Top news stories for Episode 5 (August 3, 2017):
2) Facebook Staffing Up: Company Actively Hiring Apple Veterans To Produce "Siri-style" voice assistant for their hardware
3) #VoiceFirst "Fake News": That Other Facebook Story Where They Shut Down Chatbots Talking To Themselves In Own Language
4) Amazon Takes Google Home / Chromecast TV integration and Brings It To Their Own Ecosystem
5) The Future Of Difficult Financial Conversations Belongs To Voice Assistants
6) When Voice Meets Impulse Buy: Guest Post on Voicebot.AI By Tobias Goebel Explores "From Thought To Transaction"
7) #VoiceFirst Publishing: Fashion Fix Daily Becomes First Fashion Magazine To Publish On Amazon Echo
Panel for Episode 5 (August 3, 2017):
Patrick leads the VaynerSmart division at VaynerMedia, a full-service digital marketing agency headquartered in New York. His group brings Vayner's mission of cross-platform digital storytelling to emerging platforms, with voice-first being one of their primary areas of focus.
Tobias Goebel is Director of Emerging Technologies at Aspect and has worked with voice technologies for nearly twenty years. He is a frequent speaker and blogger on topics around customer service and, more recently, the (re-)emerging chatbot, NLP, and related AI technologies. He now leads the global chatbot program at Aspect. Have a look at www.aspect.ai and check out Aspect’s NLU Lab, or say hi to Ivy, Aspect’s digital employee, at m.me/AspectSoftware.
Brian just published issue number 6 of Multiplex Magazine called The Enchanted Loom. He explores a new AI concept for Voice First systems called Artificial Understanding. Get the Read Multiplex App at the iOS store and subscribe for this and the entire catalog of magazines.
Bradley Metrock: [00:00:09] Hi. And welcome back to This Week In Voice for Episode 5, Thursday, August 3rd, 2017.
Bradley Metrock: [00:00:17] This episode is sponsored by VoiceXP. This is our first episode of This Week In Voice where we have VoiceXP signed on as our exclusive sponsor. Please take the time to go check them out for us at VoiceXP.com and learn how they can create an Alexa skill for you or your organization.
Bradley Metrock: [00:00:36] We're very very very pleased very grateful to have joining us today. Three amazing guests. Patrick I will start with you. We've got Patrick Givens from VaynerSmart - Patrick, say hello.
Patrick Givens: [00:00:48] Hey Bradley and everyone out there - glad to be on today.
Bradley Metrock: [00:00:52] Patrick thank you very much for setting the time aside Patrick leads the VaynerSmart division at VaynerMedia which is a full service digital marketing agency headquartered in New York. His group brings Vayner's mission of cross-platform digital storytelling to emerging platforms with voice-first being one of their primary areas of focus. Patrick, thank you very much for joining us.
Bradley Metrock: [00:01:13] Our next guest is Tobias Goebel. Tobias, thank you very much for joining us.
Tobias Goebel: [00:01:17] Great to be here. Thank you.
Bradley Metrock: [00:01:19] Tobias is Director of Emerging Technologies at Aspect and has worked with voice technology for nearly 20 years. He is a frequent speaker and blogger on topics around customer service and more recently the re-emerging chatbot, NLP, and AI technologies. He now leads the global chatbot program at Aspect. Have a look at www.Aspect.AI and check out Aspect's NLU lab, or say hi to Ivy, Aspect's digital employee at the link provided on ThisWeekInVoice.com. Tobias, thank you very very much for being part of this with us.
Tobias Goebel: [00:01:55] Thank you.
Tobias Goebel: [00:01:57] And we have Brian Roemmele - Brian say hello.
Brian Roemmele: [00:01:59] Hello everyone.
Bradley Metrock: [00:02:01] I feel like I say this every week, because I do say this every week: if you have not downloaded the Multiplex app, you need to hit the pause button on the podcast. You need to go to the web browser. Type in ReadMultiplex.com. Go and download the app, and then subscribe, and you will be much, much smarter. Brian, thank you for joining us.
Brian Roemmele: [00:02:24] Thank you Bradley. And I got to say I'm extremely excited because Patrick and Tobias are just brilliant and I'm going to be basking in all of their knowledge today so really exciting show here.
Bradley Metrock: [00:02:36] I'm pretty excited too. So with that we will get to the news.
Bradley Metrock: [00:02:39] Our first story this week is from Wired magazine and this one has been getting a lot of attention. A hacker turned an Amazon Echo into a wiretap. And if you read through the story it's actually a pretty interesting account of how a hacker took the Echo apart physically and installed some hardware additions to it. In addition to a little software and basically turned it into a listening device that could not only be activated remotely as that person saw fit but also it had complete control over the surrounding environment. And Brian I want to start with you on this story this week. Privacy is a big issue. We talk about it explicitly every week. If we don't do that it's sort of inherent - part and parcel of every topic. What do you make of this particular story and where do we sit with privacy? Is this story going to turn people off from the Echo and the Amazon ecosystem, or is this sort of not a huge deal?
Brian Roemmele: [00:03:46] Well, Bradley, thank you. You know I can't say this enough: privacy and security are going to be the defining moments of this generation. Whatever took place in the past is going to look mild in what the context that these systems have into the future because for them to be immensely useful the deeper the deeper in the deeper context it has about you. And it would be your personal assistant in what I call its true definition. And that's not what we have today in Amazon's Echo / Alexa platform, or Siri, or Cortana, or Google. They're just Q&A devices. But as they start breaching the deep context of humanity and they start going into the deep detail to create usefulness, fear is going to increase, and security and privacy is going to be bridged. You know getting to the specifics of this hack, if you will...it really is based upon iteration number one of the original Amazon Echo. Below the rubber flap of the original Echo was sort of a 'access point,' if you will. There is an SD card slot and there's other things that you can go in and mess with. And in that particular version you can hack the software the record all of the things that it does, you know, the microphone picks up onto an SD card and you can remove that card and obviously inspect it at some future date. That is worrisome and probably shouldn't have been there for any number of reasons. And there should be ways to disengage that.
Brian Roemmele: [00:05:19] I think in the future the things that we're going to do and the things that I've done in my own voice-first devices that I've created and also what I've done in everybody else's voice-first devices is - and you can go and look at this, all of them allow this - is it you have a tone that turns on whenever the system is actually recording your voice to be sent into the cloud. That notification into some cases is maybe annoying but it creates a feedback loop, maybe a Pavlovian feedback loop, that tells you that this system is now taking whatever you're saying, recording it, and sending it to be inspected by the cloud. And in theory only under those circumstances could that be hijacked. But in reality we learn that a microphone is a microphone, right? But at least that goes some of the way. I believe that we should turn those things on by default. I think whenever you say Alexa, like we hear it here, you just heard that tone if it was close enough. There's a tone that pops up and that's turned off by default and that is not a good idea, I think ultimately. So that's under normal use cases. Under you know hacking use cases, I believe that ultimately we're going to have to have fully encrypted microphones with signature encryption so that if anybody were able to take anything out of that microphone it would be useless. They literally would not have anything from the diaphragm of that microphone backwards meaning that the microphone itself is an encryption mechanism. I believe that that is one of the easiest ways to stop that sort of hacking.
Brian Roemmele: [00:07:03] Now of course somebody can always plant a microphone in a room - that's never going to stop. But to be able to go to the consumer - and I'm releasing something here that I do have an application for a patent in some ways for - there are ways to do it with not through my mechanism but to encrypt the audio coming directly out of the diaphragm of the microphone meaning no wires lead out of that microphone that has any analog audio. To do that, in a epoxied self-contained device, to do that gives you a general sense of security that it is about as secure as a financial transaction which is by the way I believe the standard needs to be when something is listening or watching you. I believe or video is going to have to go down the same road, ultimately, when it's unattended and ambient in this regard. So, yes, this is a big problem. It is only going to get worse before we all put our minds together to make it better. There will be structural advantages to some companies that get this quicker and I don't think we could talk about it enough.
Patrick Givens: [00:08:06] Brian I couldn't agree more with the necessity that the hardware itself, and then the handoff of the voice recordings, be secure. When we look at this at Vayner from a marketing standpoint, we're really starting to look at what is adoption of this space look like, from the users out there? Because ultimately we are trying to find a place to engage an audience. That's really why we would bring brands to operate, or personalities to operate, in this space. So I loved, Brian, your point about the feedback loop. And to me that's a step in the direction of these platforms providing not just great transparency of exactly what is being collected at each moment, and who has access to it, but beyond transparency and the ability for someone to look into these details...actually proactive honesty. Really reaching forward and telling people exactly what's going on at every moment. We are asking people to start bringing a new level of connectivity into their lives their homes today and well outside of their homes in the near future. And what we need is a comfort level with that. I think it's about instilling confidence and that's confidence to, as you began your comment, really to be willing to share not just tertiary or superficial dated information, but share some real meaningful data and personal information with the confidence that that's not going to be misused or lost. Because it's only with that data that we can actually move from the novelty experiences that we're seeing all over the platforms today, into real utility.
Patrick Givens: [00:09:41] That's where the value is from my perspective and from how we coach our clients here at VaynerMedia - it's to look for a way to engage in a conversation so that you can deliver something truly useful. And the way you get there in most cases is to have some real data to back up that experience. If there's not confidence in a platform, if the audience out there either doesn't adopt - which I think we're seeing great adoption so far - but as that adoption happens, how ingrained those get in our lives? If there's apprehension to use the thing, then for us looking at it says the beginnings of a marketing channel it's like shouting into an empty room. If you've got people who either aren't going to engage in a conversation because they are confident that what they say is going to be treated respectfully or might engage at a superficial conversation, but aren't willing to get personal, and get deep, that's going to really make it a challenge for us to deliver anything of value. So there's the primary hardware security and making sure that everything is designed in such a way that it's as secure as possible. Absolutely critical. But then even beyond that I think there's a huge onus on the platforms and on developers alike who are working there to be extremely forthcoming with how they're going to use every bit of data that they are collecting the experience. So there aren't surprise cases that wind up sort of poisoning the water.
Tobias Goebel: [00:11:01] Yeah I think Patrick actually raises an important point, which is in the user experience. And on the front end making customers and users and people aware of, you know, how the thing works and what the Echo does with what would you tell it. And I see this related kind of effect on on Web sites and other places where customers are asked to, in that case, provide sensitive information or private information like your e-mail address or phone number. It confuses me and surprises me why companies aren't doing a better job at explaining why they are asking for this information. What are they going to do with that information? Because oftentimes it's actually for you know for the better of the user. It's oftentimes used and collected you know to serve them better. But they're not explaining that to me so I don't feel comfortable at that point when they're asking me for that information. So I think that's a similar effect here.
Tobias Goebel: [00:12:08] We need to be - we as an industry, and Amazon, everyone building these devices - need to be up-front and you know really open and transparent about how it works and that it's not always on, always listening by default. In the early times, that's been floating around, and still is sometimes today, and it's been miscommunicated sometimes even by experts.
[00:12:36] So we're still very early, even though about 10 million devices apparently are out there in the US which is mind-blowing. As for this particular incident I'm surprised it took so long. Quite honestly, I'm glad that it happened. I am hoping that the manufacturers take good notice and analyze how they did it and improve the systems. It's important for adoption.
Brian Roemmele: [00:13:05] You know Bradley, I've got to say Tobias and Patrick, brilliant, brilliant insights here. I think you need to conceptualize even a little further is that you know the Web was risen on this concept that we can track you, we can follow you, and we can do this surreptitiously. And you know because we are smarter with cookies and all these other things. I think the pendulum swinging the other way. And Gary Vaynerchuk, and VaynerMedia, and a lot of these more sophisticated marketers are starting to realize it's the permission, it's the welcoming of coming into your home, and wanting to share what that brand is about. And so what it really is doing is it's moving things away from sort of a "hack" of saying wow I know that you've looked at 10000 Web sites for dancing shoes or something, to more of I've built such a great equity in my brand that you magnetize towards my brand, and I don't have to throw it at you. You are coming to me in a natural process. And it has to be that way in a sense that I don't think the open tracking that we've seen in the past is going to translate into this new modern augmented world. And this is voice augmentation, visual augmentation...as we start moving in this space, it's going to be much harder. And I'd love to hear what Patrick and Tobias might say about that, with how brands need to become more, really, super brands. Does that make sense.
Tobias Goebel: [00:14:37] Yeah it does make sense. And this reminds me of an initiative that was started I think about 10 years ago by Doc Searls out of Harvard. It's called Project VRM. VRM stands for vendor relationship management, which basically is the idea of flipping around the pyramid of CRM, with the model of CRM, where it's the business tracking you. VRM is about "look, I am a consumer, I have rights, I have terms and conditions, just like you have. I dictate to you, I tell you, dear vendor, what my terms and conditions are, and on which you may communicate to me and serve me." It's interesting because it touches preference management. It really shifts the balance you know completely over to the consumer and that's an ongoing trend. Now the Project VRM community, if you Google it, is still fairly small. I think from what I've observed - I monitor and watch them more closely a couple of years ago and I've stopped doing that cause there's so much other interesting stuff out there to pay attention to - but I think it's related and I think it's a slow shift in the society, in the world, away from businesses towards consumers.
Patrick Givens: [00:15:52] And I think - I love that, I would actually even add to that - it's really interesting when you do look at while legacy systems that are in place, the media buying structure, or even just ad-supported publication networks, may not incentivize such a one-to-one and honest and open relationship. Ultimately as we're marketing on behalf of our brand clients, the most efficient programs we can have are those that someone wants to be a part of. Trying to chase people around the Internet and drop our ad in front of them in just the right moment is really hard, and it's expensive. And we would much rather build an experience that merits people wanting to be part of it. And so that's exactly how we're looking...we're so excited about the various opportunities that AI starts to enable for us across platforms, and so voice is obviously a great place for us to start in that space, with the whole goal being we built something worth coming to. So instead of it being like we've you know it's a tax on the content that you're going to see to work around our ads. We actually want the marketing materials - that's why we talk about storytelling rather than ad production - we want to make something that you care to come find, and engage with a little bit more. So that's absolutely the model that we are trying to shift towards.
Bradley Metrock: [00:17:14] Awesome. All great commentary. The only thing I would add to any of that is just sort of how amazing it is the adoption of voice-first technology right now, given humanity's natural inclinations to resist invasions or potential invasions - anything that could possibly be an invasion of privacy - into our lives. A lot of credit goes to the big tech companies - Amazon and Google and Apple and Microsoft, and others - marketing, getting the message out, and spending the money on marketing, and doing it in ways that lifts the entire effort. Because when you read articles like this in Wired, my first thought was it's almost hard to believe that Tobias's, as you referenced, there's so many devices that are already the installed base of this technology.
Bradley Metrock: [00:18:15] We will move on to story number two. This is an interesting one. Facebook is staffing up. This article talks about how Facebook is actively hiring veterans from Apple to produce a voice assistant that will go along with their hardware. Last week and possibly even the week before we talked about Facebook's Echo Show rip-off, basically - a competitor, if you want to put it that way - and we talked about how viable that may be in the marketplace. And obviously Facebook is going to go hard on this, and this article speaks to that they're looking to put the manpower behind it.
Bradley Metrock: [00:18:51] And Tobias I will start with you. What should the lay person, or just other people in the industry, take away from Facebook's efforts right now to get into the voice-first space?
Tobias Goebel: [00:19:08] Well think about what they say they do, right? They are in the world to connect everyone to everyone. And what better way than to do that on a device that sits in your home when you're on your couch. I mean I've always kind of you know looked at the Echo, when it came out, as a product that closed the loop, that closed the cycle, and closed the gap that we had in terms of technology connecting us to others. You know when you're driving and you know you can't do anything but phone calls when you place a phone call you have to have the phone in that environment. When you are at work in front of a computer you have the web browser you have the you know your desktop your laptop when you're out and about you have your smartphone you can use you know texting and use phone calls as well. When you're at home and you want to connect or you want to get stuff done, you don't have anything. You have to pick up the phone, after a day of staring at screens, and you're kind of tired, laying on the couch. Now you have a way to get stuff done. And we've seen that with how Echo has introduced the the drop-in, and the call feature, and the messaging feature, which I'm using more and more with my folks over in Germany and my brother there. Now we have a way to do that at home. And, you know, what better way to extend a social network on a voice level? I mean think about coming home. You know tired of staring at screens crash on the couch and say "give me a nice three minute rundown." Facebook always has been the newspaper for my friends and family. People you know love to fear - FOMO, fear of missing out - people love to stay on top of things, especially what's going on in their circles. Taking that to voice is a great idea, and I wish that Facebook will pull it off well. What's ironic is that they're turning to Apple to pull it off well, because while Apple is probably the most brilliant company on the planet when it comes to building tangible products, they are not so brilliant when it comes to building non-tangible products.
[00:21:15] You know Siri has been a surprise, quite frankly to me, because they were the first. Usually they're not the first to do anything. Now you could argue both ways: you could say they were you know the first to bring out voice tech, but it was basically a new way of an interface after everyone else has tried interfaces, Apple came and did it right. You can argue that way. Or you can say no, they were the first to bring out a voice interface. But look how poorly they've evolved it over the years. They have not opened up. They've added a handful of new third-party integrations, like OpenTable and others over time. Fandango, I guess, for movie booking. Very poor, very disappointing, and I think it won't change until Apple hires a chief voice user interface design officer. They have a chief design officer - that's Jony Ive - voice user interfaces are a different animal. And I don't see that they're doing it well. So that surprised me that they're hiring from Apple, but it doesn't surprise me at all that Facebook sees the potential of a social network based on voice and hearing. Hearing your friends and family give you updates.
Patrick Givens: [00:22:24] Sure. And Bradley you plugged Brian's Multiplex publication in the beginning of the episode, and I'm going to give it another shout now, because I'll say Brian, your write-ups on how Facebook's voice system will be radically different, and the ways that you anticipate their positioning, have been super-informative to me. It's something we've been reading and passing around the office over here.
Brian Roemmele: [00:22:48] Thank you.
Patrick Givens: [00:22:48] I start, really, Tobias, from very much the same place that you do. What is Facebook's business? Ultimately, isn't it making connections between people and then finding opportunities to monetize as those connections are made? Whether that be advertising placements? Whether it be commerce in those connections? And just like we're talking about here, they see the same shift in where that attention, where those connections are happening more and more voice-first, more and more of these things moving from a screen-based system, or at least screen-first, into something where it's voice-first and certainly visually accommodating, but operating that's based around you. So not a tremendous surprise, I'll say, that we see something coming from them in this space. But very interesting and very exciting to see what differences they start to build into their version of experience are. Tobias, I share some of your skepticism about Apple being the source of necessarily the talent to get them there. But no doubt there is a lot of history of thinking with audio first in a lot of areas within Apple. So certainly something that I could see coming to bear, bringing about some different thinking than maybe where Facebook has operated in the past. Not sure yet from our side exactly what this is going to look like and how the difference in what Facebook ultimately offers will stand out from where Amazon and Google take things, where Cortana comes in the market, etc., but really something we're interested to find out - and coming from the marketing side obviously - Facebook has substantial support there for a lot of the efforts that we put forth. So I'm really interested to see what unique turn they put on the types of voice, what interactions they're encouraging, and then what that winds up opening up for us.
Brian Roemmele: [00:24:41] I've got to echo what Tobias and Patrick are saying. You know this is what I know right now: there are not enough individuals on the planet right now that understand voice to the level that's necessary. In fact I believe the deficit is larger than it was when the web was being brought around. In fact some of the legacy skills that one might have from building web pages don't immediately translate into building really good voice applications. I'm sure Patrick and Tobias would attest. And it creates a completely different mindset. In a lot of ways the creativity that we're pulling out of people, and like Patrick says, the storytelling that we're going to be doing both for brands and so on, just building these applications and these voice apps is really building from a different side of the brain. It's a less mechanical side of the brain. It's more organic; it's understanding of human nature. And a lot of what I write - in fact I'm putting out my July magazine today because I've been busy this month - I've been talking about Maslow and the Maslow hierarchies and how we build to meet around that. And it sounds esoteric to understand that we have about 65000 thoughts a day and about 90 percent of those 65000 thoughts are the same thoughts we had yesterday. And when you can typify human thought and potential human behavior in those thoughts, you've now reduced it down to a substantially lower number of what's really going on in somebody's brain. Now is that mind reading? No. It's to try to understand volition and intent that a person is trying to reach in their life. And Maslow is believing that we're all trying to transcend to a much higher awareness state. When you look at what Facebook is, it is part of that Maslow pyramid. In fact in many ways it's the central part of that pyramid. Certainly isn't the physiological needs that people have, although the addictions that you know I will say Ahmed would normally bring up...
Bradley Metrock: [00:26:55] Yes he would.
Brian Roemmele: [00:26:56] ...the addictions that we have become physiological. That is a good and a bad thing. But I think for all of humanity and even for brands, addiction ultimately is not a good thing, because it really ultimately breaks that person. And by the way they probably will have less money to spend in the future if they're truly addicted to something. So in the long run it's not a good thing. But Maslow is very clear in identifying that we are transcending and ascending to these higher states to be some...let's say become self-actualized. It's all sounds like new age thought until you realize it's exactly what we're doing. Facebook is part of that journey and we're infants in it and we really don't know what we're doing because it is in a way something new. We have the ability to create a persona of ourselves a reflection of maybe what we hope to be self-actualized. You know,look at me on Instagram, and look where I'm vacationing, and all these other things that we do which may not ultimately be where these things wind up.
Brian Roemmele: [00:27:58] Now when you're dealing with Apple, Apple has an incredible brain trust. They have some of the most brilliant people on the planet. I would argue that Apple is a brilliant company doing brilliant things but I think a lot of people would agree with me that they didn't do very well in extending Siri. Siri...the Alexa market was Siri's to lose, and they lost the first round. That technology was far more advanced than anything in the market at the time. We're talking almost seven years ago the last act - I say this a lot - the last dying act of Steve Jobs was to acquire Siri because he saw it as a future of the company, in quotes, period. And he saw it even more profoundly important than anything that came before. I don't think that we could say that that's what has happened over the last few years with Siri. So that does create internal frustration and it's liberating some of the talent. But here's the interesting part. If we take everybody that is available in this industry and line them up it is woefully inadequate for what we're going to need right now, and what we're definitely going to need in the next 10 years. But most definitely the next three years. The next three years are going to be very, a very crazy time where people are going to start trying to adapt the new skills. We're in excess of some, I would say, over 15 million single standalone devices in the United States right now that are voice-first. And that number is going to increase orders of magnitude after HomePod is out because that will have a raising of the harbor - the 'full moon' effect on the harbor, if you will, for all of these devices and all the other ones that are unannounced including Facebook.
Bradley Metrock: [00:29:45] Hopefully that's the role that VoiceFirst.FM can help play, is putting quality information out there so that as people exit one sunsetting type of technology or look to make some sort of other change in their lives that this can be a source of information for people to get caught up and learn about this emerging sector.
Bradley Metrock: [00:30:06] But one thing that we do disagree on: you know it is possible that...you know I can see a future where you know you have different devices and so forth. But with the Echo Show specifically...you're not going to have two of those things sitting on your nightstand. You're just not. You're going to have one. And so who's it going to be? Is it going to be Amazon? Or is it going to be Facebook? Or is it going to be somebody else who comes up with something even better than both of them? So while it may be true - and this is just my opinion, of course - while it may be true that there's different use cases and places to plug in in your house, and this that and the other, and you'll have several different voice assistants doing different things - one in your car, one in your kitchen, one in different places - you'll only have one on your nightstand and you'll only have one in your kitchen and you'll only have one in your car and so there's different battlefields that have to be won. You know what I mean?
Brian Roemmele: [00:31:05] You know, Bradley, I got to say this just to make it more clear, and put a finer point on this. I fully agree with you under that context. Yeah if the battle is for a standalone device with a screen you're going to bias yourself into one direction based upon how you use these systems. But as the technology improves and the screen that you need is virtual and maybe it's a VR...maybe it's a VR hologram...because that's where this is going. The idea of glasses - glasses are a stumbling block. They're not the end point. The end point is it's going to be holographed in front of you. Now it sounds futuristic...I've already seen technology, I've been in some of these labs - I can't say which - where you can display imagery right in front of your face and you can have it is large or as small as you want, and the screen is virtual. So if you want to see a screen, or 10000 thousand screens, they can be arranged along your physical space, and you just turn your head look at them. It's quite fascinating.
Bradley Metrock: [00:32:04] Hey, that'll be cool when it comes out. I want to believe - it's like the X-Files slogan - I want to believe in what Facebook's doing, and our next story ties directly into that, so we're going to go onto the next one. But, yeah, you know it's easy to sit here and hate on what Facebook's doing, but the reality is I hope that they are successful.
Bradley Metrock: [00:32:21] The third news story this week is very interesting to me personally and perhaps to y'all on the panel. We had our first bit of 'fake news' in this show, as well as voice in general I think. There's a big story that we covered about how Facebook had...as the story went, Facebook researchers shut down a conversation, I guess, between two chatbots that had started talking to each other. These two chatbots - and chatbots not exactly the right term - I guess primitive AIs or whatever you want to call it. Certainly some of the articles referred to it as AIs. These two things we're talking to each other - two computers - and they decided, ah you know what, we don't need English. What a garbage language. And so they just made up their own much more elegant solution. And Facebook looked at this and said this has gone far enough. We're shutting down Skynet before it starts up here. And and so there was backlash: I started to see it a little bit on Twitter, and then there's been formal articles written - one of which is on our site - saying look, this is not how this went at all. And Patrick I want to start with you. What did you take away from this, if anything? What are your thoughts on how this shapes our perception of Facebook, and what do you think about this in a larger, more broader context?
Patrick Givens: [00:33:57] Sure. So first off, I was awed by the irony that Facebook pumping up fake news about Facebook, and what sort of internal logic loop that sets in place. Looking at this story - and I'll say it trended on my Facebook feed - as the horrific, apocalyptic future version of the story. And I was a little taken aback. And then reading into it a little more, started to get to the bottom of things.
Patrick Givens: [00:34:25] I think there's maybe less to be taken from this specific case of what happened with the two AIs and a little more the two directions that I get with this are one, the willingness of all of us to hear this story and project our worst sci-fi fantasies forward. And the overall skepticism of AI. I don't want to totally bring us back around to our first story's point, but I actually think a lot of this does come down to transparency and honesty about what we're trying to achieve with these technologies. And some alignment on that, some broader cultural conversation around what is the benefit of AI in this case? Automation in general? And what are the upsides of this? It's very easy to look at the the potential negative downsides of these things, and it's important to consider them, no doubt. There absolutely must be conversation around what negative outcomes we could be affecting with this stuff. But on the positive side, what are the potential benefits? And in this case, specifically, with a program like this, what are the ends that Facebook is trying to get to? Some conversation around that, early in a process - and of course this is an early stage research project, it's not going to be public facing - but what is the broader goal of AI work, and what are some of the positive cases that we can see coming out of it? I think it's to be a much more interesting conversation and can be based in facts. And then it does give you a lens through which to evaluate cases in which potentially things do go wrong, and you need to be cautious. I mean this on the back of Zuckerberg and Musk, their back and forth over the last week or two all playing into each other about do we have this 'trust us, the future will be great' view, or do we have a super-skeptical cautious view of everything that's coming? I'd like to think there's a middle ground between the two, which is optimistic about the future, but a little less of a blind 'trust us, we'll roll something out to you' and a little more co-created, a little more open in discussed.
Patrick Givens: [00:36:27] I think it's really interesting looking at the intents of the big players in this space. So, for Facebook, if we root back to this idea of improving connections and being present and facilitating connections everywhere...what is the role AI can play in that? I can imagine many cases where AI could scale, from my perspective. Get on the marketing side here, where AI can start to scale out the type of a connection a brand might want to have with their customers. Totally valid - lots of beneficial use cases for that. But I can also imagine places where AI comes in, unseats a bunch of people from positions where they were representing that same brand in a different context, and diminishes the experience - isn't actually super-dynamic. Is not really meeting all the needs along the way. So there's plenty to be explored here, but understanding what we're going for, and being a little more transparent about the intents behind things, I think buys everyone involved a little more leeway if there are missteps along the way, rather than priming the whole public and everyone to be really teed up to react negatively to something, which I think is ultimately what you see in this case. You hear a story - maybe it's initially misreported - but everyone is so primed to expect the worst at this point, because of so much secrecy and so little transparency that they're more than willing to run with it. And then when the Snopes does their run on the story and comes back with the detective version that's actually gotten down to the bottom of it, obviously a small fraction of the people that heard the horror story up-front find the second half of it.
Tobias Goebel: [00:38:09] Well I think it was a very human reaction that we witnessed out there. We love horror stories, right? We love to be shocked. And I think it's born out of a lack of understanding how these things work. And what's interesting to note is that you know even the generation of my parents, in their late 60s, know programming and they understand...they don't know programming, what they do know is you know they've heard of 'if then else.' So the 'if then else' programming is kind of pervasive now - everybody has a rough concept of how computers are programmed. Machine learning / deep learning are new ways of programming computers. I mean here you know it's kind of interesting that to see that people use you know very kind of anthropomorphic terminology when they say you know the chatbots decided that English wasn't worthy, and they developed their own language...that's not what happened, obviously. They were just following some programming. And you know, even if you do 'if then else' programming, you can...if you have loops, the moment you have loops, things can happen you didn't expect. That's very common for every programmer out there. So classical case, I think, of lack of understanding imposed on something that's fascinating is that we've been learning about through literature and sci-fi flicks, and it'll probably take a long time - 10-15 years, maybe a generation - until this new style of how computers work has really permeated society, and everybody understands what's really going on.
Brian Roemmele: [00:39:49] Absolutely. You know, we are conditioned by our entertainment and our media. And as far back as we can imagine, even the late 1800s, science fiction always sort of gave a dystopian view of the world. And the opposite of that is a sterilized white world where everybody's wearing white outfits, going down white car doors - you can imagine that picture in your mind. They're always wrong. It will never look that way, because life is not black or white - it is shades of gray. And we are emotional creatures who build our world moving forward with everything that we knew in the past. And we reference that into the future. And most of what we build is in fact everything we build is a simulation of what we've seen in nature - it's the nature of being alive. We look at things - we say, "oh we can communicate better than the way we do now" and then we go around, and then we come back around again, and say "well, that worked OK, but it wasn't that good." And that's kind of where we're going with this.
Brian Roemmele: [00:40:52] But you know there was an interesting study in December by Max Planck Institute and University of California Berkeley. And this is something that is interesting: we build machine learning, yet in that machine learning, we have absolutely no idea when it's really working - how it arrives at an answer. And a lot of people in the computer science world will say "oh, we can figure this out." In reality, you really can't unless you build algorithms that are designed in real time to try to figure out what's going on. And then over time that becomes clogged and then it becomes even immensely more confusing. So the idea of what learning is, whether it be human learning or machine learning, is always a black box. There's inputs and there's outputs, but we may never really truly understand what the middle is. And now I add one more thing: quantum computing. Where we're literally playing with Schrodinger's cat. I mean we're putting in all these inputs - it is looking at every possible world that ever could exist. And I really mean that - that is exactly what's going on in quantum computing. And then it materializes into a reality when it's being observed. This is a known element of quantum physics. That's what quantum computing is going to be about. So this is not going to get any more crystal clear any time soon. It is going to become more strange, more complicated, more potentially creepy, and there's going to be a whole lot of science fiction written about it.
Bradley Metrock: [00:42:29] This Week In Voice is brought to you by VoiceXP, blazing the trail in voice technology. VoiceXP is taking the lead in developing Alexa skills for the best brands in the world. With VoiceXP, all you have to do is say it, to revolutionize your marketing strategy.
Bradley Metrock: [00:42:47] Let me tell you something: we've been doing these VoiceFirst.FM podcasts for the last few months, and it's just been a miraculous ride. Tons and tons and tons of new people listening, talking to us about the podcasts, and sharing this experience with us. And the folks at VoiceXP, who are doing fantastic work, are our first official paid sponsor of This Week In Voice and The VoiceFirst Roundtable, our interview show. And we are very grateful to them, and in turn, we would ask you to go check them out, and take a look and see what they're doing. And if you work for a small, medium, large company that needs an Alexa skill, call them up, look at the website, reach out, and see if they can help you, because they probably can.
Bradley Metrock: [00:43:41] We now turn to the fourth story this week which is that Amazon has taken the Google Home and Chromecast TV integration and brought it to their own ecosystem, which is just a continued part of the ongoing competition in the space. Tobias, what do you take away from Amazon's move with this?
Tobias Goebel: [00:44:02] Well I want to cover a particular aspect about you know controlling TV with your voice first. It's actually quite interesting. I wrote an article back in November 2015 about the missed opportunity that Apple had, in my mind, when they released their new Apple TV. And you know the Echo was just new that year, and it's actually quite funny because the post was picked up by Pulse and then went viral in my humble sphere, it went viral. But what I try to explain is that you know the Echo with, you know, how it was new at the time, was really how the Apple TV should have been. You know TV, being kind of you know at the center of a household, in terms of the physical location oftentimes, where the family gathers. I've gotten some comments from people saying that's kind of sad, but I think it's reality. And you know when you when you compare that to how TV makers have tried to make the usage of TV easier over time, you know when the TV remote came out, we probably had some naysayers there that said "well, I can get up and switch channels just fine. I don't need another device in my household." So that went away - everybody got it - and, you know, cannot do without anymore. Controlling the TV over your voice makes a ton of sense. I had thought - and maybe that was in Steve Jobs in his mind when he kept saying and talking about TV is a next challenge, and he couldn't fulfill that vision that he probably had - but it makes tons of sense. It fits into the user experience, at that point in time. From a UI perspective, it makes sense to leverage that as a methodology to control the TV. Others have done a poor job I think. Apple, you know, again I was somewhat disappointed when I saw what they did with Siri on Apple TV. They gave it the same name as what we know is Siri, but they completely crippled the functionality there, which is something you just don't do in the world of UX and UI because they set expectations around what they did there, with the introduction of Siri and the Apple TV, and they under-delivered. But integrating the competitors' systems - again unifying that into one and the same platform that people have been having in their homes for a while now - I think is a very smart move.
Bradley Metrock: [00:46:47] Moving on to story number five: the future of difficult financial conversations belongs to voice assistants. This is a fascinating look at the USAA and how voice assistants are being brought in to have conversations on spending and saving and different, you know, difficult financial topics with some positive effects. And Brian, I turn to you for this: what do you take away from this article, and do you think this is part of the future?
Brian Roemmele: [00:47:19] Well I absolutely think it's part of the future, and there is a whole lot going on in this type of technology. You know the banking world and the, let's say, #VoiceCommerce world are going to converge, and the idea of being able to actually speak to a banker and to have financial information to have all of these things presented to you in a private space...you know this is where near-field communication really plays a factor, and Apple dominates in the near-field. So I think it's a great beginning. I think we're going to see a lot of banks moving in this space, and I'm really excited to see to see a charge up. And it is amazing to actually see how quickly they have been able to get to this, sort of this place. I had really guessed it would probably take maybe another year to see this type of application. So it's awesome.
Bradley Metrock: [00:48:18] Turning now to story number six - this is written by one of our panelists that we're very fortunate to have on the show - Tobias, talk to us about "When Voice Meets Impulse Buy." You wrote a fascinating piece on Voicebot.AI - really pleased to shine a light on it - but talk to us about your thought process in writing that, and the experience you went through that led up to writing that article.
Tobias Goebel: [00:48:45] Yeah. First of all, thanks for picking this one for this week. I wrote it on an impulse. You know, I didn't...it didn't only make that purchase decision on an impulse, but I also wrote this article on impulse, because I realized that that is the key. You know, the way what I describe in this article is two things: first, I shared something that literally happened in my household - you know, two stories, actually. One was the fly swat, and I made an impulse buy decision to get a fly swat when I saw that fly in the kitchen. And the other one was the broken cable, and turned my head around and said "Alexa, re-order lightning cable." Sure, this only works for commodities, but, you know, think about it: it's truly amazing and it's a dream come true for marketers. We've been trying to reduce that time from...you're essentially condensing that entire customer journey from researching and deciding to buy and then finding a vendor and comparing offers and making the decision getting it delivered. Except for the getting it delivered, and Amazon's working on that, this all happened in 20 seconds for me. And, you know, in the second part of the article I then basically try to explain that look, it'll take time for people to change their behavior and think of the Echo and remember that they can do that with the Echo. But as I try to argue in the article, history has shown over and over and over again that we do make these habitual changes when they bring us an added value from a convenience factor, one from a factor of you know getting stuff done or accessing information when the user experiences and user interface is right. And I think again all of us coming together with #VoiceFirst. And that's why I think - I'm convinced - that this will happen. People will be making purchases in the future from their home, from their couch, within 10-15 seconds.
Patrick Givens: [00:50:44] Yes absolutely - Tobias, really a wonderful piece. It was actually great to read - it put into words a lot of what we have been thinking and you know discussing around the edges of the office, and what voice commerce - #VoiceFirst Commerce, I should say - really winds up meaning to us who are in the business of building brands. Because...I love your metric of 'thought to transaction' and even your point in the beginning, that this works for commodities. I'll say the exciting and terrifying component to this is does it just make many things that previously were not commoditized commodities? Does it start to reduce the role of brand? Or the way that I like to think of it, with a little more of a sunny outlook, is what is the role of brand, when we're going to an 'ask and you shall receive' shopping model? So it is tremendously interesting for our side - it's this, you know, big...particularly I look at this in the CPG space - the consumer packaged goods. And you have your big players - your Unilevers, and your Procter and Gambles - and others in this space who spent so much time over the last, you know, many decades, diagramming every step along the way to the purchase of their brand itself. And suddenly that purchase, in increasing numbers, isn't happening anymore. So what is that new journey to making the decision? And, to the points in this article, the moment when you make that decision can now be the point of purchase, wherever you are. And so what is our new model for building brand affinity earlier? For getting people...the new game for top of mind attention is going to be tremendous in this space, and a huge question from the brand side of "what is the right way in the market side?" "What is the right way to position yourself in that decision making process?" And then, as the platforms mature, from the platform side of how much preference do they give to their other products? Obviously looking largely at Amazon in that space, if I had been marketing call it Duracell or Energizer batteries, suddenly I'm sitting at home trying to use whatever device out of batteries, I'd just say "Alexa, order me some batteries." The result that comes back probably isn't going to be the same one that I would perhaps have picked if I'm standing in front of a shelf of batteries, and choosing between all these options, with the different connotations that those brands spend tons of time and money building in my head beforehand. It's going to be whichever one Alexa chooses to offer up first, and say does this work? Chances are I'm going to say "yeah, sure, send me that one." So a huge new good opportunity for ways to think differently, and capture - from our standpoint - a disproportionate share of that new purchase volume, but also, really, a challenging way to rethink the whole what we've forever called the path to purchase.
Brian Roemmele: [00:53:31] Yes. Tobias, incredible article. I've been spreading that around, and it really really not only magnifies, but focuses on the whole challenge that brands really have. And echoing what Patrick is saying, and then some, I think that anybody listening to us today that is selling anything - and I would say most definitely any CPG brand - if they don't really have a voice-first narrative and storyboard and persona developed, it is getting very late in the game. This is going to be equivalent to what brand-building was in the 1950s and 1960s, when the TV has arisen and it is just a big big opportunity. So, yes, brands are possibly going to be passed by. But it does give the opportunity for the brand to reinvent itself, so Tobias has brought attention to that.
Bradley Metrock: [00:54:36] Excellent. Yeah, great article, Tobias - really well-written.
Bradley Metrock: [00:54:40] With that, we turn to our final story of the week, which is about the intersection of publishing with voice-first technology, discussing how Fashion Fix Daily has become the first fashion magazine to publish on the Amazon Echo, and this is a really interesting one. Patrick, to close this out, I turn to you: what are your thoughts on this story?
Patrick Givens: [00:55:02] Sure. So thinking of #VoiceFirst platforms - general Alexa, in particular - a publishing channel is dead in line with how we are really excited to explore this on behalf of brands, but actually for publications as well. So earlier this year VaynerMedia actually acquired PureWow - a women's lifestyle publication - and we've been working closely with them for what their approach to the voice-first space will be. And then, at the same time, on behalf of brands we've recently launched a flash briefing for The Digitalist, which is an SAP property, as well as Gary has a 'Gary V 365' himself. So we're looking at this from the publisher side, from the personal branding side, and from the large-scale publication side, and for all three of them, really excited about the opportunity to publish through a channel where the user really is in control. I think we're in a super interesting moment here, where we went through our video revolution maybe a decade ago - put users in control of what they were going to watch, first with YouTube, and then through the different user controls in OTT platforms. And we've seen this renaissance in the quality of video programming and exploration and innovation in the kind of narrative structures that people use, but also how you can really deliver a message over video when you're putting users in control of that experience. It's super-exciting to see that happen in video, and now I think we're getting there in audio. You know, we had music, notwithstanding given that iTunes obviously spun that out, in to a place where people were in charge of their music, but the rest of our audio experiences, up until a few years ago, have been pretty much broadcast-based. You're basically tuning in to a channel, and being given something to hear - maybe you're seeking out a publisher, but pretty rarely. And now at the intersection of what podcasts are evolving into, and where the voice-first platforms are going, we have this incredible opportunity to take audio content and put it out in such a way that either a, in a flash briefing model, users are opting in to hear something from us everyday. And then what's that little nugget of information - something's that's gonna set me on the right start to my day, or a little update that I want to share on a daily basis, or even interactive skills where we're still using them as a publishing hub...people might come and find a story, be able to seek out exactly the story they want to have, or, as we advance a little more here, go from seeking out a story, to turning a story into a conversation. I think that's where this gets super-exciting, when we look at the ability to publish, again, whether that's on behalf of an actual publisher, or on behalf of an individual brand, or a personal brand - something where it's someone speaking for themselves - creating a destination hub where you can effectively foster conversation in real-time about the topics that you're an expert on, is just so exciting for us. This is a place where we're investing a ton of resources over here exploring the new opportunities that starts to open up.
Bradley Metrock: [00:58:01] Well and this is exactly how my company, Score Publishing, got into the voice space...is clients asking us "alright, we've got this existing content...we've got this voice ecosystem over here, these various voice assistants...we don't know even the first thing about them, or what to do with it. How do we leverage our content in this new world? And so it's fascinating for me to see this as well, and appreciate your take on that.
Bradley Metrock: [00:58:25] So gentlemen, thank all three of you for your time this week.
Patrick Givens: [00:58:28] Thank you so much - it was wonderful to be on.
Brian Roemmele: [00:58:31] Thank you so much.
Tobias Goebel: [00:58:33] Thank you, Brad.
Bradley Metrock: [00:58:33] Thank you very very much for sharing your time and expertise, not just with me, but with the audience as well.
Bradley Metrock: [00:58:39] And with that, this is Episode 5 of This Week In Voice, for August 3rd, 2017. Thank you for listening, and until next time.