Empower Apps

Damian Mehers shares the story of his app Voice in a Can from inception to shutdown. We talk customer stories and the future of voice assistants.

Guest

Announcements

Links

Related Episodes

Social Media

Email
leo@brightdigit.com
GitHub - @brightdigit

Twitter
BrightDigit - @brightdigit
Leo - @leogdion

LinkedIn
BrightDigit
Leo

Patreon - brightdigit

Credits

Music from https://filmmusic.io
"Blippy Trance" by Kevin MacLeod (https://incompetech.com)
License: CC BY (http://creativecommons.org/licenses/by/4.0/)
  • (00:00) - What is Voice in a Can?
  • (11:46) - Working with the Alexa Voice Service
  • (16:25) - Customer Stories and Shutting Down
  • (21:21) - Future of Voice Assistants
Thanks to our monthly supporters
  • Bertram Eber
  • Edward Sanchez
  • Satoshi Mitsumori
  • Steven Lipton
★ Support this podcast on Patreon ★

Creators & Guests

Host
Leo Dion
Swift developer for Apple devices and more; Founder of BrightDigit; husband and father of 6 adorable kids
Guest
Damian Mehers
Swift, SwiftUI, .NET MAUI, C# and more from Crans, Vaud, Switzerland. My apps include Voice in a Can which brings Alexa to the Apple Watch and Vision Pro: http://viac.app or http://viac.app/testflight and more https://damian.fyi/apps/

What is Empower Apps?

An exploration of Apple business news and technology. We talk about how businesses can use new technology to empower their business and employees, from Leo Dion, founder of BrightDigit.

[00:00:00] What is Voice in a Can?
---

[00:00:00] Leo Dion (host): Hey folks, thank you so much for being patient with me. I'm in the middle of moving as I'm gonna keep pumping these episodes out. These were all recorded last month I will be speaking at the Server site Swift conference. . You can use the promo code "EMPOWER APPS" to get 15% off the conference. Check the show notes for more details. that's it. All right enjoy this episode with Damien. Bye.

[00:00:20] Leo Dion (host): Welcome to another episode, empower Apps. I'm your host, Leo Dion. Today I am joined by Damian Mayers. Damien, thank you so much for coming on.

[00:00:34] Damian Mehers (guest): My pleasure. Thank

[00:00:35] Leo Dion (host): Yeah, I'm really excited to talk about your app. The Life and Death of Voice in a can. We've done a. A couple of episodes where people would shut down apps or sold off apps. We infamously did our episode with Christian about Apollo. But I'm really interested in talking about voice in a can and how that whole ecosystem panned out.

[00:01:00] Leo Dion (host): But before we get into that, I'll let you go ahead and introduce yourself.

[00:01:04] Damian Mehers (guest): Sure. Yeah, I'm been programming since I was 14 years old on my one K zx 81 19, 19 81. I'm 57 years old now. Seen, I've seen the industry change rather a lot, started programming on in basic Ada C plus C, and Java as well. Of Swift and Swift in my own time, in my day.

[00:01:30] Leo Dion (host): Okay. Awesome. How did you said you it was a Zed. What what do you have experience with outside of the Apple space in your whatever, 40 years of programming.

[00:01:43] Damian Mehers (guest): Yeah, I like I said, I started programming on these kinda APIC computers, these at 6 5 0 2.

[00:01:47] Leo Dion (host): Okay.

[00:01:48] Damian Mehers (guest): And I, that was when I got my, my my start. And I'm pretty convinced that sitting in front of my BBC computer at the time trying to. That was actually the not because I wanted to do to get them, to not pay for them, but because I wanted to have infinite lives, but spending hours and hours debugging and decrypting, and disassembling the code. That kind of taught me how to program because it, it gave me the sense of.

[00:02:12] Damian Mehers (guest): Ability to focus for immense quantities of time. Believing that one day you'll actually get what you want do done. So anyway, that, yeah, and, but my first kind of serious job was on, on computers from a company called Deck Digital Equipment Corporation. As a student, I went to work for them on the axes and and at the time it was amazing when I got that internship, I thought I made it because, was the biggest, second biggest computer company in the world.

[00:02:33] Damian Mehers (guest): It.

[00:02:33] Leo Dion (host): Yep. Right after IBM, right?

[00:02:35] Damian Mehers (guest): Exactly. But now of course they've gone, so it's a little bit of a warning in the back of my mind. You working hotness and the greatness, but you never know.

[00:02:42] Leo Dion (host): Yeah. Yeah. Okay. What do you do right now if you mind me asking?

[00:02:51] Damian Mehers (guest): Yeah. So until the end of last year, I worked as a consultant at Nestle, which are a big Swiss company working on mobile apps, Android, iOS, using zamarin, and the replacement, which is Donnet Maui. And now I'm working for a private bank in Geneva, again, as a consultant doing Donnet Maui. So in my evenings I do c say Swift and Swift I, and during daytime I'm doing C and it can be, it can break my mind a little bit sometimes trying to switch.

[00:03:19] Damian Mehers (guest): Kind of good to, to have a foot in both worlds and to understand. There was some interesting concepts, the whole thing with with awai, that's been there in c for a long time, but more then the new model in Swift is is something that doesn't exist in, in c the runtime or the time I should say.

[00:03:33] Damian Mehers (guest): Sorry. Checking,

[00:03:34] Leo Dion (host): we, we talked a few episodes ago, I forgot with who. It was, about how they're changing the name from Zain to.net, Maui. What is the difference between these two technologies exactly.

[00:03:45] Damian Mehers (guest): Zamarin was the original technology that Miguel Deta and Na Friedman created back in the day. And then Microsoft acquired Zamarin. And for a long time things were looking rosy and great. But then Miguel left and so did Na. He was wanting for a while. I think Microsoft decided they basically wanted to rewrite it.

[00:04:02] Damian Mehers (guest): So a lot of the underlying foundations are the same, but the way rendering for the, for some what was called zamarin forms, which is the kind of the UI little one on both iOS and Android is is being rewritten in Maui. And the big thing there is really that it's.

[00:04:17] Damian Mehers (guest): It's not doing what I think Flatter does, which is to render the UI components themselves, they actually render as native components. So a button in actually renders as an iOS or an Android real button. But so yeah, so it's a kind of an evolution if you like.

[00:04:32] Leo Dion (host): Okay, that makes a lot of sense. So let's get into it. Let's talk a little bit about voice in a, can you wanna explain what this app does and how you came up with the idea for it?

[00:04:44] Damian Mehers (guest): Yes, does is not quite right word. What it did I should say is

[00:04:46] Damian Mehers (guest): Anyway, yeah, it's. So the kind of the, this, the backstory is I was working at Evernote and I'd just been let go along with the CEO and of Evernote Phil Libin and he we had to tea in San Francisco and he was just, we were just talking and he talked about this new device that was coming out had just come out and it was taking the world by storm a little bit, which was the echo, the Amazon Echo.

[00:05:07] Damian Mehers (guest): And so he ordered one from Amazon and.

[00:05:09] Damian Mehers (guest): I took it home and most technology that Damien brings home is not well received by the family, but this time it was actually sat in the kitchen and we used it to play radio. We used to, add items of the shopping list, set timers. So it was actually the rare success in the household that the whole family come got on board with.

[00:05:25] Damian Mehers (guest): And I was browsing around on the internet. I was I had a long commute working in.

[00:05:33] Damian Mehers (guest): I was looking around to see if Amazon had any APIs, and sure enough they had an actual it was called the Amazon Voice Service, which was designed not only for device manufacturers, who wanted to create their own echos, but they specifically allowed apps as well. So I started playing trying to basically send an audio recording to their service and get back a.

[00:06:00] Damian Mehers (guest): Amazon has some very specific requirements. Like you, you can't send audio through you prerecorded. It's gotta be live streaming from the microphone. So I was doing some pretty low level stuff, to, on the audio devices to grab the sample audio samples and stream them up. And once I got working on the iPhone I then there's actually real no real point in creating an Amazon client for the iPhone because, but.

[00:06:25] Damian Mehers (guest): Up and running on the Apple watch as well, because Ammar was supported on the Apple Watch. So that's, but this was the very first apple Watch. Maybe the Apple Watch two was around then as well. It was very slow. Horrendously slow. Yeah. Yeah. There was no Swift ui. It was all, it was, I.

[00:06:42] Damian Mehers (guest): There was very little ui actually, it was mainly the engine underneath I got the app actual running on the Apple Watch, but it was horrendously slow because it was very early hardware.

[00:06:50] Damian Mehers (guest): But but it was, I was delighted. And the nice thing was I was able to reuse a whole engine that I created in c for the, on the Apple watch on the iPhone. I was able to get it running on the iPhone, on the Apple Watch. Sorry. And so that was and eventually I released it and you never know what's gonna happen.

[00:07:04] Damian Mehers (guest): But but Tom Ver, Tom Warren from the Verge picked it up and wrote an article about it, and that was and your wildest dreams come true. And then you realize maybe it was a bit of a nightmare because the first version is never that, great. There were always issues. And so my sales did I think I made about $3,000 in the first day.

[00:07:21] Damian Mehers (guest): And part of me was thinking, Ooh, I can retire now. I've got, it's gonna carry on like this forever. But of course it didn't, it dropped down to, I think I always paid for my lunch, so 10, $15

[00:07:32] Leo Dion (host): how'd you monetize it, if you mind me asking?

[00:07:34] Damian Mehers (guest): No. Yeah. What I did at the time was a one-time purchase of $2.

[00:07:38] Damian Mehers (guest): So that was the, the model. I don't, it's such a difficult thing to know what the, the right thing to do is generally, if you've got people complaining about the price and people saying it's too low, then you're in the right spot. And I did have, I didn't have many people complaining about the price, but but I probably could have charged more, but it really wasn't a thing to, to make lots of money.

[00:07:56] Damian Mehers (guest): It was as much a passion project as anything.

[00:07:59] Damian Mehers (guest): So I didn't yeah, going today, maybe I would've done the $2 a year or something, but

[00:08:02] Leo Dion (host): What was the most, okay, so did you do this all on C?

[00:08:07] Damian Mehers (guest): Yes, I did it all in C Yes, I did initially all in C until I rewrote it in Swift and Swift.

[00:08:12] Leo Dion (host): Okay. What let's talk about that. What was that experience like? Switching from Zarin on C over to

[00:08:18] Damian Mehers (guest): Yeah when after Microsoft acquired Zamarin, they decided they were no longer gonna support the Apple Watch for for Zamarin. And that was a, there was a at that point I had, because Zamarin ran pretty much everywhere. I had it running on my Xbox, on my Samsung watches on, on Android watches

[00:08:36] Damian Mehers (guest): pretty much everywhere you can imagine.

[00:08:37] Damian Mehers (guest): But I, but really all the money, all the intention was. Was coming from the Apple ecosystem, so I decided that it might be a, it was a good time because they, I couldn't create an Apple Watch app anymore using zin. That was a time to to rewrite it. And fortunately, swift UI had just, was out by then.

[00:08:55] Leo Dion (host): Yeah.

[00:08:55] Damian Mehers (guest): So the experience was, a complete rewrite of thousands of lines of code. But I did the, I'm happy, the approach I did in the end, which was to create tests where I basically fed input to my C-sharp engine and captured the bite stream output. And then I, when I was building the Swift version, I had all these unit tests

[00:09:14] Leo Dion (host): Oh wow.

[00:09:15] Damian Mehers (guest): that I had backward compat, there a bug in the old one on the.

[00:09:20] Leo Dion (host): It was the same

[00:09:20] Damian Mehers (guest): you are, working at a quite low level, that I was that it was working. So it wasn't, it was as much about learning Swift as it was about, c and learning the the the patents, if you like, of the new programming language. But but even.

[00:09:35] Damian Mehers (guest): I think Don do net's framework libraries, the stuff that just basically comes for free was much richer than the much more evolved, I would say, maybe more clearly more mature than the swift ones. So there were some things that I could do, I can't remember off the top of my head, but things I could do in C Sharp, that were just, that were built in that, that weren't quite as easy and swift but I got there in the end.

[00:09:55] Leo Dion (host): What were one of the things that you could do in C that you

[00:09:58] Damian Mehers (guest): I wish I could remember off the top of my voice. Off the top of my head. Sorry. There's something called Link in, in, in C, which is language integrated query, which is really nice. And yeah, and it's not as rich, you've got, I've usually got map and filter, in, in Swift, but it's not I guess some of the some of the

[00:10:15] Leo Dion (host): Because basically what you could do is you could do like a SQL query on a data structure essentially, and can figure out all the functional programming for you that you would essentially have to hand write and swift Yeah, no, I

[00:10:27] Damian Mehers (guest): It translates into a meth chain. It's synthetic sugar. It looks in a th blood to

[00:10:31] Leo Dion (host): Like I would say like at Predicates are probably the closest to

[00:10:35] Damian Mehers (guest): a little bit.

[00:10:35] Leo Dion (host): Yeah. But not, but a lot more robust I would

[00:10:38] Damian Mehers (guest): yeah. Yeah. It's been there, I think since 2008 or something like that. It was a long time that Swift had been there.

[00:10:44] Damian Mehers (guest): Yeah. I also, but at that point, Amazon were releasing something called the Alexa Presentation language, which was a UI for the Echoes, for the new Echoes. And I wanted to add that to the to my new app. And so basically the way I worked was that the Alexa sent down a service, sent down a.

[00:11:04] Damian Mehers (guest): Structure, which describes, all the elements and the rea and the event handlers on all sites, sorts of stuff. I initially had dreams of implementing my own pasa for that, but it was, it's really rich and therefore complicated. But there was a library from Amazon, which I could basically feed in the JS O and it would back events and.

[00:11:28] Leo Dion (host): Yeah.

[00:11:29] Damian Mehers (guest): Unfortunately that was written in c plus, and at that time there was no swift to c plus integration there is now. So I discovered that lights have objective wrapping, c plus calls in objective C and then calling them hit from Swift. So I did that for basically all the stuff I needed to do and essentially I was.

[00:11:46] Working with the Alexa Voice Service
---

[00:11:46] Damian Mehers (guest): Dynamically generating the Swift UI on the fly. Not a compile time, but a runtime dynamically generating the ui. Maybe the shopping list contains a, a frame with, toggle elements that could be swiped, and then I was actually getting the Jag gesture feeding that back to the Alexa presentation or the Alexa presentation, language library, and getting.

[00:12:06] Damian Mehers (guest): But it all worked amazingly well, and it worked on the Apple Watch as well, so it worked well. So that was yeah, so that, so the whole experience was, part of it was really just the delight of learning something new in, in Swift and Swift ui and, and then also it's true, it's definitely the case that building native, especially if you're working at quite a low level, is easier, than than using something like c and summary.

[00:12:31] Leo Dion (host): This watch app, was it like an independent watch app? Was it a companion app?

[00:12:35] Damian Mehers (guest): It was a,

[00:12:36] Leo Dion (host): component?

[00:12:37] Damian Mehers (guest): yeah it was a it was a companion nap. I don't, I think I might. I don't remember if I decoupled it from the watch app, from the phone app, but it worked totally independently. I was out for a run, in the middle of the fields with just my Apple watch listening to podcasts, and I could add items to my shopping list or turn on my electric blanket, whatever I want to do, using the Echo voice in a account on my Apple Watch using the the data connection.

[00:13:02] Damian Mehers (guest): So it was independent from that perspective.

[00:13:04] Leo Dion (host): You also build a Vision OS app. Is this correct?

[00:13:08] Damian Mehers (guest): Yeah, that's right. So voice on a can was doing well and the vision, apple Vision Pro was announced. I'm sitting in Switzerland. There is no Apple Vision Pro available. I applied a few times for the for the events that they were holding for developers, but it never I never got accepted, but but still, but I still, I had this vision.

[00:13:24] Damian Mehers (guest): Forget the, forgive the pun of coming into an office, like I'm sitting in a, completing your office. I put on my Vision Pro and I just pick up my virtual echo and place it on a desk, next to me. And then whenever I want to use it, I just look at it, tap my fingers, it lights up, and I can say, add things to my shopping list or, turn on my electric blanket, or, and then my vision was that the u the UI would pop up.

[00:13:47] Damian Mehers (guest): From the unfer, from the Virtual six cylinder. So I got all this working in the simulator and it wasn't that complicated a Vision Pro app. It was basically a, in terms of the ui, it was a cylinder, that I like when you look at it, it lights up. I did take the tap gesture and I changed the UI to the cylinder's rim to light up to the listen.

[00:14:09] Damian Mehers (guest): And then I was able to project a Swift UI view, for example, the shopping list, you could scroll the shopping list in, in the echo, in the virtual Alexa, wait, I'm gonna set off my Alexa. But and then it would go away. So I got it all working and and in the simulator. Never got a chance to to try it out for real until, I dunno, a couple months ago when I actually got a Vision Pro.

[00:14:30] Damian Mehers (guest): But but I, it was accepted by Amazon. You never know. Sorry, apple, it was accepted, you never know. But but they accepted it and and so I was there launched a, so when the. Developer conference came up and they showed that massive wall of all the icon vision on the far right side, half my icon visible.

[00:14:47] Leo Dion (host): Your head was peeking out. Yeah. That's hilarious. Speaking of like getting accept, what's the different, like kinda what's the difference in character between App store review and Amazon accepting your Amazon like Alexa app? What's

[00:15:03] Damian Mehers (guest): I was, yeah, I was really early when I released my app. So if for the echo, there wasn't approval of process, but I just sailed through however, and then said they accepted it and I could, I created all my other ones. Never an issue. However, driving home from work one day got a notification on my phone from something about legal, Amazon, blah, blah, blah. I couldn't look at it. Of course, I drove on until I could pull over safely and looked at it and wasn't sure enough. A, an email from a lawyer from Amazon with a long checklist of things that they wanted to verify that I was doing. For example, that I wasn't saving the audio locally all this kind of stuff.

[00:15:40] Damian Mehers (guest): But I was playing it by the book and we exchanged emails backwards and forwards. Stern, but it got friendly towards the end. And the final thing you said to me was how cool he thought my app was. The lawyer said that. So that was nice. So all in all, I would say it was pretty straightforward from the Amazon perspective, from the Apple perspective.

[00:15:55] Damian Mehers (guest): When I had my RIN version of the app, I tried to release it for the Mac and they, they rejected it and they said it was, I can't remember what the reason was, basically too generic or something. I can't remember what, but when I rewrote it in Swift and Swift ui, I did it for the iPhone, for the Apple Watch and the Mac.

[00:16:12] Damian Mehers (guest): And then again, it went through without any problem. So as we know, it depends on the day, the reviewer you have, so it's just the way it's, yeah. Apple still a bit of a we all know the stories, It be a bit, you just have to be lucky,

[00:16:25] Customer Stories and Shutting Down
---

[00:16:25] Leo Dion (host): yeah. What so you got this email from your, this lawyer, were there any other emails or contact you got from customers that made you think, wow, I'm really onto something.

[00:16:35] Damian Mehers (guest): Yes. Yeah, early on. The first hint of it was somebody who was blind. I did actually, I think I put accessibility in there, some voiceover stuff. But I hadn't done a very good job. Somebody who's blind contacted me and were very kind. They gave a i a two star review, but they contacted me, which is the main thing.

[00:16:50] Damian Mehers (guest): And, really well as far as I could. So that was initial, that was a little bit of a wake up call that, oh, this little fun thing I'm doing on the train, going to work every day. Maybe it's actually a little bit more meaningful than that, and then I started to get emails and one in particular still haunts me.

[00:17:05] Damian Mehers (guest): Hs, not the word. Yeah, maybe it, haunt is the right word, but it's a guy, stage four. Cancer veteran who's basically thanking me for being able to I interacted with him a bit to get him to help him. But in the end he's, he says, I have stage four cancer. I'm gonna, what did he say? Fight it or die trying or something. I can't remember. Anyway, it's quite poignant and and he was using my app to control all his home devices and he had load loads of them. So that, that really changes things from being a fun little thing that's paying my lunch, and eventually it's paying my mortgage.

[00:17:34] Damian Mehers (guest): That really put things into perspective that you don't know, you know how people are going to use what you're doing and accessibility. Everybody says yes, it's very important, but it really is important. You don't, and you never know how people are gonna use what you're doing. So it's not a should never be an afterthought.

[00:17:48] Damian Mehers (guest): It should be built in from the beginning. And I try to do that now when I create my apps.

[00:17:54] Leo Dion (host): That's

[00:17:55] Damian Mehers (guest): Does that answer your question?

[00:17:56] Leo Dion (host): Yeah, no, I'm just, I'm reading on the webpage what he said about how I'm not a whiz, but I have all these smart devices to help me as the cancer gets worse. So far, I have 29 smart devices installed, which I can control with my voice, iPhone, iPad, and laptop. And now Apple Watch.

[00:18:12] Leo Dion (host): Thanks to you. That's, yeah, that's really awesome.

[00:18:16] Damian Mehers (guest): I have to say I was in tears when I read that

[00:18:18] Leo Dion (host): Yeah. Yeah. I would imagine that would be the case. Yeah. It's like we always think these things are like little toys and then you see somebody actually use it. Like, why did I think of that? That's amazing.

[00:18:28] Damian Mehers (guest): It was obvious that, that the people would be using it in that kind of situation. But at the time I didn't think of that.

[00:18:34] Leo Dion (host): Yeah. If only sir. Worked the same way.

[00:18:35] Damian Mehers (guest): Wait for Apple Intelligence, I'm

[00:18:37] Leo Dion (host): That'll fix everything. It fix it for me. Not for you. 'cause you're in, that's.

[00:18:41] Damian Mehers (guest): I'm, but not in the European Union. I have not actually looked to see, Switzerland isn't part the European Union, but I dunno if Apple knows that. So we'll

[00:18:46] Leo Dion (host): Yeah, that's true. That's true. Yeah, so it's going well. Everything's going great. It's all blue sky, right? There's nothing

[00:18:55] Damian Mehers (guest): No. Yeah. Until, I think, yeah, towards the end of last year, I did get an email from Amazon saying something about not supporting the Amazon Voice service going.

[00:19:06] Leo Dion (host): Okay.

[00:19:07] Damian Mehers (guest): And I thought, and they assured me that, or it was a general email, but saying that it was going to be still supported. They'd support existing devices.

[00:19:15] Damian Mehers (guest): And then I started to get several emails from different people saying that they were getting an internal service error from the Amazon services. Of server when I was, when they were using my app, something new, and at that and that point, and then I looked and joined off all the Amazon Voice Service documentation was gone.

[00:19:33] Damian Mehers (guest): It never existed, so as far as, so all my links were broken. And at that point I real and it was clear, it's clear that Amazon are refocusing, they fired a load of people and in Amazon they've, I think $25 billion is the whole thing on the verge about it. They spent on it and they're really trying to work out how they're gonna make money.

[00:19:49] Damian Mehers (guest): And clearly my kind of app isn't their focus anymore. It was, I visited the Amazon offices and I met some really nice people there, so I, so there some good times early on, but it's changed. And it just, sometimes you know what the right thing to do is, and it just felt like the right thing to do was to pull it from the app store before things got worse, even though it was still making not enough money to pay my mortgage every month, so it wasn't nothing but it was, but it just felt like it was, things were only gonna get worse and I didn't good conscience still sell the app when I knew that things were probably gonna go downhill. So I just made the decision to pull the app and I felt much. Afterwards you know about doing that.

[00:20:30] Damian Mehers (guest): I'm still getting emails from people now saying, oh, you gonna bring it back? What's happening? But I have a form letter now that basically says, it just doesn't look very likely. If Amazon decided that they really, oh, we need Apple Watch app, they reach out to me, I would be absolutely delighted, to work with them.

[00:20:45] Damian Mehers (guest): But I just wanted to be sure that the customer experience is as it should be and I just don't feel right about it now. Yeah.

[00:20:51] Leo Dion (host): Yeah. It's, I wouldn't say Amazon is more friendly than Reddit was with Christian because they kinda pretended like it never existed

[00:21:00] Leo Dion (host): weird. But

[00:21:04] Damian Mehers (guest): It's not really personal. They've not really, sorry, they've not, it's not really personal in a sense. They don't probably,

[00:21:08] Leo Dion (host): They That's true. The CEO doesn't lie. And

[00:21:11] Damian Mehers (guest): No, exactly. Yeah.

[00:21:13] Leo Dion (host): Yeah, that is

[00:21:13] Damian Mehers (guest): not been recording per, phone calls with them and playing them back.

[00:21:16] Damian Mehers (guest): Yeah. Yeah. Yeah, it's not it's not quite the same experience. I think I'm just unintended, casual casualty

[00:21:21] Future of Voice Assistants
---

[00:21:21] Leo Dion (host): right, exactly. I. It's so interesting to me that voice was such a big thing. And, things haven't been great for tech this year and, there's a lot of things that are just like getting pulled or getting taken away and stuff like that. It's so interesting to see Amazon because one of the things I remember people loved about Alexa is the fact that you can build so much on top of it. It was a lot more open. And now it's yeah, I Do you like, where do you see voice assistance going forward? Because used to be everything was echo. Oh, the echo's amazing. Why can't Siri be more like the Echo?

[00:22:02] Leo Dion (host): And now it's

[00:22:03] Damian Mehers (guest): Yes, I think, the the Amazon ecosystem is still alive in the sense that you can still create skills for it, so you can still plug into the,

[00:22:11] Leo Dion (host): Yeah, that's true. That is true.

[00:22:12] Damian Mehers (guest): your own, your, my, my kind of use case, creating a, an engine that kind of use the Amazon Voice Service was a bit different, but, yeah, I think we're all waiting to see what the impact is of the gpt. 'cause they've not really hit the the large language models. They've not really hit, the echo yet, although there's talk of it coming, we're waiting for Apple intelligence. So I think, it's a natural fit. And I think it, I hesitate to say it'll change everything, but I think it probably will change things just given my experience using them for programming. Now of course, programming is a specific use case that's very, special and, but I still think it's yeah, it's gonna be a, it's gonna be interesting to see because of the hallucinations, programming you can see on API doesn't exist that it tells you to use.

[00:22:53] Damian Mehers (guest): Yeah, I don't actually, I don't know. I think it's gonna be a big change and I suspect there's going to be a backlash and there's going to be services and apps. Specifically, we don't use LLMs, and that's our feature. We're just a simple, basic app that doesn't try to make things intelligent that don't really need to be.

[00:23:10] Damian Mehers (guest): But the voice I say, I still use the Echoes everywhere and the family still uses them. We use them, for our shopping list, smart home control. And that for me, they're definitely in my use case, in my scenario, they're more powerful than home kits, in terms of controlling my devices.

[00:23:24] Damian Mehers (guest): Yeah, I think let's see. Ask me in a year's time, once the language model, large language models have really had their impact.

[00:23:30] Leo Dion (host): And it, do you think the apple, or excuse me, the Alexa voice service, is that just something they don't want to maintain anymore? Or is it

[00:23:37] Damian Mehers (guest): N no. I don't. I just, I don't know for sure, but there's a guy from Microsoft, I forget his name off the top of my head. He was basically left micro Microsoft. He was head of Windows like devices. He just joined, he joined Amazon. It'll come to me as soon as we finish the call, it'll come to me. But anyway, he's clearly been given the remit of making money from them and turning it into something, that makes money.

[00:23:56] Damian Mehers (guest): So I, I think that they're going to, my, my guess is that they've basically been given one more shot to make money, so they're going to be coming, I think they'll come along with a paid version of Alexa with which will be using a large language model. Yeah. And and they'll be trying to make money whole of.

[00:24:12] Damian Mehers (guest): By the way, did you know you could buy toilet paper? You know that all these are things that came back with, it was really frustrating. I just hear everybody complaining about the, by the way, promise that the Amazon, the Alexa came up with, so that way of making money didn't work. So I think they're going to try and I think they're gonna have one more shot of making money.

[00:24:28] Damian Mehers (guest): But I think the ecosystem is so big and they've sold so many of these devices that.

[00:24:35] Leo Dion (host): Yeah. Do you think do you think there's a I guess we talked about it being a future voice assistance. You think those are just gonna have to be attached to LLMs of some sort, whether it's Apple Intelligence or Google

[00:24:48] Damian Mehers (guest): I think people are going to expect it. They're going to expect the, the more you're not gonna be, you're not gonna expect to have to phrase your. Your queries in very specific ways, which you do now, you're gonna expect it to understand, the context and the gist and the, sometimes I'm programming and I just say to, to JGPT I dunno, swift invert list.

[00:25:09] Damian Mehers (guest): That's all I say. And it, automatically generates, the code for me. So it's a. It's much more intelligent. Although the echo, you can say things like five minute timer. You don't have to put all the ceremony of, please set up blah, blah, blah. It'll work. But but I think the gp, the language models are gonna change how we interact with those assistants.

[00:25:29] Damian Mehers (guest): I think.

[00:25:30] Leo Dion (host): Do you have the slightest desire to jump back in and do anything with Alexa? Programming

[00:25:36] Damian Mehers (guest): No. I've moved on new shiny things, I jumped onto it when it was a new thing and I created a skill to let you notes to a note and one, one note, and I, I created voice, but I feel like psychologically I've moved on now. Sometimes when I look back at all that code I wrote for this, for that for the engine and for PAing, the APL L and then generating ui, it's all managed.

[00:25:59] Damian Mehers (guest): I should probably open source it or something. But I do have a tinge of nostalgia, but I think I'm the downside, is closing the app and losing the revenue. I enjoyed interacting with people who, had questions about it, but the upside is I free up all my mind, the opportunity cost of.

[00:26:20] Leo Dion (host): What before we jump onto what you're working on next, what kinda is your process of shutting down the app per se? Exactly.

[00:26:28] Damian Mehers (guest): I have actually shut it down in two ways. Early on or earlier on when I moved to using Swift, I shut down the Samsung app, which was a, which was popular. I had a discord group, with people chatting about it. I was having sales every day. But I shut that down and I expected it to be the same as as Apple, which was, it would still be there for people that already bought it, but, oh no, it would just disappear from the store.

[00:26:51] Damian Mehers (guest): And if you bought it before, you're outta luck. So that experience was tough because I had to tell these people, I'm sorry, there's nothing I can do with the Apple one. So far it's been pretty smooth. I like the fact that Apple leave, so I basically flip a switch. Double checking everything in the store, and and off it went.

[00:27:09] Damian Mehers (guest): And then I had to update the website and to, to explain what had happened. But it's been, so far, it's been pretty good. I'm waiting for the other shoe to drop in that dunno what's gonna, I expect at some point it will bake for existing customers and, and then basically, that's what I'm gonna get.

[00:27:22] Damian Mehers (guest): The angry emails. People can't leave one store, one star reviews anymore because the store's closed.

[00:27:27] Leo Dion (host): Yeah, exactly. Exactly. So what are you, what's the plan now going forward?

[00:27:34] Damian Mehers (guest): Yeah, I think I said I was working, I was, I'm 57 years old and I've got all this legacy stuff, services I've used, over the last, I don't know, 20, 30 years from emails, from the 1990s eGroup, which got bought by Yahoo and turned into Yahoo Groups. I have all these things, these exports.

[00:27:54] Damian Mehers (guest): So what I wanted to do was create a integrated on. Tens of different services, whether it be Twitter or Facebook or LinkedIn or email, or pretty much any service you can imagine, although the only service that I've used in the past last fm, you name it. And they integrated all, suck it all in and show it in an integrated search timeline so that I could go back on my Apple photos, my WhatsApp messages so I can go back.

[00:28:24] Damian Mehers (guest): I made in the, with a family, I dunno, in. 12 and I can see the places I was at because I'm importing my Google location history. I can see the emails I sent around that time. I can see the Facebook posts. I made the apple photos that I have all the integrated, searchable timeline. So it gives me a way of basically seeing my line, my digital life integrated.

[00:28:49] Damian Mehers (guest): On device. No, no cloud component. And and I, and it's my data, so why shouldn't I be able to own it and search it, so I can see I'm, so the real scenario is I'm connected to someone on LinkedIn and so I search in my app for that person's name, and I can see the emails I sent the conference we're at the same time.

[00:29:05] Damian Mehers (guest): All this stuff.

[00:29:11] Damian Mehers (guest): I'm not sure I'm gonna release it. I don't, I dunno for sure. I, it's in test flight at the moment. I'm still in two minds. As to whether or not to release it because there is a massive, I don't, massive is the right word, but a significant mental cost to releasing and supporting an app.

[00:29:27] Damian Mehers (guest): It's not free. Even if the app was free, which it probably won't be I'll, there's still the support emails and so on and

[00:29:32] Leo Dion (host): How about like open sourcing parts of it,

[00:29:35] Damian Mehers (guest): That's true. That's a possibility. Yeah

[00:29:37] Damian Mehers (guest): It's. Yeah, you, it's a responsibility, to, to release something. And I wanna be sure that it's something I really want to do.

[00:29:45] Damian Mehers (guest): I've got, I know, just under a hundred test flight users, but yeah, so we'll see if I actually release it or not. But the test

[00:29:51] Leo Dion (host): the name of it?

[00:29:52] Damian Mehers (guest): available, it's, so the name is closed ppl. ODI means absolutely nothing. It's just one of those. Domains that we all have, that we dunno why we registered them, but we have it.

[00:30:01] Damian Mehers (guest): And so pz e.com, you can see a video, you can download the test flight and I'll be absolutely delighted to get any feedback anybody has about

[00:30:09] Leo Dion (host): Awesome. We'll put a link in the show notes for folks.

[00:30:12] Damian Mehers (guest): I am thinking about plugging in a local large language model as well, so

[00:30:15] Leo Dion (host): I was about to ask that. Yeah. That's the next thing

[00:30:18] Damian Mehers (guest): yeah. Having said all that stuff about the feature being not no, LM

[00:30:22] Leo Dion (host): You know when you leave this world that way people can talk to you even when you're dead, right?

[00:30:28] Damian Mehers (guest): Yeah. But also, yeah, no, I think I, I, somebody from my family my father, passed away recently, and I, it does think, strike me that it would be a good way to, to be able to pull together someone's legacy, and put it in and have all of his emails and photos and, all the stuff in one place that

[00:30:43] Leo Dion (host): Have you ever read those stories about people who, like they import the emails and text messages to their late partner or whatever, and then it gets

[00:30:53] Damian Mehers (guest): Discover things I didn't want to know.

[00:30:55] Leo Dion (host): Yeah. Yeah. It's

[00:30:56] Damian Mehers (guest): no, there was the Black Mirror episode

[00:30:59] Leo Dion (host): there's a black, but it's been done in real life. Not as, not like buying a robot or anything, you're not getting whatever.

[00:31:06] Leo Dion (host): Yeah. Your boyfriend back ordered from Amazon or anything, but there is LLI, there's article, there's been articles I've read about, girlfriends talking to their dead boyfriend 'cause they

[00:31:17] Damian Mehers (guest): That's not something I'm interested in, but I'm interested in being able to query it, but you, but because the quantity of data is so massive. It's gigabytes. Because email, my email, my Gmail archive is 30 gigabytes. I would've to use retrieval augmented generation to, to, to put, do a local search, to pull in useful information and then do the LLM query on that, because.

[00:31:36] Damian Mehers (guest): Unless the context window gets enormous, I'm not gonna be able to to have all of the data in the context.

[00:31:41] Damian Mehers (guest): So it is something I'm thinking about, but yeah, we'll see.

[00:31:45] Leo Dion (host): Oh yeah. That's gonna be a thing, that's gonna be a

[00:31:48] Damian Mehers (guest): Yeah. I suspect so.

[00:31:50] Leo Dion (host): If there's a VC that's willing to spend money on it, it'll be there. Yeah, that's a,

[00:31:54] Damian Mehers (guest): as as

[00:31:55] Leo Dion (host): Go ahead.

[00:31:56] Damian Mehers (guest): I just remember, Phil Lib in the former avenue, CEO, was saying, people offering crazy amounts of money forever night, but then you end up with a crazy amount of money. Crazy guy on the or lady on the board, and so it is the same with this, VCs, if they offer got of crazy money, you maybe you've gotta, you're dealing with a crazy person.

[00:32:17] Leo Dion (host): Yeah, exactly. Anything else you wanna talk about Damien, before we close out?

[00:32:21] Damian Mehers (guest): one, one final thing I would just say, and this is people creating apps in general and also people who are using apps, is it's really easy to get upset with customer the app developer and then return with the with the customer. And, the one star review would leave zero stars if I could, I wrote a blog about it and I think you really, it kind.

[00:32:41] Damian Mehers (guest): To try and put yourself in the shoes of the person, empath empathic with the person you're dealing with, and assume the best, so even if you get the horrible email from the person wanting all the review, be kind, assume the best and try and help them. You never know what's going on in their lives.

[00:32:57] Damian Mehers (guest): And I've had cases where, one star review has turned into a five star review, and those people are the most rabid or rabid, I should say supporters of your app once they flip over, so anyway I would just say try to be kind, to when you're responding to reviews, even if you really are angry, which we've all been, and upset.

[00:33:15] Damian Mehers (guest): Yeah.

[00:33:15] Leo Dion (host): Yeah. Yeah, that's a great way to put it. Damien, thank you so much for coming on. Where can people find you online?

[00:33:22] Damian Mehers (guest): Yeah. Damian, FAYF yi, sorry, D-A-M-I-N fyi. That has all my my links to everything so yeah, Damian. Yep.

[00:33:31] Leo Dion (host): and we'll put that in the show notes as well. People can find me on Twitter at Leo g Dion. My company is Bright Digit. Leo g Dion at C am is my Ma on. Thank you for joining me and I look forward to talking to you in the next episode. Bye everybody.