Apple's path to ambient computing and truly personal assistants

Links

• Ben Bajarin: Twitter
• Creative Strategies
• TechPinions
• Vector 018: Don't believe the assistant hype, with Jan Dawson

Sponsors:

• Mint Mobile: Voice, data, and text for less. Get free first-class shipping with code VTFREESHIP.
• Thrifter.com: All the best deals from Amazon, Best Buy, and more, fussily curated and constantly updated.
• Interested in sponsoring VECTOR? Contact sponsor@mobilenations.com

Transcript

[background music]

Rene Ritchie: I'm Rene Ritchie, and this is Vector. Vector is brought to you today by Mint Mobile. Mint Mobile is like one of those big bulk discount warehouses but instead of giant packages of paper towels, you get a US wireless service at low, low prices. Right now if you get three months, you also get...if you buy three months you get three months for free.

You'd also get first class free shipping on any Mint Mobile purchase by going to Mintsim.com and using promo code VTfreeship. That's VT as in Vector, VTfreeship. Thank you, Mint Mobile.

Ben Bajarin, welcome back to the rebooted version of the show. You were kind enough to be on previously. How are you doing?

Ben Bajarin: Yeah. Doing well. Looking forward to my return to Vector.

Rene: Awesome. If people aren't familiar with you...and shame on them quite frankly. If they're not familiar with you, can you just give us a little bit of your background and what you do?

Ben: My name is Ben Bajarin. I am an industry analyst with a firm called Creative Strategies. I do global market research on consumers. My focus is largely been on the consumer tech, personal computing types of devices. Obviously, that spans a lot of different categories. Obviously, because of the nature of the consumer, Apple's a big focus of ours from both an analysis and a market research standpoint.

Rene: I've said this before, but what I love about you is a lot of people will just say they feel some way, or they feel another way, or they like something better, they don't like it. You bring so much data to the table. You can say why [laughs] something is, not just because you feel that way, but because you have numbers.

Ben: That's always been our goal. It's hard. We work with executives at pretty much all the major tech companies, and it's hard to run in there and be like, "It's my opinion that so and so." You aren't going to change anyone's mind with your opinion.

It's nice to have an informed opinion, but also some really good quantitative data that backs it up and gives you some credibility that this is why things are happening. Here's the data to show it.

Rene: I did a podcast a couple of weeks ago with Jan Dawson, who's done a lot of work with you on "Tech.pinions" as well, where we talked about the home automation, the home speaker, home assistant market. Our opinion was that it was still nascent.

I know Neil Cybart -- also a friend of the show -- wrote something very similar, but you said, "Nay." You said you had an alternate opinion, so immediately [laughs] I wanted to get you on so you had a chance to talk about it.

Ben: There's a lot of the ways that you can skin this. We've done a number now of research reports specifically on not just home speakers -- Echo owners, and Google Home owners -- but really just smart assistance as well.

We've been trying to understand how people use Siri. We've been trying to understand how people use Google Assistant and how that's evolved because it was OK, Google and stuff before. We've been trying to track the behavior side of this.

The thing that I did not expect was truly how mainstream the Echo went. This was really before the so-and-so, sub-$100 pricing. I was generally surprised that you saw pretty normal consumers, your mainstream, non-Silicon Valley, non-young techie early adopters starting to embrace these products and really be quite happy with them.

Now again, they're using them in a somewhat limited context, mostly music, perhaps some reminders. The thing that I think is really interesting about this is that it's already gotten to a point where people are already more comfortable using this type of a product than they are the voice assistance on their phone. It's not to say that that does something dramatically different.

I did an analysis where I basically said, I scored the Alexa and Siri on the basis of a number of tasks, and for the most part, they do the exact same thing.

The difference is you may not always have your phone with you. You may not always be in a position to speak to it in terms of where it's at. In some cases, it doesn't always work. Siri's built for a screen where Alexa's built for something without a screen in most of its implementation.

The way that it was oriented just had this really natural expression for people even if they weren't doing these things with their phone. There was enough in all of that for us to realize that there's really something here. There's something big here.

Yes, it's smallish. I think our estimates in it now that in total there's more than 30 million of these things have been sold, and again, that's vast majority of these before they were under $100.

It's also closely aligned to some trends in the past around Bluetooth speakers that had a nice run, just being able to connect your phone and play some songs in a device that was somewhere else, portable Bluetooth speakers.

Even the iPod you could argue was probably a precursor to this, the value of just having music at your disposal wherever you are. Now these things manifest themselves in new ways because they're all over your home.

The benefit here is this. It's this theme that we talk about called ambient computing. Some device that's listening to you, some device that you can have a conversation with, some device that might even know more context about you, what room you're in.

If it's got a camera in it, it might know a little bit more about you, what you're doing at the time, and feeding value back to you from a more contextual environment in this ambient area.

That could manifest itself in a speaker, in a wearable, in something that goes in our ears, in our cards. That whole thing, there's something big there, and our first true quantification and measurement that this is a big deal is what's going on with the Echo.

Now we got to wrestle with no doubt this is a big opportunity, no doubt the home is one of the next big battlegrounds for all of consumer electronics. Now we say, "OK. Well, who's best positioned to win that, and why, and what might their strategies be for this battleground?"

Rene: One of the things I think that got Neil's goat, and I know it bothered me previously, is that you take essentially the same US media, and you feed them the nonexistent numbers on Amazon Echo, and you feed them the nonexistent numbers on Apple Watch. Apple Watch is considered a flop where Echo was considered a breakaway hit. That was perplexing on the surface.

Ben: You guys might have seen that a little bit more than I did in terms of the overall arching narrative.

They're two very different products to begin with. You've got to look at them dramatically different. You look at something that's as simple as a device that's intelligent enough to play some music. Who's not going to like that? No matter where it comes from.

Rene: [laughs]

Ben: If that's all it did, you're going to like that. Then some people were just really turned off by the Apple Watch. That's been a tough pill to swallow for a lot of people because they're coming at it from a very different perspective.

It's different when you look at a product and you say, "That's not for me." The danger with media and pundits is always to say, "Well, if it's not for me, it's not for everybody." Which is just a load of crap.

Anybody could look at the music-playing device and instantly relate to that use case. The watch was a little bit trickier because you had to use it to really understand its value where you could already get the value from a simple demo of an Echo, even if all it does is play music.

I understand how it happened. Again, you've got a lot of pundits that really do put their own self-centric colored glasses on the world and view things within what they want and like that missed the watch.

That being said, I know Neil has said this and others is that while the Apple Watch probably becomes the ambient platform for Apple. That might be the case, but my perspective is I also want to see them compete in broader ambient and areas as well. That's not just with HomePod but other things as well that I think will be necessary for them.

Maybe not everybody has an Apple Watch. Then even if you do, why can't one of these ambient speakers or some environment add even more benefit to the context of your phone, your TV, your watch, etc. The more you have, the better.

Rene: That's my big dream is that because it applies to any company. Google and Apple are the best positioned for this because Amazon, aside from Echo, hasn't had a huge success rate with hardware.

If you take the sum total of what they're doing, things like assisted mesh networks, sorry, assisted-mesh networks start being really interesting where you have...I know Apple does a little of this already and so does Google, but where you have all of these products that right now have siloed intelligences on them.

When those intelligence are no longer so siloed, and they are exactly what you said, ambient, where it doesn't matter if you're accessing them through your living room, or your wrist, or your phone, or your tablet, or your television, or your car, or anything, they have an awareness beyond the device that they're currently locked to.

Ben: I actually don't think Google's well positioned here, and I'll give you why. The best way to view these products -- speakers in particular so smart speakers -- is as an accessory to an ecosystem. I don't care how we define the ecosystem. Just think of it as it's not the central point, but it's adding value to other points that you have.

If we just use that line of thinking, we say, "OK. Well, it makes a lot of sense that the Echo does really well and that, you know, our research and others' confirms that Echo owners tend to actually shop a lot from the product."

There's a level of convenience there to just saying, "Alexa, buy me blah, blah, blah, blah, blah" instead of writing it down, or pulling out the app and going through four steps, or trying to remember to buy it the next time I'm at the store. I was in that moment in time. I know I need this. There's levels of friction.

It makes a lot of sense that the Echo is an extension of Amazon's ecosystem from a commerce standpoint. Then yes, their services like music and others come into it.

Apple actually has a real big differentiator here. That's Amazon's differentiator. It's commerce, and then services are basically equal to other things in terms of what they can do with smart home. There's nothing unique there. It's the Amazon ecosystem that's unique.

Where Apple has a unique advantage is in communications. I've tried to use the Echo to call people, and it's miserable. I would seriously use this thing all the time to text my wife, or call whoever, do conference calls. I don't care, whatever it is. Because I'm an iPhone owner, the communications angle of something like the HomePod actually becomes really, really differentiated.

For Google, the only differentiator that they have is search. You have to ask yourself then so is this just becomes an accessory to search, and if so, is that enough? It's going to do all the other things the same. It's going to do home control the same. It's going to do music the same. It's going to do videos the same. That's really all the same catalogs of stuff.

If it's just search, is that enough for you to buy one of these things because it does a better job searching the Internet than other things. I don't think that is. I don't think that's enough of a differentiator.

I don't think their Assistant is really going to catch on and be that big differentiator either to the same degree that the convenience of or depth of Siri, and maybe even where the Echo goes.

I think that Apple and Amazon are a bit better positioned within the view of that that it's an accessory, and then you just work out what is an accessory to. Apple's got a lot of legs they can run with if they use it with that sort of analogy, which is that it's a driver for services.

It's a driver for music. It's a driver for communications. It's a driver for video. It's a driver for whatever, whatever they want to build on top of that as a part of their ecosystem. Those two ecosystems are stronger than Google's holistically.

That's why I think that Google doesn't have as strong of a leg. We'll see. Right now, I lean a bit more toward Apple and Amazon.

Rene: The other area of pushback was that people in the US were so in love with Echo, and rightly so.

Even early on, it was plugged in. It had all these beamforming microphones so it could really do very good voice capture in terms of understanding what you were saying. Because it had a constrained skill set, there was a much less chance of error. It seemed almost magical compared to Siri at least how well it would perform.

Outside of America, it just wasn't available. It's just come to Canada now. I think it's in Germany and England as well. It's in very, very few countries. Again, now it's being licensed, but previously, if you did something with Echo and then left your house, you couldn't scream to tell it to change something or do something else. You were cut off from that connection.

It looked like they had come at solving similar problems but from drastically different directions.

Ben: It's just, again, they took a much more voice-first type of an approach. You run into all of these barriers still with Siri where it assumes you have your phone to look at. The Echo never does that. It knows you don't have a screen. You start doing these things that you couldn't quite do with Siri in a hands-free environment. You're like, "Wow, that's really good."

Honestly, those people who get a HomePod, it's going to do everything with the exception of commerce as good as an Echo. That's why I say view it from this ecosystem perspective, which is that's their hook.

The hook is going to be the commerce side that takes place in the home, whether that's groceries with Whole Foods. They're going to have an angle there. Believe me, there's going to be some value there in this battleground.

When it just comes down to an evaluation of tasks that it can perform, you're going to see these things perform relatively equal. That's why I say you got to look at that. What's the leg that they can stand on that the consumer says, "It's going to do all those things the same, but it's this one thing that it does or these two things that it does that the others can't that matter most to me."

[music]

Rene: We're going to take a quick break, just a very quick break. I'm going to tell you about our sponsor for today. That is Thrifter. Thrifter is a new way to save money on everything from gadgets to home goods by shopping based on value and not hype.

What that means is there's just this awesome team of people that day in, day out scour the Internet from Amazon to Best Buy to Woot and find all the best deals, make sure they really are a good deal, that they're historically low prices, or they're hard to find, or they're back in stock, or whatever makes them compelling to tell you why they're a great deal.

Then they give you the best price they can find for them. It's just an awesome service. Go to thrifter.com and check it out. All the best stuff, none of the fluff. Thank you, Thrifter.

[music]

Rene: When the HomePod delay was announced, and I talked about this with Jan and a little bit with Gartenberg, the assumption was immediately it's Siri's fault because one of Apples' struggles going forward is going to be to turn around the messaging and the perception of Siri.

It was immediately that something must be wrong with Siri where Siri had shipped for HomePod, and AirPlay 2 had not. It was pretty obvious where the stress points were with that product.

My understanding is similar to yours that from an assistant point of view, it will be better at some things, worse at others. Overall, they're going to be similar products.

Ben: Honestly, I don't know how widely this is, but the whole everybody thinks Siri is crap, and it's got a bad rap, that to me reminds me of when everybody said, "All consumers hate Windows." Then you did research, and you realize that people actually didn't really hate Windows or Microsoft.

Rene: [laughs]

Ben: All of our research has suggested people don't have that unfavorable of an opinion of Siri. I don't think that's a problem.

Now I do think though that Apple does need to bring those differentiators of surprise and delight to Siri in order to cement it as the one thing that you can't live without from an assistant standpoint. That's what I think they've got to work out. I don't know if it's there day one.

Now again, I don't even think that Amazon necessarily has that either. I just think there's a lot of convenience to this ambient-computing environment in your home.

To me, Apple's long-term hook is literally just to get Siri right as a true personal assistant. It's like this theme I've been talking about lately called personalized machine learning, which most people aren't talking about. They're talking about communal machine learning.

Even Apple's doing it by saying, "Well, we'll know what's trending in words or emojis." I'm like, "That's all fine and good, but I want something to truly know me inside and out and become the most valuable assistant I can have in both work, and life, and everything else."

That's the battleground that makes super sticky from an ecosystem standpoint around Siri. That's the thing if they can get that, if Siri can move in the direction of really becoming a smart assistant that's helpful to you in life, then that's the kind of thing that cements this from an ecosystem standpoint.

Then, to my point, is that the HomePod would become an accessory to Siri. That is ideally where you go because that's essentially where you need this interface for ambient. You need this intelligent interface that's both unique to you but might also know enough about your family and your family dynamics and things like that to be truly valuable.

If they can pull that off, they're going to be in a really, really deep spot in terms of just keeping this loyal base even more loyal than it is today.

Rene: I think that's very true. Switching topics just very slightly, Apple's "Machine Learning Journal" has been putting up a lot of entries lately around this topic. Yesterday, they were talking about differential privacy -- at least yesterday based on when we're recording, we're talking about differential privacy.

We had a short discussion on Twitter about how...Right now, Apple and Amazon and Facebook are huge ingestion engines. They just harvest data. They use that data to make better products.

Where Apple has taken a stance and made it a differentiator that they don't want our personal information, that they'll go to great lengths to collect as little as possible. When they do collect it, to collect it as privately as possible and then to get rid of it as quickly as possible.

That runs counter to what you want which is Siri to take that information, learn from it, and become better at doing things for you based on your personal information.

Ben: The challenge here is...I think again when Apple's using differential privacy as a technique that they're employing machine learning on, they are talking about broader social trends. They're saying, "What are the things that habitually we need to understand about our users?"

It's essentially saying, "We need analytics, but we want to do so anonymously with as little bit of data as possible and then use that then to better make our software and things like that." There's a communal side of machine learning. That's talking about the communal side of machine learning.

Then there's the personalized machine learning. This is the most difficult thing because if you think about I want Siri to know everything about me, I'm fine with my iPhone and Mac OS and iPads and whatever to know my calendar, to know my email, to know who I'm talking to, to know the context of that, whether it's work or personal, and start to act on my behalf.

That doesn't necessarily have to go public. That could be localized data. The problem is how do you train a machine with that little of data? It's really hard if not impossible. Nobody's doing that today. They need massive data sets, terabytes of data sets.

I'm asking for something that can actually get personalized and be trained to become my assistant with a much more limited data set. Those algorithms don't exist today. But to me, Apple I think is trying to solve a harder problem which is that. I honestly don't know how they do it.

I do think that to some degree they need a little bit more clarity into that data. I'm not entirely sure that differential privacy is going to work in what I'm just describing, at which point I've always said, "Look, I trust Apple with my data. I trust them to keep my data safe. And so I would prefer them have it, protect it, but use it to make better stuff for me."

I know they're not going to sell it to anybody. I know they're not building a profile about me. I know they're not going to sell ads on me. I have no problem with them having that information to deliver me much more personalized services. If that's an option, I'd take it.

But what I'm looking for is that highly, highly tuned and personalized service. If you look at it today, I've always used this theme of an article I've been constantly meaning to write called "Apple's Machine Learning is Hiding in Plain Sight." If you know what you're looking for, you can see it start to show up in iOS.

You can see it start to show up in automated responses. You can see it start to show up in some of these clever things it'll do around location or bringing in the last thing you were searching for on Yelp and having that be a first response to a person when you say, "Meet me here." It's just really smart stuff. If you know what you're looking for, you see it.

But what's happening is I think iOS is essentially learning about me. Apple isn't just building a flat, vanilla operating system. They're building an operating system that has the underlying frameworks to become customized to its user. It's doing that right now in a text-based fashion.

The next step of that will be how does Siri start getting involved by surfacing some of these things to me or pushing some of these responses to me or alerting me of some of these things intelligently?

Like if you remember one the best ways we had talked about the Apple Watch notifications early on because all of a sudden now we had a screen. It was like a polite tap on the shoulder, "Hey, I think you should know this." There's so much value in that with the Apple Watch.

The next step of that is Siri to say, "Hey, you know, you got an email from one of your colleagues, and this just came up, and it's urgent. How can I help you take care of that?" Then you start now interacting in this much more elegant and efficient fashion. But that requires again very, very personalized, very customized types of experiences that's not communal. It's unique to me.

IOS is becoming this highly personalized learning operating system that starts to become very, very customized and unique to you so that my experience with iOS in terms of something is a little bit different than yours because we're different people with different habits. Now Siri then becomes the middle agent of that that's helping add more value in all of these ways that it's learned about me.

Like I said, that's a small data set. That's just data on Ben. That's not data on common things Ben does that's common with everybody else which is what they're doing today. That's a hard problem.

To be honest with you, neither Google or Amazon are even remotely positioned to do that because Apple's the closest to the user in their ecosystem. Google's not. I use very few of their services. Facebook sees a facade of me. Facebook doesn't actually see all of me because I'm very selective of what I share. Amazon only sees my commerce behavior.

Nothing actually has a holistic picture of me except for iOS. That's why I think they're well positioned for this. But I don't want to see them not succeed on this because they're purposefully trying to not know anything about me when I'm like, "Look, please know about me and deliver me more personalized stuff."

Rene: The thing that I find so interesting is that Apple often starts assembling different technologies, sometimes intentionally, often though it's just serendipitous. When extensibility was created, it eventually became an incredibly important core technology. But that was because, "Hey, we need something. Oh, we can use this. That's great."

We see those pieces now whether it's the original purchase of Siri or it's things like the secure version of CloudKit that they're using in place of iCloud to handle things like face sync and eventually messages sync because they don't want those going over more open channels. You were starting to build this network.

Because one of the problems we had previously was when everything was on device and not syncing, you spent all this time training it. Then you'd get your new iPhone. You were starting over from scratch again.

We're seeing those pieces with SiriKit and secure CloudKit and all of those things. Three years ago, starting to build a neural networks on silicon, we're starting to see those things come together.

Ben: If you notice, because I started paying specific attention to this because of this data collection point, I think the last three major operating systems, every time that you agree to send analytics to Apple, it includes more stuff. They're essentially widening the net that they're getting with each time that you opt into that.

I would argue that most people opt into that. If you held a flame to me to guess, I'd say it's north of 80 percent. People agree to share with developers and to send analytics to Apple. They're crawling more stuff. It is iCloud now. It is parts of your email. It is some things.

They're using that with differential privacy but to make these points. It's like with each release, they're actually getting more and more and more data. Yes, maybe this is again like a lot people to say an unnecessary move to hamstring yourself when your customers would delight...

Most of them would happily hand it over. It's fine. But the point is that they are I think increasing even with differential privacy.

It remains to me again an open question as to how far differential privacy will really take them. It's important for some things. I'm not entirely sure it's important for everything. But you can already see them making some tweaks in that strategy about, one, how they use it and also how much data they're then getting to apply that to. It's gotten bigger each time.

Rene: It's super interesting because a lot of the technologies are not always applied in obvious ways. Tim Cook has nebulously said that Apple uses AI to improve battery life.

But some of the interesting things that they do under the hood, I think this stuff is incredibly [inaudible 26:50] , is, for example, let's say every morning you look at Twitter first, then you look at Instagram, it starts to understand that. It will start preloading Instagram once you start looking at Twitter because it believes you're going to go to that next.

If you don't, if you throw it a curveball, if you go to messages next, it's fine because that's how a computer's supposed to run. It's not supposed to know what you're doing. The worst case you get is default computer behavior.

But if you really do go to Instagram next, then it's not awake from sleep state anymore. It's not a slow load state. It's like instant performance and all the power efficiency that go with it. That's from learning your behavior and starting to do almost prescience which I think is one of the things that are super interesting about where Apple is taking machine learning.

The benefit is you get much better performance, much better power efficiency. The worst case is you get the same power efficiency and performance you always got. It's so under the hood that I think it goes completely unnoticed.

Ben: Like I said, that's why I liked the title of my article which is It's Hiding in Plain Sight because it's a lot of stuff that you don't know until you're looking for it. It's the little optimizations that get even better now that they're tying so much more of this in dedicated silicon and new algorithms that they're doing to efficiency manage and use that silicon.

That'll just keep going down their roadmap because of how good they are on the semiconductor side. It just keeps getting better and better and better. But that's why I said you can see how they're building this software that's learning about you, how they're building an engine that basically makes that faster.

But again, that's I said, the hard part here is just that they won't be working with a lot of data to get to know me. That's not what machine learning does. If they can crack training a system with very limited data sets that's still highly personalized to a user, they're onto something that I don't know what anybody else is doing along those lines.

Rene: The other thing that I'm still really interested in, and I know Google and I think Amazon too have started a progress towards it, but one of the early issues that I heard about with HomePod was the idea of multi-personal assistants. Because Siri right now and probably all the other ones when they started were personal assistants. But HomePod is a family product.

Even Apple TV is a family product. We saw Apple struggle with how do we handle iCloud? There's no messages for example in Apple TV.

But it needs to get to the point where if you ask for your messages, it knows it's you and gives them to you. But it will not give you your wife's messages unless she asks for those messages. It starts to understand that it is a personal device in a semi-public or at least familial setting.

Ben: That's always been one of Apple's challenges is that their focus had always been on the individual. The home is so different in that it is a very communal environment.

But again, I don't even think that Google and the Echo have actually solved this well to the vision that we can all create about it truly knowing us and delivering us the information that we specifically want, not what somebody else in our home wants. Nobody's there yet. But that's really I think the challenge.

It's not been one that Apple's been typically good at because they like to deliver stuff specifically to you. The communal bit I think will be a challenge for them. But it's one that I think they're vested to solve. I hope they're vested to solve.

Because again I really do think that the home is an important battle ground for consumer electronics. That extends beyond things that we have on our person and wear on our wrists or faces or ears.

I want to see them make a go at this and do so with a bit more of a strategy than just what's positioned right now around HomePod which is that's geared a little bit more toward high-end audio as the differentiator. It needs to get broader than that.

I think they have time. But the aggressive pricing that an Echo and even Google Home but really Echoes are going at, it doesn't really help them because they just continue to sell more. The question's going to be what would get them to switch from what they're happy with today to something else, and would they pay a dramatic premium for that or not?

Rene: That's absolutely true. Just to round us out and finish this up, that Apple did not ship HomePod this year, do you think that's a hit to them? Do you think it makes it harder next year?

Ben: No, not because of their positioning. The way that they're positioning this and the way that they've got a loyal enough base for, I think they've got time to chip away with this. The concern is that somebody who's buying one of these things for anywhere from 30 to 80 bucks is not going to buy a HomePod.

Apple either has to address different price points, which they probably will, or hope they've layered enough value on this that they can get people to spend $340 or something around there as pricing can vary over time or not. That's the challenge.

They are unfortunately for a mainstream customer going up against people who are perfectly happy with the capabilities of their device for 50 to 80 bucks. That's [inaudible 31:53] .

There's a decent chunk of the market that's happy with HomePod, but I would prefer Apple to have a much more deeper we want to win every room of the home because we know how important the home is as a battleground. Right now, that's not the strategy for Homebot. It needs to become a bit more robust if that is something that they care about.

Rene: Very true. Ben, if people want to follow you on Twitter, or they want to just read more of your articles, where can they go?

Ben: At @BenBajarin on Twitter, and Tech.pinions is the site that I started and write for with a number of colleagues so pretty much most of my written stuff is there, and obviously, I pontificate a lot on Twitter.

Rene: [laughs] No, it's all fantastic stuff. Thank you so much for joining me. I really appreciate your time.

Ben: My pleasure. Glad to be here.

Rene: You can find me @ReneRitchie on the Twitter, or Instagram, or the social things. You can email me at rene@imore.com. If you haven't already, you can subscribe to this show, even leave a review if the mood strikes you on Apple podcast, on Castro, Overcast, Downcast, all the casts. I want to just thank you so much for listening. That's the show. We're out.

[music]