Lessons Learned from the History of Telephony - Watch This Space Podcast, May 2021
2:52PM Apr 30, 2021
Speakers:
Jon Arnold
Chris Fine
Keywords:
voice
packet switching
people
technology
big
technologies
jon
telephone
machines
speech recognition
ai
siri
speech synthesis
podcast
thinking
pstn
companies
search
business
telephony
Welcome to Watch This Space, the podcast about future of work. Every month, we bring you insider perspectives on how digital transformation, emerging technologies and generational change are shaping the future of work. We are two analog guys finding the groove for all of this in today's digital world. i'm Jon Arnold. And these trends are my focus as an independent technology analyst for my company, J Arnold and Associates.
And I'm Chris Fine. And I'm an independent consultant and advisor, focusing on enterprise technology, IoT security and the future of work. My company is Integrative Technologies, LLC. Good morning, Jon, how are you?
Yes. Great. Thanks, Chris. And back to you. Good morning. We are, we are an international team here. I'm in Toronto, and you're in New Jersey. And we don't have to cross borders to do this. So this is an easy gig. So let's talk about what's what's happening now. And I think I think the the topic for today is we're going back to school.
Right? Yes, right? Yes, we're going to be we're doing a little experiment, we're planning to use Jeff Pulver's new education platform to teach a little course on the future of voice or to have a little session, we should say, but to try to talk about where today's rapidly growing voice industry came from and where we think it's going. Right, Jon?
Yes. So we mentioned this, I think in a recent podcast, there have been a couple of hiccups along the way. But we finally have a date for this. So if you're interested, just so you know, off the top, it's Wednesday, May 13. It's a one hour course from 11:30 to 12:30 Eastern that you and I will run as a live event and people will attend it's a it's a small payment, it's $20 to participate. So the idea here, and then we'll talk a bit about the topic itself. But the idea that Jeff is pioneering, it's not brand new, we did touch on this on our last podcast, but the idea is that it's short, digestible, but professionally produced forms of content. And that's a bit of an understatement. This is more of a higher level of, you know, prepared information and teaching. So people can learn kind of in a short amount of time, practical information, that they're just not going to get by, you know, scanning the web for articles and headlines and search words, right, Chris?
That's right. And also, we're going to make the case or we're going to point out some examples of where as big as you think this is, it's bigger, where, you know, a lot of publicity comes to things like Alexa, Google Voice, Siri, etc. that are obviously huge markets and big industries now, but when you think about what's going to be done with voice, our thesis is that it is the fundamentally most natural way for human beings to communicate. And the better that machines and get in the more voice and video, certainly as a minimum voice can be used in actual collaboration, the better the better things get in terms of communications. Right, Jon? I mean, it's really growing.
Yes. And and, you know, the the short term for that is, you know, as we were saying in our talk is that voice is bigger than ever. And it should be a head scratcher for a lot of people, maybe millennials in particular, right? Because the last thing they're gonna do with that smartphone is talk to people. And you know, it's I it amazes me, how long is this term, the VA stay in use, when you look at the utility of, of these mobile devices and what they're actually used for? It's not to make phone calls, that's for sure.
Well, phone calls are always available. I guess this is the way I see it is that the technology of being able to connect to people directly or a group of people? Heck, it gets better all the time, but it's been around for quite some time. And Voice over IP revolutionized it at the costs with zero etc. But what's really revolutionising voice is the interaction with machines. And that is where a lot of the action is going to be. Because if you think about how we communicate with machines, it's always been in a way that we've done it for so long. We think it's natural, but it's not like typing. And that's not how we would communicate with people. So what's really interesting is all of the work that's being done to make that sort of the defacto way that you communicate with machines.
Yeah, and you know, certainly the what we were calling in our talk the theme of the new voice. And what that really means is adding a layer on top of what we normally do, which is P2P - person to person. But now it becomes this machine to person or person to machine, use of voice and up till recently, prior to AI being the driver. That form of voice interaction with machines has really been query based. Right, Chris, it's you ask Siri a question. Where's the restaurant? What's the temperature? Right? Those kinds of things where you ask a question, you get an answer. And that's very functional. And that's no different than text search, right? And that, by the way, is a big reason why there's such an investment in this voice technology.
Because the next kind of frontier for the search business that Google builds empire on is voice-based search, mobile search, right, where you're not going to type in a search term - you're going to say it, so the better those engines can be in giving accurate results, the more people are going to use those search engines, the more benefit the advertisers are going to get from, you know, doing their thing on those platforms. So there's a practical business reason for it. But I think where it gets more interesting, and this is where this idea of voices bigger than ever comes from. Now, with AI, we're shifting from not just query based uses of voice but conversational. Right now we're going to engage with machines in a similar way that we gauge when we're speaking like we are now, right, we're conversing. We're talking, we're, we're sharing ideas, we're building on each other's sentences, that kind of thing.
Agreed. And so equally, when we say voice, it's not only voice recognition, but it's speech synthesis. And so both of those have made such unbelievable advances. And when it talks, when you talk about voice recognition, you're right. One of the original uses of this was search and query. And that is going to become the norm, I think. But I also think that voice is getting more and more capable of being directive. And so that the knock on voice always was speech recognition technology was it was hard to get it even, for example, to write down what you said, to translate it into ASCII into the characters, unless it was speaker specific, like dictation systems. Or it was very limited vocabulary, like airline voice powered phone systems that you call. But what's happening is all of this is blending.
And the big ingredient that's allowing this to happen is AI, because speech recognition was always statistically based. But now it's AI based. And the power of that is incredible. Because what it allows you to do, for example, is to become, to let a system become an expert in the areas it has to be not just determining individual speakers, or being speaker independent, but being a subject matter expert, if you want to call it that. And what that allows is directive speech, you know, do this, do that - much more than you used to have. And would people spent decades laughing about that, with Siri and even with Alexa, right, because if it's just the basic layer of speech recognition, that's pretty powerful. But when you add context via AI, it becomes very powerful. And it screens out the error, too. So, it's just a big thing that it's, it's like so getting to be so pervasive, that people don't understand how big it is, I think sometimes.
Yeah, and it also it creeps up on you in the sense that, you know, we're very conditioned already to using these smart speakers in our home. And we don't think twice about it. And before long, it's just it's in the background. And you may not even realize it, but you know, it can be listening and picking up everything you say. Now, there's some privacy and security issues around that, of course, but what happens is, as it becomes just part, part of the landscape, you don't think too much about it, you know, you're kind of it's an implicit sense of trust that you're okay with it being there and it becomes more and more of it becomes more and more of a go-to because the better it gets at giving you what you need, the more you're going to use it.
And the underlying idea there, which is also the AI story, is that I think we've covered this earlier, but that the big value here for machine learning is that it doesn't make the same mistake twice. Right so when you correct grammar, we correct an acronym that it didn't get right, when you correct an inflection of speech for you know, a foreign language or something, it gets it, and then it continues going forward, it will not repeat that mistake. And so the more it gets to be accurate, and it develops patterns of not just recognition, but also engagement with you, the feels natural, it becomes your go-to, and you know, when it's just how we're thinking of using it, like in the workplace or at home, that's one thing.
But when you think about the other use cases of this, for people who are alone, think of seniors who are living by themselves, this becomes a really important lifeline for them, you know, it gives them a way to engage and not just talk to a machine, but it takes on a lot more meaning for them in those kinds of situations. And, you know, people are trying to learn new languages, within you know, real time speech translation, it's pretty these use cases as they grow, just make it more value valuable. I mean, it's the classic network effect with this technology, right?
Yes. And the more research we do for this little session we're having, the more fascinating it gets, right? And, you know, I just didn't want to leave completely your point about telephone calls, because there's still an awful lot made every year in whatever the technology is, or the equivalent of telephone calls. So going back to the original start of the telephone, that what people thought was the miracle of it wasn't even, I can instantly get in touch with so and so, it's that I can speak to them. Without voice - if telegraphy had advanced, let's say it would have gone to where it eventually went to with teletype machines, and other ways of communicating text from one place to another, you know, facsimile, it would have gone there.
But the miracle of the telephone really was voice. And that was that was what propelled that into being the thing that it became, I think, more than simply the idea of getting a message from point A to point B. And ultimately, what happened to the companies that owned the telephone business was they didn't see even the potential, and they tried to hold on to it. And they placed too much value in this huge infrastructure. They built it to do one thing extremely well. But it couldn't do five things. And technologies came that essentially revolutionized and crushed the cost of providing that service. And so ultimately, they, like many other historical pioneers, notionally lost the business in the sense, at least in sense of monopoly. Although what's also interesting is these companies all ended up pivoting to being essentially data companies, whether the data was carrying voice or not. And they're bigger than ever. So if you take what was AT&T in the United States, or Bell Canada, and you add up the components of them now, they're, they're worth more than they were when they broke them up as a monopoly.
Only because they've been able to pivot like you say.
Yeah, because they deliver what became a much bigger market. Right? They were able to use a lot of the infrastructure that they had, and they acquired companies that were further along than they were in, for example, Internet Protocol. But there, it It ended up that communicating was still a gigantic business. It just wasn't just phone calls, as you say.
Yeah, yeah. But that's the Reader's Digest version, of course, because we could go a lot further in the course and beyond. Because, as we also know, they fought these new technologies tooth and nail because it threatened their monopoly.
Yes.
And that's a whole other saga, you know, yeah, aspect of the whole situation there. But you're right, this idea that it's become a bigger business. And that's not the only example. You know, you look at the way, you know how GM, you know that film Who Killed the Electric Car?, right. You know, if they, if they decided back then that EV was the way to go, we would all be driving electric cars by now. But they killed it because it was a threat to their model. Now, it's just inevitable they have to embrace it. So now they're talking about it as if they invented it yesterday, right? It's the same scenario all you know, every industry goes through this.
One of the most. I mean, the whole history of telephony and voice has got a lot of irony in it, as well as triumph. And what was interesting, what I find fascinating about the decline of the telephone, traditional approach to telephony, was that the technology that ended up really killing it, namely packet switching. All of the predecessor research was within The Bell System, within the telephone world, except actually making the breakthrough to packet switching, which was an external project, which was government funded, because one of the vulnerabilities that the telephone system had was that if you hit one place, it could take it out. Right.
So the whole invention of the internet and packet switching was to make it so that if the Cold War enemies nuked one city, the network would just adapt and work around it. And the data would flow. And it was distributed control. Whereas a monopoly is very hierarchical. But a ton of the work that preceded that contract that yielded packet switching, actually came from AT&T, and some of the other companies, Bell Laboratories. And if you look at what underlies all of voice today, possibly with the exception of AI, that's being applied to that, all of the theory comes from the Bell System, you know, the communications theory, the science, the transistors, right? And all of it. So it's just, it's just a really rich story. And well, it'll be it'll be fun to cover that and then look forward.
Yeah, yeah. Well, you know, successor technologies win out because in some way, they're superior, right. They've made some kind of advancement that makes it better. And you know, I'll just say quickly that, you know, VoIP was initially looked at as an inferior technology. So it could never replace the PSTN that was the position that was held for a long time. And because that way, it could never replace it. But as we know, it is actually packet switching - is a superior form of technology in certain ways. And when the network is optimized to support it, the voice quality is actually better than the PSTN, which was an unimaginable concept back then. When we only had dial up and DSL, you know, we didn't have broadband and the speeds you need to do it. So the technology and in fact, was superior. And that's why it's won out in the end, right? Even though it was always positioned as an inferior alternative to PSTN.
Well, that's actually a good thing, we could take it out with a thought here, right? So the same exact thing was true of speech recognition, speech synthesis. Right? So that was always you know, like, they, you know, haha, I said this, and it did that, you know, we're Haha, sounds like the robot, but give something that's fundamentally that useful long enough. And it's going to develop into what it's starting to develop into. And the same exact thing was true with Internet Protocol voice, right, that it was a best efforts, low rent kind of world. But ultimately, it eliminated the old world almost completely. Yeah. So it's just an interesting story. So we'll look forward to talking more about it. Right, Jon?
For sure. And yeah, as takeaways for our listeners and lessons learned, you know, history repeats itself, right. You know, the trajectory that we start with, in our course, about the foundation of the Bell System, etc, what went through what the the lows and the highs and where it is, today is a recipe that we're going to see, as I've said, with the EV - electric vehicles - we're going to see this in other sectors as well. With energy, with healthcare, healthcare is going undergoing an incredible, you know, spurts of growth and innovation, mostly because of these technologies. And we're going to see it in other industries, too. We're certainly going to see we're seeing it now in education. I think government is gonna find its groove with this, these technologies and reinvent itself. And so I think, you know, we should be thinking here about lessons learned, right? What has - what is it, 150 years of telephony history? - and the things we've learned going through that journey? I think there's some good takeaways there for how to think about emerging technologies and emerging opportunities. Right?
I agree with you. And it's always a tough, it's always a risky thing to take a bet. I'm thinking about EVs now, but to take a bet against a technology that either seems like it doesn't work well enough, but it's transformative, or it's too expensive. And it's transformative, right? Yeah. So you look at EVs now, and you can already buy if you have the money, you can start to buy EVs from these independent companies. These the electric cars have to have even more range than a gas car, if you pay enough for the batteries and for the systems but the minute you have something like that, you know that that price is going to come down because they've achieved the science.
Yeah. And by - just look again - like with telephony, it's the last mile, right? That kept competition out of the marketplace. And in the auto industry, it's the same thing. Most of the manufacturers control the distribution network - you have to go to one of their dealerships to buy a car. You can't buy a Tesla in any of these dealerships. So they make it really hard for those alternatives, even though they're superior, just like with VoIP - just like that. They make it hard for you to get to market to buy it. So you have to find other channels. And that's, again, the markets responding, there's lots of ways to buy a Tesla, you just can't do it in the conventional way. Because the dealership model is a closed model.
Yeah. But back to our original point about where voices going. The newer generations aren't going to care. They don't care that you have to go to the local Chevy dealer like they're just want a car.
Yeah, exactly.
So with that, perhaps we should, we don't want to teach the course in advance. But it's very interesting to preview what we've been thinking.
Yes, yes. Okay. So in case you're wondering, and this will be our kind of exit, if you want to find out more about attending our course. Certainly, you can find it on my website and my blog. And also you can go to pulver.com, right, where you can get information about not just this course, but the broader PulveREDU program. So we are one of you know, several pieces of programming that they're offering in this learning model. And of course, Jeff has a host of other ventures and initiatives that keep him going and keep a lot of people like us very interested in what's going on in the world. So Jeff's quite the guy. That's another topic for another time. But yes, easy to find information on pulver.com about it.
And if you can join us, that would be great. And if not, we will just find our way out of today's podcast. So that brings us to time. And we want to thank you for listening today. We hope you enjoyed our podcast and that you'll continue with us as we explore the future of work here on Watch This Space. You can access all of our episodes at WWW dot watch this space dot tech, or wherever you subscribe to your podcast - Stitcher, Amazon, Google. iTunes, it's all there. And also we'd love it if you leave us a rating. And you can also pick up transcripts of our of our spoken word here on my website, if that's of interest, and that is all I have to say. So I'm Jon Arnold.
And I'm Chris Fine. Thank you, everyone. Thanks, Jon.
Okay, you're welcome. And thanks, Chris, and we'll catch you next month.
