« All Episodes

Interview with Kevin Kelly (Part 2)

Published 8/11/2017

In today's episode, I talk with one of the most influential voices in technology in the last 20 years - Kevin Kelly. Kevin is the author of "What Technology Wants" and "The Inevitable", co-founded Wired magazine, and is now leading the charge of optimism as it relates to the future.

Today's episode is sponsored by Rollbar. With Rollbar, you get the context, insights and control you need to find and fix bugs faster. Rollbar is offering Developer Tea listeners the Bootstrap Plan, free for 90 days (300,000 errors tracked for free)! Head over to rollbar.com/developertea now for the free 90 day offer!

Transcript (Generated by OpenAI Whisper)
I want you to take a second and ask yourself what your job title is. And I don't want to use the terms developer or strategist. I want you to think about what your real job is, what value are you providing. And in today's episode, we're going to talk about how this is changing. We're talking with editor and author Kevin Kelly. Kevin wrote what technology wants and more recently he wrote. The inevitable. Kevin is also the senior Maverick editor at Wired, which he co-founded in 1993. This is the second part of my interview with Kevin, by the way. Go and check out the first part if you missed it. You can find it at spec.fm. And of course, if you're subscribed in whatever podcasting app you use, it should be the one right below this one or above this one, however you have it laid out. But Kevin is a voice of reason and he's also a voice of optimism when it comes to the future of technology. And the future technology is going to look, well, it's going to echo very similar things from the past. Some of our job titles are going to change some of the things that we do on a day-to-day basis to create value and an economy. Those are going to change as well and we're going to talk about that in today's episode. I'm excited to be joined by Kevin. Make sure you subscribe in whatever podcasting app you use while you're listening to this episode if you don't want to miss out on future episodes. Let's get into the interview with Kevin Kelly. Have this struggle of telling people what I do and it's already that way, right? We're already beginning this kind of strange path of brand new jobs that my parents didn't even have in their perspective when they were my age. And continuously that's becoming more and more thing. And obviously, before 2007 or so, somebody like a social media strategist, nobody even knew what that meant, right? And some of the fundamental concepts are still the same. You have to understand communication, you have to understand one to many or many to many or whatever it is that you need to understand. But there's titles and the specifics of the job. They become new. It constantly new versions of that. And it's so interesting to hear that kind of wrap up that ultimately becomes, let's find ways to hire new technology. That's what it is, right? Exactly. Right. Yeah, I'm not really worried. I'm not that worried as many people are about employment and automation because I think, like you said, we will invent entirely new job careers titles that we can't even think of right now. You know, like, I don't know, artificial intelligence strategist, right? I was like, yeah. There's somebody, or like, I whisper, somebody who's really good at understanding how AI is working. They're AI repair person or AI psychiatrist. You know, there's just so many new niches that are going to come along that are very hard to imagine and they seem ridiculous to us. As you were saying, the most web designer, you know, mortgage broker, these wouldn't even make sense to the farmers 150 years ago whose jobs have all gone. And so I think we're not going to, these new jobs aren't going to make sense to us now. Yeah, I think you mentioned the web designer or rather AI whisper job. And I think this is actually going to, the way that this will affect us earliest and you crosscheck me. You're, again, you wrote the book on this, but I think this is going to begin by web developers, kind of the people hiring web developers or software developers, adding new bullet points to that requirements list, right? I'm saying, okay, well, if you're going to be a software developer here, then yes, you need to know Java and X, Y and Z, but you also need to have experience with machine learning or you also need to have experience with this, you know, whatever AI platform API thing that is out in the wild. I think that's kind of the start of that because, you know, as we saw with software developer titles in the beginning, it's software developer because there's really only three languages, right? Assembly and C and, you know, whatever other thing that existed. And then it continued to branch, it continued to grow and expand. And as it expanded, each of those individual areas got large enough to justify a person that is solely dedicated to that thing, right? So now we have people who are even subdivided inside of a given programming language or subdivided inside of a giving tool set that are really focused on one particular aspect of that software development process, whereas previously it was really all one title. I think that's going to continue that direction. Right. I think so too. So you talk about AI and you mentioned a few times this term bots and because I've read your material, I know what you mean by this. I think some people listening, maybe envisioning the Boston Dynamics robots, you know, coming in and being cashiers. I want to kind of correct that perspective and talk about this concept you use towards smertness to describe or cognifying cognition to describe AI. Can you kind of walk out what a cashier bot would look like? Yeah, so the main concept I think is important to put out is, can begin or be done at the beginning with a very simple thing of just talking about AI in the plural, so it's AI's. That's just that what we're going to basically invent is a whole zoo of different types of thinking, different types of minds, different types of artificial minds. And I use robots to mean both AI's which don't have a body and robots which do have a body. So bots for me includes AI chat bots, AI Alexa, which don't have very much of a body and then AI's have a body which are robots. Oh, there's Alexa. I woke her up. Alexa, quiet. And so I think that the image that you want to have your mind is of many, many different species of thinking and some of these are just types of smertness that don't have consciousness. They're maybe they're kind of like a vegetable, vegetable level. They're smart, but they don't have very many dimensions. Your calculator is smart in the arithmetic. It's smarter than you are in arithmetic. And the GPS navigation is really smart and spatial navigation, but it doesn't have anything else. But we will also make more complicated and complex varieties. And so it'll be like a whole ecosystem of different types. And some may have very large amounts of consciousness and maybe they aren't so smart. There will be, you know, AI's that are extremely perceptive or have great perception. There may be a little bit of language recognition and they've got, you know, a little bit of spatial, but they're memory short, whatever. So the idea is that there are many, many, many different types of these species of thinking. And there's some set of primitive elements of cognition, which we're in the process of identifying and will invent more of them that can be recombined and remixed in many different ways. And that each of these will be engineered to specific uses and the ones that are in a kind of a car driving may not want to be conscious. And we may find that it's a total distraction to have consciousness in a car because that's what we have and we're distracted. And so no, you, you, these are conscious free drivers. And then there are other places where we, you know, maybe someone, some bot that's making mortgage decisions or something. You want to have some sense of empathy with people. So maybe there's a lot of empathy, maybe some consciousness there. And so it's plural. There's AIs and there'll be thousands of different types. Yeah. And what's so interesting about this concept to me is the idea that, you know, that we're trying to replicate some human form is really short-sighted in that there's so much more that we can do if we remove certain limitations that the human can't remove, right? So and you mentioned this, I believe in yourself by Southwest talk. But you discuss this concept that, okay, well, the thinking portion of the human or the learning process is really the part that matters and the thing that we want to utilize is that or the thing that we want to mirror perhaps is the learning process, but not the humanity side of it, right? Because, you know, a human has quite literally physical limitations that we may not want to impose on, let's say Alexa, for example, right? Very, very interesting opportunities that begin to show up when we say, okay, no, we don't want to make this more human. We just want to take the good parts and then lean on the things that the technology can do that a human cannot do, the things that the technology can do better than a human could do. Right. The other aspect is that we don't need to make humans because they're pretty easy to make right now. And for me, the chief benefit of AIs is this because they think differently than us and that there are certain problems they can science or business that are so going to be so difficult, maybe quantum gravity, whatever is that our own intelligence is alone, working alone may not be sufficient to solve it. So we have to invent other kinds of thinking. It's not that we want them smarter than humans because I think, you know, like, as far as the humans is probably still going in the same direction that we're going, you want something that's coming very different approach. So you want a different type of thinking that would help us solve some of these problems. So we're going to solve in a kind of a two-step thing or first invent another kind of thinking that together working with us can solve that. So I think it's when we're trying to be innovative and thinking different that it's actually very hard to do. Thinking of it's very hard to do when you're connected to everyone else all the time. When seven billion people are connected continuously day and night, there will be groupthink. There will be, it will become really difficult to come up with a truly different idea. And I think AI can actually help in that assignment of even spurring us to keep thinking different and working with them together to think differently. When we have this massive, super-organism of all the humans together in this kind of converging global culture, it's going to be a challenge to keep coming up with wholly different ways of a problem. And I think the differences of AI, the non-human aspect of them will help us do that. In a way, it could assist with our creativity and the inefficient things that we do need to continue with. Those are the things that are truly unique. Right. So the future of software programmers, I think, is going to be, there are going to be a lot of AI going on where you're going to be working with some kind of a system that may be proposed to you different solutions, whatever it is, I don't know how you call it, codes, systems, loops, all this kind of stuff. You may work with it and it's going to be proposing things as a colleague might or completing things that you do in a rough form. And together, you are going to form the center, they call it, this team of human plus AI working on a problem. And some people will do that better than others. But then some people will be involved in making these AI's that are more creative in a certain direction than others. You might come to favor, like I really like working with prospect over here because prospect is a good compliment to the way I think. Your fellow colleague is really going to work with this other AI because it suits her approach better. And I think working with the bots is going to be how we're going to end up as something that's going to replace everything we do. I think we're going to end up working as a team. We're going to take a quick sponsor break and talk about today's sponsor, Roe Bar, and we'll get right back to the interview with Kevin Kelly. Today's episode is sponsored by Roe Bar. Roe Bar is an excellent service. For those of you who write software on a daily basis and you have people that are actually using your software, that's going to be most of the people who are listening to this podcast, that software is going to have bugs. I hate to break it to you. You're going to have bugs at some point. And in fact, you're going to have bugs that you don't find until after they've shipped to production. Now the problem is tracking these bugs down can be expensive and it can be very difficult. In fact, it can also be costly because sometimes you're even relying on your users to tell you about those bugs. And that's a worst-case scenario because by that point, you probably already lost money. You may have even lost that user entirely. On top of that, you're going to end up going and digging through your logs, writing extra code to try to do some of the things that Roe Bar is actually going to do for you. And that's exactly what Roe Bar has specialized in. They're going to help you track your errors better than you could in your own custom way. That's because Roe Bar has thought of pretty much everything. They integrate with all of your existing services. For example, you can send error messages to email. Of course, you can send a email, but you can also push it to Slack or you can push it to Trello. You can push it pretty much in the service. You can imagine really the list is very long. You're not going to cover the whole thing here. But imagine integrating all of the errors in your production applications directly into the channels that you're already looking at when you're working on those projects. That's what Roe Bar is going to do for you. We use it at whiteboard. I highly recommend it. Go and check out what Roe Bar has to offer. Roe Bar.com slash Developer Tea. That'll get you started on the bootstrap plan for free. That's zero dollars for 90 days. You can get started tracking your errors today for free for 90 days. Roe Bar.com slash Developer Tea. Thanks again to Roe Bar for sponsoring today's episode of Developer Tea. Yeah. Collaborating with both other humans and other bots in that. The word you use in the book is beginning this global system together. Really a very exciting prospect. And probably the one where people create that dystopia that is so prevalent in this collaboration is so often turned into a binary viewpoint where the limitations of those AIs, they create the bad. That's not the picture of the future that we're looking for here. Instead it's connecting humans better together and connecting humans with intelligence that can assist them. I actually have coined this white leacoins this term because I'm not a futurist. I'm not writing books about this, but this term of assistive intelligence. Which is a different system. Assistive or assisted? Assistive. And what do you mean by that? Effectively the idea that the intelligence of that thing is pointed at providing assistance to a human. Yeah. Specifically rather than solving something on their own, solving something only when interacting with another person. Yeah. And we, you know, there are probably going to be a couple of different modes that works. One is where they're proposing all kinds of like, you know, it's like brainstorming. Yeah. Where they are kind of brainstorming with you on these things and you may be kind of doing the curation. Like a caddy. Like a what? Like a caddy. Yeah, exactly right. Well, why don't you try this or try that or how about this? Oh wow, this is a crazy idea. And then you're kind of working with that. And then the other way they could do it is like an assistant where they're kind of finishing some of the details that you work with where you kind of do a broad outline and they do that. And there also might be ways in which there's more of a conversation about say like the structure of something where you are in a conversation about, you know, I think we should do this. And it's like, well, what about, you know, or do you do remember the other thing is, what about like the thing about AIS is they can have a complete mastery of history? Yeah. It's like, well, you know, when Samuel Jonson did that in 1870, when it didn't go so good or whatever. And so there could be this historical viewpoint. Didn't have to, you know, I mean, I can obviously in programming, there's going to be other more recent examples, but they could say, you know, over, you know, Tim over in Trinex, he, he, that loop worked, but the problem with that loop was this X and Y. And that kind of a comprehensive command of what had been done would be also make you smarter as well. So I think we could kind of imagine a bunch of different ways in which that assistive AI would work, just even just in your own little industry of, you know, coding and programming. It would, you know, and I would imagine that could be one of the first industries that would, that would use the assistive AI. Almost certainly. Yeah. And we like to build our tools for ourselves. Right. So why would, other than maybe, you know, a monetary or commercial reasons that are compelling. Are you aware of anybody who's doing that now? That's a good question. I know that some, there are some things like, for example, the Google drawing tools. These are things that use machine learning to try to figure out what your, the shapes you're trying to draw. There are some layout tools that, as you begin to create an interface, I can't quite remember exactly what they are. They're still very much in their infancy. But they aren't, they're not conversational. They're not the cognizance of these things is not visible. It doesn't feel like AI. It feels more like a tool, honestly. Yeah. I know that Facebook has claimed that they're using some of their AI to design other AI. And I don't know exactly what that means translated into, you know, an operational practical way. Yeah. So that is a claim. Yeah. There are some groups that are working on finding, you know, consistent algorithms that, as you're writing your code, the machine is trying to figure out, okay, where are the weak spots most likely to be? There's some things like grammar checking. It's like, it's like, we're checking your grammar as you go along. It's more like structural. Right. The grammar is structure. Testing, yeah. So, very powerful possibilities, obviously, in that arena. And, you know, it's very interesting to know as a developer, some of the pieces of the puzzle that make this stuff tick. And it's exciting. And really, I know we're on the very beginning of this bell curve for AI, specifically, a machine learning and those kinds of things. And it is very exciting, and the future is looking, you know, you discuss this idea that AI is going to be a service that is offered, you know, in large. And that's already happening. That's immediately it started happening already. We have image recognition technology that's available through an API. You hook it up and basically give it an image and it'll tell you what's in that image with levels of confidence. You have things like video monitoring or checking the video to see what items are in that video. You can find locations and all kinds of things. And they are available to Developer Today for pennies, you know, per image already. And it's only going to increase that level of contextual AI. Yeah. The complete, the actual thought for the benefit of the Rear. So, I haven't seen the book. The idea was that just as the Industrial Revolution made artificial power beyond the natural power of muscles of a human or an animal, and we could extend our reach of making things with artificial power to build, you know, skyscrapers or railways across continents and factories turning out miles of cloth. Well, because of artificial power, that power was distributed on a grid and it came in commodity. So that anybody could buy as much artificial power as they wanted or needed and to do whatever they wanted and to invent new ways, you know, electric pumps harnessing that artificial power. And now we're going to do the same thing with artificial intelligence, which certain varieties of that will be delivered on the grid called the cloud to anybody who wants to buy as much AI as you want or need. You don't have to generate yourself. You can just purchase, you know, TensorFlow, whatever, or go to Microsoft or IBM and it'll become a commodity. It'll become a utility. Maybe it's a better word that's available to anybody to, you know, plug and play. Yeah. It's so exciting because, you know, and one of my questions that I had for you in my, I have a very long list of questions, certainly didn't get to all of them today, but, you know, we could have this conversation for days on end probably, but one of the questions that I had for you regarding this specific subject, you know, you talk about 30 years, but for developers who are listening to this today, what does 2018 look like? What can we do to begin to bring this optimistic, you know, I really think developers are kind of at the ground floor of this as we already discussed. What can we do to start thinking at least towards this optimistic view of the future? No, I think we kind of hinted it. I think it started to tinker with these new tools. There are no AI experts given where we'll be in 30 years. I mean, there are some highly paid people now who know more than other people. There's a lot of money being diverted into AI just to take one example, but there are no AI experts compared to where we'll be in 30 years. We know nothing really. We're just at the first hour, the first day of figuring out the dish. So you're not late, you have a chance to become that expert and it's going to be, you know, like tinkering around with the stuff that some of the biggest breakthroughs are going to be found to. Very similar to the state of the Delta revolution. We had all these amateurs with their, fully around with electricity. Just, I mean, we forget that people had no idea what electricity was in the beginning. They didn't know the natural force. They didn't have idea of electrons. It was this amazing thing and people are tinkering, figuring it out. And we're in the same stage with intelligence. We really don't know what intelligence is. People think around with artificial intelligence will be one of the ways that we're going to figure out this. As I said, you can buy some now cheaply and you can start to mess around with it and do stuff because it can be done. And that is the, I think that's where all the big breakthroughs are going to be found. The easy hanging fruit are just sitting there waiting for someone to come along. And people, the gray beards in 20, 47 will look back and say, oh my gosh, I wish I could have been alive and young in 2017, when nobody knew anything and all these things were just waiting to be grabbed. And so it's, it's no better time ever in the history of the past because the tools exist now. And compared to what we're going to future, this is really the best time ever to make something happen. That's absolutely excellent advice. For developers listening, just understand this idea that the future is created by people just like you, people just like me. And the movement, these inevitable things, they are not inevitable in a negative sense. They're inevitable in a sense that you now have a picture of the direction, the vectors of the future. And it should be an empowering thing. And I'm really excited for the future. I'm really excited for seeing that 20, 47 moment come to life and recognizing, hey, some of these people actually did take advantage of that opportunity. Kevin, thank you so much for joining me on the show. Yeah, I really enjoyed it. Thanks for having me and for the great questions. Absolutely. And of course, people can buy the inevitable pretty much anywhere, right? In fact, I just posted a picture of, I found it was found in Costco at the discount book table on the new paperback version next to the Stephen King novel. So yeah, anywhere at this point, yeah, you're right. Excellent. Yeah, thank you, thank you, Gankin. Okay. Bye, mate. Thank you so much for listening to the interview with Kevin Kelly. I hope this has challenged your thinking and excited you about the future. But also created a roadmap in front of you of a lot of work, a lot of exciting work. The UNI gets to engage in on a daily basis. Thank you again to Kevin for joining me on Developer Tea. This is the second part of my interview with Kevin, by the way. Go and check out the first part if you missed it. You can find it at spec.fm. And of course, if you're subscribed in whatever podcasting app you use, it should be the one right below this one or above this one. However, you have it laid out. But thank you so much for listening, thank you again to Rollbar for sponsoring today's episode of Developer Tea. Of course, you can get started with the bootstrapped plan for free for 90 days by going to rollbar.com slash Developer Tea. And you can get started tracking your errors in any production environment. Go and check it out. Rollbar.com slash Developer Tea. Thank you so much for listening. And until next time, enjoy your tea.