for the platform going forward, it's big.
John Gruber
What do you … so let's just say three years from now we're all using iPhones that are using APFS instead of HFS+. What would be a noticeable improvement to the experience?
Craig Federighi
So, it'll help with performance, it'll help with things like how we do software updates and other things, 'cause we can snapshot volumes and other things, we can roll things back. I mean, there are a lot of important attributes there. It's important when you think about multi-user — like how files are protected between multiple users on a Mac, because we actually have file-system-level encryption now standard across both platforms. So I think from a security point of view, it's big. And I think performance, I mean now you do a copy or even like the safe save operation, when you save documents in a lot of apps it's like, "Move that one aside, create another whole copy of all of that, now overwrite some of it, now delete the old directory." Now that's atomic and the clone file makes all of that super fast. I think it's just going to be great across the board.
Phil Schiller
I want to answer in a very different direction. Of the keynote, the thing we haven't talked about that to me was really amazing was we had a bunch of demoers who had never been in a keynote before. It was their first time. [Applause] And they were fantastic. Stacey did a great job, Bethany and Imran did a great job. Boz did an incredible job. [Cheers] And Cheryl did an incredible job. And all of them work on the things they demo, and they were fantastic. So that's my sort of unsung thing of the keynote was those presenters.
John Gruber
I said mid-keynote — I was sitting with Ben Thompson — and I said I can't believe that none of these people have ever done this before, because they are amazing. And they really did kick ass up there. That was great.
Craig Federighi
They sure did, they sure did.
John Gruber
That's it unless you guys have anything else from me.
Craig Federighi
Just thank you for having us.
John Gruber
I want to give some thanks here. I want to thank our sponsors: MailChimp, Microsoft, and Meh.com. Go there and buy some junk. [Laughter] I want to thank Jake Schumacher and Jed Hurt. They're doing the video here, so if you're watching at home, you can thank them. They are the co-makers of the upcoming documentary, App: The Human Story, which has been in the works for awhile. I've seen a rough cut, it is amazing. It is really coming along. Appdocumentary.com if you want to see more. Drew Bischof from Hybrid Events is here running whatever apparatus is involved in doing the livestreaming, which I've heard is very hard. [Laughter] I want to thank Mezzanine and the entire staff here who has been— they're led by Megan Rogerson she's been here all four years that I've been here. The staff is great, the bartenders are great, security … I mean, just a really great place and I really appreciate it. I want to thank Paul Kafasis and my wife, Amy Gruber of Just the Tip fame, their podcast that is on, I don't know, some kind of hiatus. But they're the ones who made this event run so that I can just sit back here and be nervous and make these cards with questions and not pay attention to any of the details. I don't know anything that's gone on out here. The fact that you guys even have seats is thanks to them. And I want to thank Phil and Craig for being here.
Phil Schiller
Thank you.
John Gruber
Unbelievable. Thank you. [Applause] Last, but not least, thank you for being here.
Phil Schiller
Now we've gotta find the way out.
