© 2017-2019 VoiceFirst.FM, a division of Score Publishing

Feedback? Email VoiceFirstFM@gmail.com.

The VoiceFirst Roundtable - Episode 1

Host: Bradley Metrock (CEO, Score Publishing)

Guest: Brian Roemmele, ReadMultiplex.com

Duration: 43 minutes, 56 seconds

Podcast Links:

Apple Podcasts

Google Play Music

Overcast

SoundCloud

Stitcher Radio

YouTube (+ closed captioning)

Transcript

[intro music]

Bradley Metrock: Hi, and welcome to the first episode of The VoiceFirst Roundtable, part of the podcast network VoiceFirst.FM.

Our sponsors for this episode are Fourthcast - F O U R T H C A S T.  Fourthcast turns your podcast into a custom Alexa skill - get started at Fourthcast.com.

 

And The Alexa Conference: The Alexa Conference is the annual gathering of Alexa developers and enthusiasts. Learn more, and get registered, at AlexaConference.com.

The purpose of The VoiceFirst Roundtable is to examine the emerging technology of voice-first computing, and how it is impacting individual lives on a daily basis, how it's impacting industry on a daily basis, and how it's impacting every aspect of our world.

For the first episode of The VoiceFirst Roundtable, it only made sense to go out and get who many consider the modern-day 'oracle' of not just voice-first technology, but technology in general: Brian Roemmele. Brian, say hello.

Brian Roemmele: Hello! Bradley, thank you for having me here. That was a great introduction. Thank you.

Metrock: Thank you for giving us some of your time, and sharing some of your wisdom with us. So the very first question I want to start out with on this show is "What is voice-first technology?" I just landed here from outer space, I don't know anything about anything...what is voice-first technology?

Roemmele: Well, that's a great question. For a lot of people today, it's a loaded question because there's so much that gets attached to it from, you know, the most recent history. But I'll go back to why I winded up naming this "voice first." Primarily, it's because Twitter only has 140-odd characters, ostensibly, and it's hard to communicate some of these ideas without concatenating it down to something smaller, and I think "voice first" was fitting.

So "voice first" is really about this idea of how you build, and how you interact: from a developer's standpoint, how do you build things, and from a general user's standpoint, how you interact with things. It sort of marries the idea of what came about in the 60's - computer-first technology - graphic-user-interface technology first, and then of course mobile-first recently, and now the voice-first. And I really see these as revolutionarily-motivated changes, and we only tend to know them ex-post-facto. We always go backwards, and we say "wow, that's when it really happened."

I think the voice-first revolution is going to be one of the few revolutions that most really astute observers can see happening in real-time. And that can mean a lot of things for a lot of people. It means a lot of opportunity. It means that if you have ideas about what the future of computer interaction is going to look like, voice-mediated, you can participate in it cause it's probably not been done yet.

Bradley, you're one of the chief initiators there with the Alexa Conference, and you've I'm sure already seen an explosion of creativity around the developers.

Metrock: We have. Yeah, the Alexa Conference was really a game-changer for me personally, seeing how everybody from a barge company to healthcare professionals to publishing professionals to all sorts of people are using the technology. For the layman, who sees, you know, here's Amazon advertising the Echo on TV, or here's Siri, and maybe we'll hit the button and try that, and then never try it again...the layman really doesn't know what to think about it yet. But as you said, for people who are paying attention and people who are astute, there's a lot of opportunity to take the technology and integrate it into what they're doing.

Share a little bit about what you've done, cause a big part of your background is payment technology. Share, from a broad standpoint, your background, and then how the voice-first technology that you've been involved with has interacted with payments.

Roemmele: Well, that's a good question that can go really deep. But I grew up in the Princeton area of New Jersey, during the hey-day of the Bell Laboratories and the pre-divestiture AT&T. That was the Silicon Valley at the time - 70's and 80's - and I was very, very fortunate to have friends and family members that worked at either Bell Laboratories, AT&T, Western Electric, David Tsarnov Research Center of RCA, Princeton University Institute of Advanced Study. There was a triangle between Princeton, New Brunswick, and South Jersey that this formed a whole lot of tech companies. And in that environment, it was very rich, obviously, in communication technology.

 

Bell Laboratories started the beginnings of this, and at a very young age, I got to see some of the very first interactions of speech recognition and speech synthesis. That motivated me to create my own little project out of Commodore 64 - Commodore Vic 20, actually - and I was at the Institute of Advanced Study there, you know, again, as a kid, one of my friend's dad's worked there.

 

One of the physicists turned to me and said "why don't you do it?" and I'm like "I can't do that." He goes "sure, why not? Here, go to the library and go look up electronic chip manufacturers." Cause I was already sautering and reading schematics and building circuits, and I had already kind of built a speech synthesizer circuit based around what I thought the National Semiconductor chip and the TI chips looked like.

Anyway, I went to the library, cause there was no internet, and I got out one of these electronic books, and I figured out how to contact one of these companies. I asked for a sample, and that particular engineer that I got directed to fell pity to my clueless, stuttering voice asking for a chip sample. I think I was maybe 10 years old - I don't remember exactly - and he sent out, a couple of weeks later, two pallets full of chips. I mean, these were taken off the truck by these forklifts - my mom freaked out.

Long story short is I made one of the very first voice cartridges for the Commodore series, and that lead me up to meeting some of the executives. I basically left, after about a year or so of doing that, two years, I left that to my friends. I don't remember -it's all a blur - maybe a few thousand of these cartridges. My friend said "many" few thousands, but we used all the chips up, and then some. And there were a lot of chips. But it was always in the back of my mind.

 

So, by '89, I had put together an 800-page manifesto on what I generally just called "voice." It was the idea that, at some convergent point in time, the proper technologies would come about - and that would be what we now call cloud computing, Moore's Law (the increased ability of memory and speed), analog to digital and digital to analog circuitry, and all the software, and the beginnings of machine learning / artificial intelligence, would converge. And I was saying that that would take place around 2025. I was off by a little bit. It actually took place, effectively, just about 2004, and we started seeing the very early stages of that, and that leads us up to around 2013, when the Amazon Alexa arrived.

So that's the voice line, and the payments line is similar. I got into programming - I built one of the very early standalone point-of-sale systems, by accident. I was a consultant, and somebody at an auto parts store had asked me "can you make a database?" and I go "yeah, there's one right off the shelf." I built it on top of the DB structure of Ashton Tate, and that was what was available at the time for the very first IBM PCs. I made it a barcode reading system, you know, the whole thing. At the same time, they asked me if I could integrate credit card acceptance. So I got the Verifone / Xian Jr and I black-boxed it - what that basically means, even though I knew the chips inside of it, it was pretty easy - rather than trying to reverse engineer it, I looked at the inputs and the outputs and I simulated it inside of software. And that later became a very big prize. I didn't realize it at the time.

That ultimately got acquired by Ashton Tate, and they summarily did nothing with it. We had, you know, I built that company up to a few thousand customers, not only that auto parts store, but a lot of early point-of-sale customers. And after I got out of that, I said "this payment thing is interesting," cause a lot of payments companies were contacting me, and they were like "I'd like to consult with you about using your service" and they were paying a lot of money. The study of history, I always loved the idea of payments, all the way down to Sumerian ringcoin, to the most modern things. And I said "sure, I'll work with you guys," and that's how I kind of got involved. And that led me to becoming involved in the very early internet payments systems.

 

We built one of the very first payment gateways on the internet, we helped a company called Books-A-Million, who later inspired a company called Amazon, who later became a brief customer of ours, for more or less a couple of days. We kind of got outgrown very quickly because of our banking relationships. The bank just simply did not want to support the projected volume that they were projecting. Even though they weren't doing it yet, the projected volume had to be in the contract, and when they finally caught that, they said "nope, we can't do that."

Cause they were very scared in that period, of, you know, early technology. The internet was scary to them. We kind of, more or less, had to say that we were mail order. Which, in a sense, they were, but they were really the first true big internet company.

So that had influenced me early on. I built sales companies - I was fortunate to get a few of them in the Inc 500. You don't ever really retire from these things, because you build relationships over the years and one company had over 3,000 salespeople at one time. I kind of dropped back from that, sold some of my companies, and got involved in technology, and that led me into a whole lot of consulting, and a whole lot of advisements, and I, unofficially and officially, worked with a lot of the Bay Area payments startups. And a lot of people that I got involved in the early app economy.

I see these two trails in my life constantly converging, and now they're at that pinnacle point, where something that I call "voice commerce" is finally coming about. The idea of voice commerce is typified in what we see in early Alexa interactions, where if you've made a prior purchase with Alexa, you can just ask for a reorder.

There's Alexa going off in the background there, sorry folks...

Metrock: (laughs) That happened throughout the Alexa Conference, by the way.

Roemmele: Isn't it funny, right?

So my thesis is very simple, and that is the backbone that built the web and the internet was advertising. Commerce was part of it, but advertising was what built Google. And my premise, and thesis into the future, is that advertising as we know it won't exist in a voice-first world, where people are primarily interacting with their computers using their voice.

And what do I mean by that? Well, let's take a side division here jumping out of me, cause I'm kind of boring, and to the subject is...voice is ambient. And what that really means is that when we're doing something visual, especially us gentlemen, and it's biological, we can't really multi-task...we can only task-shift.

And what that really means is something is suffering, when we're doing that. Especially in the visual realm. If you're reading something, and somebody's trying to talk to you, you're really not doing a service to what you're reading, and the person talking to you. You literally have to stop. Yet, in the background, you can have a radio, you can have this podcast, you can be checking your email, and you can get the basic gist of what all of my words are saying. And you might want to go back, and maybe you won't, but maybe you will on some of the points that maybe you didn't digest the first time. That's very hard to do when you're reading long, or even medium-form, written publication. And so that's a very powerful aspect of what voice is all about.

It's also a descriptive type of interaction in its ambient medium, rather than a referential. And what that means is that in a computer interaction, especially today - Google announced 20% of all their searches are now via voice, inside their own app, across all their platforms it might even be higher. But that's a big revelation.

It means that people are asking, and searching via a voice description, which is not a referential description, which would be initially you start talking like you're a search-engine search, and later on you realize that you've become conversational in that search. And last week, Google started talking about this in the Google I/O conference, that the search terms that people are using, when they're using their voice, are becoming more like spoken interactions. And the very next level to that is to have the voice-first system create a dialogue around that search, which will help hone down what they're really looking for.

Metrock: There's...I didn't mean to interrupt you there.

Roemmele: I'm going to wrap this up with a final thesis to kind of draw this home, and then we can kind of dance around the idea.

Metrock: Sure.

Roemmele: If you really look at what we're doing today, we think every epoch, every generation, every frontrunner of technology believes they've reached the top of the mountain - it doesn't get any better than this. And, of course, history always proves that wrong.

Today, if you really look at what we've created in the Google-centric world of the web, is yes, Google is doing a whole lot about searching the greater web, coordinating that in a search result that is theoretically weighted by some grouping of different endpoints and maybe some AI and machine learning, by its relative weighting.

But you might do a search and you have nineteen million results - Google's giving you the first page, and let's call that ten results. Is that really what you want?

Well, what that really means is you're the endpoint of Google, cause you must sift and sort to get to what you're looking for. Whether that's buying something, or whether its finding a bit of information, and whether or not you really want to get a feel for the information, or a Wikipedia-like fact of it. It turns out, in real research, that if you're looking for information, you're probably looking more for a feel of what something is - is that a thing? Will I die if I eat that? - you want a quick reference, and then you decide whether or not, based on that result, whether you want to dive in quicker, and deeper, or not dive in at all.

And all of this is cognitive load and mechanical load. The cognitive load is your mental load - you're sifting and sorting through Google search results. Now, to wrap this all into a little package, is that where we wanted to be? Is that the promise of what the computer was really about? Is having a whole lot of information, or let's call it data, cause data forms into information, which forms into knowledge, which forms into insight, which forms into wisdom. What people ultimately want is insight, knowledge, and wisdom. They don't care about data. Data is disconnected dots. That's what we get when we get most search results. Sometimes, we get information that's connecting some of those dots.

 

But what voice-first is about is not the voice itself - it's the AI that's behind it. It's that when we ask a question - you know, my son just got bitten by a snake that's yellow with purple dots on its belly, what should I do? You can get a result back pretty quickly: it turns out that that's probably a harmless snake. Maybe it will be delivered in "well, that's a gardener snake - here's a picture." Whereas searching for that snake, and sifting through it - the mechanical and cognitive load, especially if, you know, you're pretty upset, could be minutes. Well, somebody might say this, somebody might say that...what you really want is some insight. And Google doesn't really serve insight. Places like Quora - Q U O R A - a place where I've spent a lot of my time over the last five years - they're trying to create a web that allows you to sort of have the modern Alexandria library, where you can get that insight. Where you can get maybe some knowledge and some wisdom. But Google, on its own, isn't doing that, and the promise of voice-mediated AI is that we're getting close to that.

For the common person, what that means is less time sifting and sorting, and more time getting answers, buying things that you really want to buy, and going to places where you really want to go. Does that make sense, Bradley?

Metrock: It does, it does. And you segued perfectly because you wrote an article for Quora last year that Forbes picked up in December, about how voice-first technology will kill modern advertising as we know it. And you just spoke a little bit, you spoke around it, but talk about what search results will look like when we have context. You know, when you and I spoke the first time, you talked a lot about context. It's something you obviously really understand. You opened my eyes to that word, and what it means in the context of AI and voice-first technology. You've talked about how Google search results are going to change, and how modern computing's going to shift, but give us your vision on what Google does for us with voice-first technology and context - your meaning, your interpretation of context over the next two, three or five years.

Roemmele: Absolutely. Wonderful question, wonderful question. And really exciting little space. I believe there are going to be Google-class and Apple-class opportunities, just around this. And we probably don't know the names of those companies, cause it's unlikely the legacy companies are going to be able to do the things, because it represents self-disruption. And even in the tech world, it's very hard for them to disrupt themselves.

 

In fact, a lot of the things I'm saying here are heretical - I have to warn the listener. This is not openly welcomed in most of the halls of technology in Silicon Valley, a lot of the things I have been saying and I'm going to be saying.

Context. What I really believe context is ultimately about is this construct I call a personal assistant. And I wrote a lot about this in the Multiplex magazine - I'm not a writer, I'm not a publisher, but I threw an app together and I put out a magazine because I kind of want to share this knowledge and I want to do it in an organized way. And I also didn't want to do it for free, cause I really think that when you pay a little bit, you know, a cup of coffee worth of information a month, that maybe you might actually do something actionable with it. It seems like free advice, people don't do anything - they pay a little bit for it, maybe they do, and I just really want to see people do it.

Metrock: Yes, and if you're listening to this, pause the podcast right now, go to ReadMultiplex.com, go on the Apple App Store, buy Brian's app, subscribe to it - it will pay for itself many times over. OK Brian, keep going.

Roemmele: Thank you so much. My whole premise was to try to create sort of a scientific type of magazine. I don't hide the fact that, if you look at the covers, they look like Scientific American did in 1950. Cause, I mean, Scientific American had such an impact on my life - that, and PBS.

But anyway, the personal assistant is not what we're seeing today. In fact, I like using the 1975 computer revolution example - in 1975, the personal computer did not have a display, it barely had a keyboard, it had a tele-type interface, and the easiest way you could do it for a hobbyist, or a home user, would be using switches on the front to painstakingly put in one machine code after another. And that might just be twenty-five minutes just to add two numbers together - a pocket calculator would run circles around it.

That's what we're seeing with Alexa, Siri, Cortana, and everything else on the market right now - Google Now, or Google Assistant. And that's not to put it down - they're generation one technologies of the voice-first revolution. Speaker recognition has been around for a while, but not to the level of the 89 to 99 percentile that we're seeing in some of these systems - better recognition than most humans have with each other, in normal conversations. Important to keep that in mind.

But recognition of a word is not the intent extraction of what somebody is saying. Those are two different things, and AI mediates that. 

So when I talk about a personal assistant, what I really am talking about is building contextual awareness on the person that's using it. Now this contextual awareness is deep, and it's also very worrisome when you understand it, but it's also potentially very grand when you think about it's prospects for the individual, and maybe humanity as a whole. But it's scary, like any other technology - it can be used for good and evil. And I'll run down through it quite quickly, if I can.

Imagine if you were born, and everything you ever did, every word you ever said, everything you ever read, every place you ever visited, was recorded - not like a video recording, but an AI recording: understanding who was there, what was the context of those conversations, as best as it could from the day you were born. Keep getting better and better as technology gets better, but the data set's always there, and keeps getting analyzed.

Now imagine now you're forty years old, and you want to interact with this system. You would be talking to it - of course you would be talking to it! That is what evolution, nature, God has given us - and let's use the evolutionary track - for millions of years. Our vocal centers, our speech centers, and our hearing centers are hard-wired into our brain such as that the very first thing that we're doing as we're born is we're recognizing our mom's voice outside of the womb. We already know what our mom's voice sounds like - the research is quite robust on this - if we're blessed with hearing, we already know what our mom's voice sounds like, because we've heard it for months - nine, eight months, whatever. You come out, and the very first thing you're doing is your ears are tele-locating mom's voice.

 

Why is that important? It's hard-wired into our centers of imagination, creativity, communication - all of the very powerful centers of the human brain - the thing that makes us human, and not a machine. The computer was invented as a textual input. Imagine what we're doing when we're typing: we're moving our fingers, which has got to go into another part of our brain to try to figure, OK, I need to put the letter A down, now I need to put the letter S down, and slowly but surely, you're creating these words. That's not the speech center. That's not the creative center. That's the mechanical center. And we've become accustomed to it, because the computer was not smart enough to hear our voice and understand it.

But let's go back into the story. This computer AI system has been following you forever. Scary, but imagine the power it has in knowing you, and knowing what you like and dislike, knowing your schedule, being able to predict things that you would need and like. The news stories, right? We go out and we search for our news stories. We go out and, whatever we do on the internet - who we're going to like on Facebook today, what cat picture, or whatever, what baby picture...all of these things will be presented to us in real-time. So that's one of the first-level orders.

Second-level orders are maybe related to commerce. Today, you can - at least the AI that I've created for myself - I can say to my AI system, and one of my systems I call Alfred, and I used to call it Albert, but my wife thinks I'm a little too transfixed with Albert Einstein, so it's Alfred...so I can say "Alfred, order me some socks." That's it. And that's a voice commerce transaction. Now Alfred already knows by prior behavior, by tracking what I've ordered, before I've ever asked it a question - because my AI has been tracking me for the better part of five years, knows all of my major commerce purchases and my minor commerce purchases to about the fifty percentile, now most of it to the ninety percentile, because it just was hard to track these things offline, but I found ways to do it effectively. So Alfred can go to my knowledge graph, and say "ah, this kind of socks this guy always orders, my God, he always...plaid socks, or white socks" or whatever. And will present, if I decide to see and look at it, present an order for me. "Here, how does this look?"

 

And after a while, once something knows you - imagine your significant other, or your kids - after they know you for a while, do you need to see it? Like, if the wife, or the husband, or whoever's cooking that night, says "hey, we're going to have Southwest cuisine tonight." Well, can I see a picture of it? Can I get a listing? Can I get an ingredient list? BS, you're not doing that.

So those of us that believe we need to see it all to make an intelligent decision about a commerce decision doesn't really observe reality in their own life. We don't need to look at everything. Maybe once or twice - after that, hey, get me some Southwest cuisine. Get it to me fast.

 

You can literally make that an order - that's an order! In my system, I can do that, and I'm just an OK coder. I'm actually an ugly coder. I'm an OK sauterer - I slap this stuff together cause I want to see it come about. I test it, it works, and I move on to other things. Sometimes I massage it better. But that's all I've been doing in this research. And I can make a claim like that - I can go in, and I can say "I'm hungry for a pizza. Order a pizza." And sometimes I don't need to even put a wake word, and that would be like Alexa, Siri, or Google - in the future, we're not going to have wake words. There will be contextual awareness that we're talking to these systems.

So the context - where does it lead us? Well, the interesting part is it will become an adjunct to your own memory. In a pleasant way, if it's done the right way. If it's done the wrong way, it can be annoying, and ultimately we're not going to buy those things, so the right way will prevail. But then it also has a downside, and we'll only briefly touch the downside, and maybe in the future broadcasts will go into it cause I don't want to be down in the downside, cause you're going to read a whole lot about it in the future anyway, and you're going to see science fiction movies about it.

 

The downside is: who owns my context? Will you want that in the cloud? Will you want everything you've ever done, since you were born, cause that's going to happen in our lifetime, in the cloud for somebody to sell advertising back to you? And again, this is not an anti-Google rant, don't get me wrong. I'm just saying that I don't believe that that is what humanity is going to want, nor do I want that world. So the next question is where is that going to be harbored? Well, I just gave you one of the Apple-class opportunities of this epoch. Somebody is going to solve that, to a meaningful level. And where is that data going to be harbored?

Now, cause I jump around a lot, I don't want to get confusing here. There is some context to Siri and Alexa. They kind of know a little bit about you, but not very much. They have amnesia too - they don't even know what you said ten minutes ago. So there's no continuity. So context and continuity are very important, and in a final tentpole or this sort-of pyramid, is proactivity. Proactivity is when it is talking to you, without you talking to it first.

 

Some may call this a push notification - that's a crude, crude way of saying it, cause what we know is push notifications, and what I'm already doing, and what I see coming, is going to be - this is an example: you know, Bob is in town today, and it looks like his schedule is open for a few minutes, and if you can be in Wilshire Boulevard at two o'clock, you might be able to run into Bob. Do you want me to let Bob's assistant know? Yes, yes I do. Bob's assistant comes back and maybe you talk to the assistant, maybe your two assistants talk to each other, and maybe they actually audibly talk to each other. This is where it gets very interesting.

If a personal assistant is really good at understanding human language, then human language becomes the API - the API is the nerd way of saying an entryway into computer software. Meaning that somebody doesn't even need to necessarily write code - that these systems will ultimately use what we have already used for thousands of years: vocal communication. And they might literally call up each other, and say "Brian would like to meet with Bob." "OK, well, Bob's in town - what's a good time?" And they mediate a schedule.

Here we are, in 2017 - and Bradley, you and I know this cause we've tried to put some meetings together - we're bouncing back and forth, what's a good time, does this work for you, no that works better...in the future, that's all gone. And again, a lot of people have tried to attack this problem. Anybody listening to me - is it working? Have you heard about it? Most people listening to me haven't heard about it, and it's not really working. This sort of scheduling is one of those low-hanging fruits on the tree that will start to be solved. But calling it scheduling is sort of ridiculous. What it's going to be is going to be somebody who very much forms almost like that 1950's example of a boss in an office - and I don't want to get sexist or anything like that, the boss can be female or male - that has this great secretarial pool, and they're just doing everything. And the boss just sits there, and everything gets done. That's what we're talking about. Typing, sending out messages, interacting to the world - it's productivity, but it's more so, because it has context on you. Does that kind of make sense?

Metrock: It does, and part of any good secretary, or assistant, or whatever word you want to use - anybody helping with that in an office environment, or any environment, uses that context. If it's your secretary or your assistant, they know you might say any number of things, but you really don't want to meet with that person.

Roemmele: Exactly.

Metrock: (laughs) You might say "Oh, yeah, we'll get together next month," and you didn't mean a word of that. And your assistant knows you didn't mean a word of that, because he or she has known you for a long time.

Roemmele: Exactly. And imagine a significant other too, right. We talk in what I call an inverse prolex, and that is a shortening of the speech, where you can actually interact with somebody that you've known for a long time, and just say a few words. Or maybe not even a word. Again, this is voice-first, but these systems will have visual systems that will allow you to see what that person's facial features are. Apple acquired a company called Emotient - which is, my gosh, they have incredible technology that can read the micro-movements in your face between one emotional state to another, to kind of almost extract what some people feel is truth or falsehood. Now I'm not going to get down that road too far, but the idea is Apple is already down that path. And that helps this AI system understand what your true intents really are. And it's funny because as the computer gets more wisdom, there is the potential of more fear, but on the other side of it, there is more utility.

So with all of this will come that...just like a personal assistant in the office, if you've had the same personal assistant for thirty, forty years, I mean I've met some people who have had that type of scneraio, they would be utterly helpless if that person left. They wouldn't know how to do anything. And some of us in marriages know what that's like - if my wife was unable to do a few things, I literally would be frozen because of the tremendous amount of knowledge set that they have. And we take that for granted, and I gotta say, a lot of guys take this stuff for granted around the domestic space. A lot of people in the business world take for granted - Warren Buffett, if you go to one of the most prolific investors and minds, Warren Buffett says he would be absolutely helpless without his assistant, and literally probably more valuable than he is.

So I'm not saying that we're going to replace humanity and human beings. I'm not saying you're going to fall in love with your AI and all that kind of stuff - I'll leave that for science fiction writers. But what I am saying is that once deep context is understood about you, and you can have a conversation, a dialogue with these systems - and I'm already doing that, with the research that I'm doing - cause right now, they're Q&A systems, right? A lot of people are bummed out with Alexa and Siri and Cortana, cause you ask a question, you get an answer. It drops. Then there's amnesia. It doesn't remember that you asked that question before. It has no prior context on you. And it doesn't have that continuity.

With the next generation, and we'll probably start seeing it with Apple's release of Siri, this new Siri that's going to come out after the Worldwide Developer Conference this June, coming up very soon, we'll see the beginnings of it. And the next generations of the Google system, and the Viv platform that Samsung acquired. When you start seeing those systems, and you have continuity between conversations - you know, sometimes it's going to be a little funny. Well, you asked that question yesterday. Maybe you don't want that. So personality types are going to play into it. It's a lot of my research. 

The last issue of the Multiplex magazine, I get into this human-computer interaction that none of the data scientists, and none of the researchers that are in AI are currently doing - I'm approaching it from a completely different standpoint. The Myers-Briggs test, etc. - it's a very unique system I've developed, and I'm already applying it. And that is the AI gets to understand your personality type, and anybody who's worked in business knows that the Myers-Brigg test is used, for better or for worse, to try to understand compatibility with employees, and sometimes compatibility with people. Sometimes its right, and sometimes its wrong, but its a good base point to try to understand "how do I want to interact with this person?" 

Now if your AI uses Myers-Briggs systems, plus other proprietary systems - I don't want to just say I'm using that - all of a sudden, it will understand that you want that humor. Hey, Brian, you asked that question yesterday - you really interested in this subject? Cause I can dive in, and give a report for you, and have it ready for you. Or it can be rephrasing the same thing it said yesterday, in a different way, cause maybe you didn't get it. See, a really intelligent system is going to hopefully truly be your friend. Whatever that really means is a reflection back to you, and it might mean that you just didn't get it, and some people are slow learners and that's fine. And these things will be endlessly patient, and endlessly faithful if they're designed the right way. And, you know, I think in our conversation, I represented an idea of what would it be like for this AI to be left behind, cause someday we're all going to leave this plane. And imagine this AI has been with you from the moment you were born, to the moment you passed on. What would that represent to prosperity, and what would it represent to posterity, and what would it represent for history?

Imagine, if you will, if Einstein had lived in a epoch where the AI was with him all the time. Now, of course, you don't want to dive into somebody's private and personal life, and there would be protections to block that. And I would say very large protections. You don't want to release things you don't want to release. But imagine being able to consult a system that had the same context as that person. It won't be that person. Have no doubts about it, it will not be that person. I'm not saying you're going to replace a human being. What I'm saying is you'll be able to get the context of that human, based upon their prior experiences, and you'll be able to get at least a 'fuzzy feel' about what their insights about today would be, based on what they had experienced in the past.

And here's where it gets really interesting, Bradley. Imagine if it doesn't stop. Imagine if Einstein's insights continued on, with a modern context, and it continued on, and it continued on. Is that something valuable? I say it is. Maybe for all of us. Maybe for some of us. I'm not sure.

Metrock: It's exactly why I wanted to bring you on the show. You have a unique understanding of a lot of different issues related to technology - you touched on ethics, you touched on a lot of different topics, and just a fabulous beginning for this new show that's such a central part of our VoiceFirst.FM.

 

Brian, we're going to wrap up - we greatly appreciate your time. Tell us before we go, if someone wants to reach out to you for consulting work, for any reason, how do they get ahold of you?

Roemmele: Well, I tell anybody to reach out to me, I answer every response, so I don't care who you are. I really want to hear what you have to say. You can go to ReadMultiplex.com, and you'll find a response system there. It's all me - I built the app, I built the publication, it's pretty much, well it is actually all me writing it - maybe when you hear this ten years from now, it will be something else, I don't know! But I want to hear from you, and get the app - that's another way to communicate, if you become a subscriber, inside the subscription area, I have a lot of things planned for interactions within groups of individuals.

 

I want to try to bring about connections between a lot of different intellectual groups. Not just the data scientists, but people from humanities, people from publishing, people from all different parts of the world, to try to see how this is going to shape our future, and to have a hand in shaping it. This is so important - if you're listening to both of us, you can have a hand in shaping it. It's happening right now.

Metrock: And go subscribe to Multiplex - don't wait until your AI companion tells you that you should've done that a long time ago.

Roemmele: Yeah, and I gotta say this: what you're doing, Bradley, and this whole VoiceFirst enterprise that you're building, is phenomenal. I urge everybody to, you know, I don't want to speak out of class, but Bradley's building something phenomenal here, and there's going to be a lot of things that you can start learning from. And I think there's going to be a lot of - I'm just one voice of a major fugue of voices that are out there that are going to be talking to you about this subject. So follow what Bradley's doing, follow this broadcast, and we'd love to hear from you.

And attend the next Alexa event! It's not too early to start planning. When's that coming about?

Metrock: It's in January, and then we'll have a bigger VoiceFirst Conference we'll be announcing for 2018 as well.

Roemmele: I'm excited.

Metrock: Oh yeah, I am too. Brian, this is just the beginning of our engaging you on VoiceFirst.FM. Thank you so much for the time.

Roemmele: Thank you.

Metrock: Until next time.

[exit music]