Artifacts and dreams of our science fiction childhood are coming to pass at a greater rate than ever before. Even the elusive flying car seems to be getting some traction. But what about the tools and technologies that hover around us that help us get our work done every day? Our phones have replaced hundreds of individual products in our lives and yet we see this as obvious today.
This week on the show author, speaker, and futurist Sid Probstein joins us to talk about the perils of prediction in technology, and the delights that come at the end of the trending rainbow.
About Sid Probstein
Product and technology leader with entrepreneurial background in startups, enterprise software, cloud, messaging & collaboration systems, search engines, natural language processing, big data, machine learning, AI, remote & distributed product development, etc. Python and recent c# hacker. Social media and live music fan. Author, speaker and futurist.
Links & Notes
Pete: Welcome to “Shared Insights,” the podcast from BA Insight. My name is Pete Wright and I am joined today by Sid Probstein. Sid’s experience as a technologist and entrepreneur across our sector is, in a word, vast. Along with his day-to-day, Sid is an author, a speaker, and a futurist. And today we’re going to be taking on one of the peskiest questions in technology. Where is my flying car? Sid, welcome to “Shared Insights.”
Sid: Thanks so much. It’s great to be here.
Pete: You know, I joke but not really about the flying car thing. I think it’s a provocative question for me because we have so many wonderful things out of my science fiction childhood, that these things that are now real, right. These tangible artifacts and tools of the present that are ripped off the pages of my favorite books as a kid. So, you know, to you, as a guy who spent so many years in technology bringing companies to life, I suppose it should come as no surprise that you would bring this topic to us today. Everything that was science fiction when I was a kid is now real. So as a grown-up kid, Sid, what are you most excited about from your science fiction youth that you can touch and use right now today?
Sid: Well, you know, I think the best thing is the fact that it is all here, pretty much, not everything, but almost everything. So we’re sort of living in that movie, or, you know, the things we used to see in “Creature Double Feature,” on Sunday afternoons or Saturday afternoons. My personal favorite, of course, I think is everybody’s, is what we used to be referred to as a phone, now I refer to…or a smartphone, now I refer to it as an appendage or an organ. I think Brad Feld, the great investor, VC, had a line about how, you know, if he got 30 or 40 feet away from a room having left his phone, he would feel physically ill, and, of course, that reminded him to go pick it up. We depend on our phones for so much and, you know, I think we could easily list a hundred things, a hundred products, if you will, that one phone replaces, right? Like, it’s our alarm clock. And it’s also tracking our health. So it’s kind of doing a bunch of device functions there. It’s replaced my computer at some level for browsing the web. And of course, it’s a phone, it’s a compass. It’s a flashlight.
Pete: Certainly radio, and television, and movie theater.
Sid: Right, I mean, you name it, it almost…it takes it up. But fascinatingly, right, we see this now and say, “Oh, this is so obvious, the App Store, the platform.” But the truth is predicting the future is hard. And it’s wrong to say anybody ever said that it’s particularly hard. Just predictions are hard. “Star Trek” for example, one of my favorite shows, one of the most influential science fiction shows out there had it pretty wrong. If you remember their view on devices, right, was that you had the communicator, the phaser, and the tricorder, and actually there was a separate medical device. And clearly, that’s wrong because they are all gonna be probably in one device or just about one device. Maybe you could argue that the phaser has special power requirements.
But I think, you know, that’s an example of how interesting science fiction is when you look at predictions. It’s easy to get trapped in what we’ve always done and hard to think about how could we really be different? Can we imagine, for example, something more cutting edge than a phone? A folding phone, not that difficult, right? But let’s imagine the phone that we can wrap around our shoulder like it was a bandana. Or pull it into the length of, you know, of a belt if we need to do it that way or perhaps if we wanted to be a sensor in that regard. Or compressing it down so we could fit it in our shoe. Or fly along above us taking photos. One device could do many things. It’s sort of the ultimate sci-fi concept. So super exciting if you can.
Pete: Isn’t it even more delightful, you know, one of the things I think “Star Trek” probably got more right than others is the idea that we can say “Hello, computer.” And suddenly the ship is talking to us. And not only that, we have this sort of integration with the vocal robot with all…that are competing with one another. How many different names can I call to activate all of my various computer agents all at once without even seeing the device that they’re connected to?
Sid: It’s one of my favorite parts too, you know. And obviously it inspired Brin and Page to create Google at some level. The Star Trek computer, never misunderstood, never had to ask about the context, never had any questions about the role, or who had full command of the ship, and all of history it seemed like it didn’t ask for your PIN number. On occasion, actually, I think there are one or two episodes where there was forbidden information that a special password was required. So in that regard, it’s certainly true. But that’s still incredibly elusive, you know, with respect to something like having a device that can do many functions. Or frankly just the fact that, you know, in a very sci-fi way my phone can track my health if I had the Apple Watch, or any of those watches not to pick one. It collects a lot of data on me that can be used for all kinds of analysis.
That seems further along than, say, Siri or Alexa. You know, Siri sometimes has trouble playing a song I actually have in my library, right. The other day I was bringing up podcasts matching the same…the song I was looking for. Not even close, right? And so I think disambiguation and some of the other problems that plague, you know, broadly this…the world of search, the world of query understanding, those still haven’t been solved. So, we’re a ways from sci-fi there.
Pete: Well, we should talk then about some of the trends that are sort of influencing the investments in these new technologies, right? You know, some of the trends around how we’re handling disambiguation, how we’re handling, you know, AI, machine learning, and how that’s influencing the future’s sort of invisible computing for us end users.
Sid: That’s totally right. And, of course, that’s solving that problem that’s so front and center. And the good news is people have been trying to understand what people mean when they type a word or two or three and for the purpose of showing them ads. And so a lot of work’s been done there in terms of, you know, random approaches, saying, “Let’s show a bunch of random ads and see how different audiences react. And then try to predict what audience of a new user is, and then showed them the stuff that produced the most response.” So-called quantum thinking, right, to now a lot of big data approaches, looking at query logs, and getting users to rank or evaluate results and then trying to you know, crunch that to find the patterns. There’s certainly a lot of progress.
The trick though, is you really need a pretty complete solution. You need something that’s capturing, for starters, all of that activity so that you can even begin to analyze it. You need to be able to classify a bunch of things, and classification, much like prediction, is very hard. You wanna know, are they looking for a list of things? For example, if they’re searching, is this related to something temporal, like is it an appointment? Or is it a research topic? Or are they looking for a specific answer? It’s sometimes that notion of, are they trying to solve a mystery? Are they trying to fit together the pieces of the puzzle?
All of those things are drawing a lot of ad dollars, particularly when there’s investment dollars I should say, particularly when there’s money to be made, like advertising or an area where the risk and cost is very high. For example, manual processing of mortgage documents. A lot of work is going into being able to use sort of next-generation OCR and then the automatic resolution of the most common errors so that humans don’t have to do that. And somebody can get a quote or approval almost immediately for a loan as opposed to potentially waiting for days feeling anxious. Right on up to, you know, compliance and knowledge work where do we really have two weeks to pull together the data to get an answer? When the position that we would take, if we thought with a particular prediction was gonna go one way or the other, you know, only lasts for an hour or a day, is it worth investing for two weeks? So sometimes the investment isn’t speeding it up, sometimes it isn’t getting that insight, but wherever, you know, there’s a compliance problem or a real revenue opportunity problem, that’s where spending is lining up.
Pete: How does this line up for you around the cultural sort of costs of these new technologies? I mean, you talked about, you know, the investment that comes from opportunity around today, advertising. But we also have a little bit of a backlash going on right now, and I say that with heavy air quotes a little bit. It feels like there is this groundswell of the human cost becoming more transparent in and putting pressure on big data companies in acting in a more principled manner. That letting advertising drive investment at the cost of human users is a detriment to the pace of just straight up innovation. And at some level, you need people to feel safe using these technologies so that you can learn how the technologies will interact with humanity, right? So what is your sense on where we are in the long arc of history around invention and innovation in this space and the role of just straight up doing good in the world?
Sid: Well, I certainly think the recent events, all the different data breaches, and selling of data, and leaking of data, and the fact that you know, leak is our household word, right, because of politics. It’s important to remember that nature hates a vacuum. And so as much as a lot of these stuff has happened, look at some of the things that have come out of the positive side of all this intense connectivity in social media. While there has…there’s been tremendous communication of, you know, the some of the threats of AI, the concern about AI, or some specific people who have been misusing it.
Now, that’s not to say that that’s always been heard, right. But with higher awareness of the issues and the possibility, the risk of what can be done with data, and what can be done with prediction, the biases that are present in some of these things, those messages, discovery of those things can be transmitted faster than ever, right? And we don’t have…that’s not something we have to imagine or it’s not a stretch. A company that, you know, is breached or does something wrong, even if it’s for one customer, sometimes can be lit up on Twitter, have their reputation immediately questioned, right? And if they really have acted poorly, usually there are significant consequences in share price or in just sales. And executive teams, you know, while it’s an axiom that the management will get away with it, it’s not necessarily the case, right? These kinds of things can be spread very quickly. And again, if there’s typically some reality to fraud or offense or misuse of data, that’s things that come back to haunt them.
It’s true also we have to see a little bit, right? There have been some major violations in Europe. And while it seems like GDPR has teeth, we’ll see if it’s, you know, pennies on the dollar for someone like Google or Facebook with vast, vast sums of cash and continue to earn it. Or if it’s a more material punishment, and if they really do make compliance an important piece of the future.
So I personally feel positive. I think it’s important to trust and verify that the progress is being made. And, you know, the truth is, these things are going to happen one way or the other, and people participate when they are publishing their thoughts and commentary in public. And perhaps that will be enough to really get, you know, something together that we ultimately see as more AI, and more intelligent, and more of science fiction. And more like that, you know, “Star Trek” computer than things we play with today in a positive way.
Pete: Yeah. And it’s certainly in a way that, you know, more people feel like they can trust to bring into their homes and bring into their lives. Which I think is one of the most interesting kind of trends that we see when we look at, you know, the future. We wanna talk about the future of search. What is the future of search look like in this space? And what are the things that you are looking forward to seeing develop and grow?
Sid: Well, it’s a great segue from Google and thinking about search and how things have changed. I was speaking with some friends the other night over dinner, and we were arguing about trivia. And of course, we instantly resolved it. And it became so clear that there is no trivia anymore, right? If we’re in a bar arguing about trivia, well, we’re just sort of ignoring the fact that the phone can answer that question. But there’s…that’s certainly an order of concern. Would you call it like a game-changing use case for search? I don’t think bar trivia is a game-changing use case for search. But there are, and where once you have the index, the capabilities to answer a question, then the question is how can you anticipate that question and not make the user ask? That’s the real future of search.
So partly, it’s always important to separate what Google does from, so-called enterprise search where you’re dealing with behind the firewall, inside the company search. Google can’t answer the typical questions there, which often are list-oriented or specific to the company and aren’t documents that Google can see. So just getting it those things can be hard.
Search by nature finds documents. And in the corporation…in the world of corporate search, very often we’re trying to find lists of things like you know, which customers bought particular products with these features. Maybe mentioned something in a discussion with our sales teams, or with our service teams. And ferreting those out can be difficult, getting those lists together can be difficult. But if we think about it, the future of search is paying attention to that and a whole lot more, like my personal context. So if I have…let’s say companies as Office 365, right? So my calendar, etc., it’s all online, and so are calendars and inboxes of my colleagues. Well, if I authorized it, then some agent could sit there looking and see who I’m meeting with, check my correspondence with them, and ultimately start to generate and then for what queries should I be running for that meeting? And when should I run them? And when would I wanna look at the results? And how would I wanna see those results?
That’s a lot of what the future looks like, where much like, you know, my phone will remind me it’s time to leave for an appointment. Actually, I was late today but that was my fault, not Siri or my Google Calendar. But that’s the beginning. That’s the first layer of the onion. A deeper layer is, you know, here’s your briefing for your meeting. Here’s alternate route information. Here’s location of a good venue for lunch during a meeting or a debriefing spot. Here’s a great article that you may want to contribute because it was authored by one of the people you’re meeting with, or one of their team members, right? Those are some of the interesting connections that search can begin to surface. And whether you’re asking for that specific information, which I think is great if you can fulfil it, but even better is to have that brought to you proactively. That’s more like the future.
Pete: One of the things we talk about on the show all the time is the Google problem when it comes to internal search, right? It’s this, the idea that as consumers at home, we’re searching Google and we’re searching Amazon. And then we bring our expectation of search performance into work and the internal search we expect to live up to our experience at the kitchen table. And I think about that in terms of YouTube’s vaunted algorithm, right, that, you know, it recommends videos based on what I’ve already watched that it thinks I may like in the future.
And the gap between insight and insanity on YouTube’s algorithm is, I think, a great example of the where we are right now snapshot that I have a fantastic list of recommended videos. And if I look at one puppy video, the whole thing is dogs on motorcycles, right? It’s adorable, but it has gone way too far in the wrong direction. And that I’m hearing more and more from folks who are aware that this is happening and have higher and higher expectations that it’s going to get better more quickly. And I think that, you know, as you say, I mean, there’s a lot of opportunity yet to be squeezed out of some of these tools to make them better and make our expectations align more clearly with our reality.
Sid: I couldn’t agree with you more. You know, I was actually reflecting the other day on this. And I believe it’s, you called the Google problem is one thing, to call it the Amazon problem is another, because I really find it almost impossible. And many people I know in search have said the same thing. It’s very hard to shop on Amazon. I can buy and I do buy, I’m a really repeat serial customer. But the end of the day when I am not sure what product to purchase, I don’t go to Amazon unless it’s something like electronic cable or something like that, you know, or if it has a title, if I know and, of course, that’s a different matter, that’s a known item. But the problem is it’s not good, and it has so many products and it’s not really good at helping you understand what are the elements of the thing I’m trying to buy that will help me navigate?
So the other day I bought a…actually, earlier last year I bought a new ski jacket, and I wanted a warmer ski jacket that was important. And I could not find, although all the deals were present on Amazon, I couldn’t find the one I wanted. In five minutes, I found a specific website that sold pretty much nothing but ski jackets. They showed me the three categories, one of which was that, you know, heat rating. And I very quickly found what I wanted, then I bought it from Amazon from the same firm I wanna say so I didn’t cut the middleman out there. The same firm also sold the product on Amazon, so they ultimately got the deal. So that’s one thing is understanding what’s important to someone who’s buying a ski jacket versus what’s important to someone who’s buying products or clothing. That’s one aspect.
The other is, once you’re right, once I started searching for ski jackets, now Amazon, and Google tool, and another lesser extent are gonna sit there showing me ski jackets endlessly even though I’ve kind of finished that mission. They lost a little bit the temporal aspect, and they understand that I don’t want more now. Maybe if you had a real deal I’d consider it, right, within 30 days. But all of these engines have lost an important thing, which is serendipity. And that’s, of course, why, you know, the market crash years ago was so unpredictable. Models only considered this housing market going up. There was no idea that suddenly randomly could turn in another direction.
Well, it’s very much the same with these. How do you…what do you infer from the fact that I purchased a ski jacket? And how could you…what could you show me that’s serendipitous and related, but not another ski jacket? Maybe gloves or a hat or something that matches that’s on sale from the same brand. But that doesn’t seem to happen. So I think those are, again, using that similarity is something that other engines that are building on top of and learning from Amazon. Like Wayfarers, for example, right, they have a whole different approach, but they sell far less. And that is the future, is a deeper understanding of how to buy something for a specific use case. And then ultimately, some generally aggregating those altogether.
Pete: Some of the things I’m most excited about in terms of our, you know, way out there technology from the future, lighting me up as a kid kind of tools are things like Waze. You know, when I can look at my Maps app, or I can go to Waze and get insight from the sort of outsource community, the outsourced transportation brain that feels like I’m living, you know, 50 years in the future right now today like when I leave my house. I know things that I never would have imagined that I could see in such an easy snapshot right in front of me on my device. What are the things that excite you to think about in our, you know, 25-, 50-year horizon?
Sid: Well, I do think, you know, the incorporation of people and their perspectives in a seamless way is very powerful. And look, it’s pure science fiction, many movies have been written about it. You know, imagine if we all wore headsets, got like the classic Google Glass but they’re broadcast too. Anytime there’s a moment, an event of any kind, minor or major, dozens if not hundreds of people can be viewing it. They can be broadcasting it. Drones can do the same thing too. We can essentially imagine being able to walk a crime back in time, right? And say, “Oh, you know, this bad thing happened. Back up and track all the people involved because we’ve been recording everything, tracking everything.” Those are scary, right? But there’s a great use case too which is the child who wanders off. Or you know, a person who needs assistance because they aren’t able to navigate through some complex problem in business.
I mean, I can come and help that up too. Here’s how you handle this exception, stepping in and interceding, maybe connecting in the expert. There are some great examples out there in health care, particularly of using a device at the edge in the situation, some actual live situation where someone may be injured or hurt. Getting the data and then transmitting it to an expert halfway around the world or two-thirds the way around the world. Or right next door, you know, if they don’t…but they aren’t able to actually go to the scene and getting the best possible diagnosis and analysis. Those are all stages of the connections between people that can be positive ones, right, and not like social swarming or something more negative.
Pete: One of the things that we get so much sort of criticism is that technology is, you know, at the heart of fracturing human attention, right, that there’s just too much signal that creates too much noise. I really like what you’re saying here because it brings to mind kind of a, something I’ve never really thought about that, you know, 25 years from now, we may just live in a period where our attention is trafficked, right? It’s not fractured, it is guided by technology and trafficked in a way like a circuit board, in a way to actually make good on the efficiencies that we’ve been hoping about all along. I think that’s a really exciting concept.
Sid: There’s no doubt that, you know, we can imagine it’s very easy for us to imagine AI taking over replacing us all these negative consequences. But the truth is they’re unlikely. They’re much less likely than a great median outcome which is AI becomes assistant to the human. Today the best chess player in the world is neither a human or a computer, it’s a combination. And the idea that you are almost in the matrix, right, and you’re Neo and you need to learn kung fu and you download it. Well, maybe it’s not quite like that right away. But you need corporate kung fu, you need parenting kung fu, you need shopping kung fu, you need travel kung fu. And the AI is there linking in experts or deriving its own intelligence from their writings and photojournalism and all of that, and they help you and you have a great outcome. If that could be something that’s, you know, sustained for a long period, at some point perhaps we lose interest in some of those functions, right? And then be happier about yielding those things to an AI that’s learned from working with us as opposed to, you know, being scary and against us.
Pete: Also, I wanna learn real kung fu and how to fly a helicopter.
Sid: Especially if I could just download it.
Pete: That’s the stuff. Sid, thank you so much for your time today. Anything that I have not asked you that you’re excited to make sure to get out there to wrap us up here?
Sid: Well, I’m very excited about some of the changes that folks are talking about on Twitter. And, you know, there’s been a lively debate about that and all the other things that we talked about. Please, you know, feel free to visit me on twitter at @sidprobstein.
Pete: And linked in the show notes. Sid, thank you so much for your time today. I really appreciate it. I hope this isn’t the last time our paths cross.
Sid: Same here. Thank you.
Pete: On behalf of Sid Probstein, I’m Pete Wright. Thank you, everybody, for downloading the show this week. We’ll catch you next time right here on “Shared Insights,” the podcast from BA Insight.