Episode 2: Kevin Kelly
Collaborating with AI, Unleashing Creativity, and Embracing the Imperfect with Kevin Kelly
Episode 2: Show Notes
Kevin Kelly is a true visionary and creative mind, and co-founder of the iconic WIRED Magazine. Kevin’s new book, Excellent Advice for Living, serves as the foundation for the conversation, as we unpack the intricacies of the creative process and innovation. Kevin draws on his professional experience and walks us through the notion of embracing a multitude of bad ideas as a pathway to uncovering true gems of creativity. We uncover essential takeaways from the book, like the power of habits over inspiration and the art of setting ambitious goals. We also discuss the multifaceted nature of imagination, the benefits of embracing imperfection, what sparks creativity, and how to reach your full potential during the creative process. Kevin also shares his perspectives on AI-generated art, how he leverages AI for his creative goals, and the limitations of text-prompted creations. Tune in to discover profound insights as we delve into the art of idea generation, iteration, and the beauty of embracing imperfection on the journey to creative excellence!
Key Points From This Episode:
The value of generating numerous bad ideas to uncover true creativity.
Embracing iteration and the iterative creative process.
Nurturing the initial creative phase by defying judgment.
Normalizing the exploration of bad ideas as a catalyst for innovation.
Building the muscle of imagination and why it is essential to creativity.
Balancing deadlines with the pursuit of excellence and continuous improvement.
Continuously learning even after creative projects are complete.
Kevin shares the creative habits that he practices on a daily basis.
The role of observation and perception in sparking creativity.
He reveals his approach to creative collaborations.
Insights into the integration of AI tools into the creative process.
Making the problem itself essential to the problem-solving process.
Creating content for an audience of one.
Advice for adopting fresh perspectives and approaches.
Quotes:
“You have to iterate your way through things, and you have to throw these things away, and the only way you can do that with ease is knowing that there’s more where it comes from.” — Kevin Kelly [0:04:43]
“I believe that sometimes the best solutions will come from a really bad idea.” — Kevin Kelly [0:07:48]
“A really, really good idea is an improbable idea. It’s not going to be the first thing that comes to your mind. It’s going to be hard to get to it.” — Kevin Kelly [0:08:21]
“So there are functions or rhythms, and one of the rhythms is that in the genesis period, you have to protect [creativity] from judgment.” — Kevin Kelly [0:13:37]
“The chief attribute of a deadline is that it takes something that’s abstract and perfect and it makes it real, which means it has to become imperfect, and so you’re going to have to shift something that’s imperfect, but that means it’s real, and so that’s good.” — Kevin Kelly [0:21:29]
“I think observation is a key skill for anything that we’re doing. The best writers are incredible observationalists, they’re able to write because they’re observing things in a way that’s almost extraordinary.” — Kevin Kelly [0:27:58]
“It’s very easy to get these current versions of AI to produce something that may surprise you. It’s very, very hard to get them to do something that obeys you.” — Kevin Kelly [0:29:51]
Links Mentioned in Today’s Episode:
EPISODE 2 [TRANSCRIPT]
“KK: There’s always some drama about something that doesn’t work. It’s the real world, it’s real life. Having done 10 Oscar-winning movies, the next one will still be the same thing. It will be a total miracle that had ever gets done and that is good.”
[INTERVIEW]
[0:01:30.4] JU: All right, welcome everyone, it’s great to see you all joining on LinkedIn, welcome to another episode of The Paint & Pipette Podcast. I am thrilled to welcome to the stage with me, the legendary cofounder of WIRED Magazine, Kevin Kelly. Kevin, thanks for joining us today.
[0:01:47.8] KK: Oh, it’s a pleasure, privilege, honor, and I’m so excited for this conversation.
[0:01:52.4] JU: Yeah, likewise, likewise. Folks who are familiar with my obsessions have probably seen one of your bits of advice countless times in various forms from me and so I thought it would be fun to start here. One of the things that you say in your excellent book, Excellent Advice for Living, is, “It takes a multitude of bad ideas to get to a good idea.”
It’s a deeply held belief of mine but I’d love to hear the story of how you discovered that and how you keep reminding yourself of it.
[0:02:25.8] KK: I think I first observed this in other people, artists working, just hearing about their journey and the number of iterations that they would go through and I think I made it personal when I started to collectively work at places like WIRED, where we’d have story ideas and realizing that quantity was very important in creativity, that you needed lots of things and most of them were going to be thrown away, in order to get to the great things.
You could get to good pretty easily, pretty fast with a few things but to get to great, you really needed to have quantity and that was born out, even myself when I started, I have been photographing all my life and I’ve done some things that were published by Tash and other publishers, LIFE Magazine, which was to get to a great photograph, you just really have to take a lot of them.
I mean, you can’t say, “I’m just going to take one great picture” and so there was a famous experiment done with some art students and a book called Art and Fear. I don’t know if you're familiar with it but they were two teachers and the story that they tell was around ceramics but I think the actual experience they had was around photography but doesn’t really matter.
What it was is they told the students at the half of the semester mark, they said, “Okay, at the end of the semester, you’re going to be graded and there’s two ways that you can be graded and you have a choice. You can either submit one or two things and we’ll grade you on the quality of those or you can give us a bunch of ceramics and we’ll grade it on the weight.”
[0:04:14.3] JU: Okay.
[0:04:15.2] KK: “How much, like if you do 50 kilos or whatever it is, you’re going to get a grade on the quantity” and they said, “The curious thing was that in every single time that the best piece of the year was always being generated by the people who went for the quantity.”
[0:04:33.3] JU: Fascinating.
[0:04:34.2] KK: So there’s something about doing things on a regular basis and I think part of it comes from, and this is my experience as well, which is that you have to iterate your way through things and you have to throw these things away and the only way you can do that with ease is knowing that there’s more where it comes from.
If that process is so laborious and hard and difficult that throwing something away after you get there is devastating, then you’re not going to be able to do that. You have to make it kind of easy where its generation is easy and that’s where it comes into making of a habit, where you kind of do something every day or you do something daily or weekly knowing that most of what you produce is not going to be saved in that sort of sense.
So this idea that you have to make lots of bad ideas to make a good one and lots of bad versions of a chair to make the great one, and lots of bad versions of a movie to make a good one is the – or the great one, is that you were going to do something on a regular basis because that is something that loses up the juices but more importantly, it enables you to believe.
I can throw this one away because there’s a lot more. I know I know that. I’ve been doing this every day and it will just come up, that active actually getting into the moment will produce things and so I’m not afraid that I’m going to run out.
[0:05:57.0] JU: Yeah. So maybe to use a personal example, I was talking with my brother earlier this week who has got a construction, a roofing business in Texas, and we were talking about how to acquire customers and things like that and I notice this phenomenon just sitting around the campfire with my family that the tendency is to fixate on a good idea and I’ve observed this in boardrooms, it’s the same thing.
But sitting around, there’s something very personal and almost visceral about sitting around the campfire with my family, we’ve got a business – you know, extensively a business problem on our hands and everybody goes, “Huh” and what I know is nobody here wants to throw out a bad idea, right?
[0:06:36.4] KK: Right, right, right.
[0:06:37.5] JU: So the question that I found myself and knowing I was going to talk to you, I kind of found myself thinking, “How do you start to normalize that?” It’s one thing for you to know or for me to know a multitude of bad ideas is necessary and especially maybe in the context of family is weird because there’s no hierarchy there but like, how do you start to normalize that?
Because I found myself throwing out bad ideas and everybody’s like, “No” and even though I know bad ideas are important, I found myself kind of sulking like, “Well, never mind” you know?
[0:07:04.5] KK: Yeah, yeah.
[0:07:05.2] JU: What do you do there? How do you help a group of people recognize the importance of that and then actually hold that value?
[0:07:10.7] KK: That’s a really great question, is normalizing it because that’s what has to happen. I’ve worked in many scenarios and workshops where we tried to elicit visions of the future and the important thing we always say is that these are plural. They’re not a single prediction, these are plural, you have to make lots of them and so the way you normalize it, I think, is role modeling. It is to say and maybe even explicit if it’s a group in which there are no bad ideas.
You need to have a good idea. I mean, you need to have some bad ideas to make a good idea but even more importantly, I believe that sometimes the best solutions will come from a really bad idea and so one of the little bits of advice, I think it’s in the book, I don’t remember now, which is that if you make a list of – try to make a list of like all the bad, the baddest ideas, the worst ideas you could think of.
It's likely that to see, they have something that actually becomes the best idea will be embedded in some of the worst ideas, these kinds of things that are probably – and that’s because what you’re looking for is something improbable.
[0:08:21.2] JU: Right.
[0:08:22.4] KK: A really, really good idea is an improbable idea. It’s not going to be the first thing that comes to your mind. It’s going to be hard to get to it and actually, going in the other direction of trying to think of something that’s kind of easier for us and you just said, you have more permission to kind of like just make list of really – it’s easy to make bad ones and so you can kind of go further by trying to actually deliberately create like the worst idea that you could think of.
[0:08:51.0] JU: Yeah.
[0:08:51.7] KK: Because then at least, you’re kind of getting away from the norm. You’re getting away from – and that’s the hardest part in the imagination business and is leaving behind what is expected, what the convention is. So having – deliberately having bad ideas, the worst ideas helps you get away from that.
[0:09:13.7] JU: You’re reminding me of a couple of practices. One is, I know it’s Second City, I had the privilege of getting to know Kelly Lenard, who runs a good portion of Second City, the famed improv site in Chicago. It’s launched so many famous creators and one thing he said that is one of their favorite activities is they call them bad brainstorms.
And once a month, they have the bad brainstorm where they basically take all the ideas that they – everybody’s responsibilities are bringing the idea that they say, “We’d never do this.” Similarly, I think Steven Tyler has talked about once a week, they would have what they call a dare-to-suck meeting, where every member of the band has got to bring just a terrible idea.
And he said, “Most of the time, they’re terrible but every once in a while, you get a ‘Dude Looks Like a Lady,’ you know?”
[0:09:57.7] KK: Right, right, right.
[0:09:58.9] JU: And there’s something about that, even Nolan Bushnell knows talks about that at Atari, having bad idea brainstorms.
[0:10:04.7] KK: I just want to follow on there because Brian, you know, he has something he does with the bands that he’s trying to produce, which is to have them at least at some point play each other’s instruments that they can’t play to make bad music because, again, for the same reason that we’re talking about, it puts it out of the ordinary constraints you have, it pushes it way out and just that release from the expected can lead to other things.
[0:10:32.9] JU: Yeah. I think even setting that expectation that the goal is to vary from the norm. In variation, if you think, the meme behavior is here, sometimes the best way to get over here is actually to sling. You know, it’s like, I don’t like these space movies where you got to slingshot around, you got to use the gravitational pull. There’s something about using the gravity of the wrong direction to get in the right direction.
[0:10:56.7] KK: Right, right, right, exactly. Yeah.
[0:10:58.5] JU: Yeah, you had said about cultivating imagination or you said in one of the excellent pieces of advice that imagination is a muscle and you need to exercise it. It reminded me of something I read. Thomas Edison said, he said, “There are three qualities essential to an inventor” and one of them was cultivate imagination and the thing that struck me is I think especially in the circles where I run, with a lot of organization and leaders, imagination doesn’t seem like a cultivatable thing. It can almost seem binary, either you have it or you don’t.
[0:11:30.5] KK: Yeah, no, I don’t –
[0:11:32.0] JU: And I, of course, I know you don’t subscribe to that book and I don’t either. How would one, if you were going to advise someone to think about building the muscle of imagination, that idea brain storms perhaps being one example, what are some other ways to build the muscle of imagination?
[0:11:46.6] KK: Yeah, and there are different varieties of imagination. It’s not a single-dimensional or unidimensional thing. Like when I think of some of the most imaginatively creative people that I know, or the most imaginative people, their imaginative in different ways. So like have one friend who is very imaginative person.
He is a storyteller, a writer in science fiction and so his imagination abilities are often about being able to fill in all the details of an imaginary place in great detail and that’s like a superpower and then I have another friend who is really good at lateral thinking. This coming off from the side, and that’s a kind of an imagination.
[0:12:33.0] JU: Yeah.
[0:12:33.5] KK: And then there’s another friend who can immediately, with a hearty thinking, rattle off variations of things. You know, just instantly making long, long kind of combinatorial, exhaustive, searching the space of possibilities. Again, just going to enumerate them very, very quickly.
So there are different kinds of imagination that I would say and I actually haven’t seen the taxonomy on that but that would be really good if somebody would work on that and so those probably take different kinds of muscles but I would say the first thing for cultivating it is that, what you said, it is a habit. It is something that you get better at by doing a lot, right?
So that’s one thing and I would say the second thing is that like we were just saying, you have to – I talk about advice elsewhere in the book about divorcing the genesis mode of thinking from the critical mode, both of which are absolutely essential in making something great. So there are functions or rhythms and one of the rhythms is that in the genesis period, you have to protect that from judgment.
You have to be utterly nonjudgmental and not even talk about bad ideas or good ideas, they’re just ideas. In that initial moment, the initial cycle, you are producing things without judgment and you’re trying to, what I would say to surprise yourself or to surprise the world or whatever it is, you’re really trying to generate something new and you have to protect that and that’s in the improv world, you were mentioning it, they are very, very, very strict in that one cardinal rule, which is that you can never say no.
[0:14:21.7] JU: No blocking.
[0:14:22.6] KK: Hold hand, right?
[0:14:23.6] JU: Right, right.
[0:14:25.0] KK: You build on, you don’t ever deny something and so in that same spirit, that’s part of the training to be imaginative is protecting that initial genesis of not saying no, of not being judgmental, of not actually trying to be critical about it. You have to do that, it’s the next step but you can go back and forth but you want to be nonjudgmental. So that’s hugely important.
[0:14:53.8] JU: Can you talk about, for a second, about space for practice and how we think about that? I was actually speaking with a leader of a large hospital system. You know, 20,000 employees in this hospital, and he and I were talking about that, this idea of practice. How have you seen it effective to actually operationalize practice?
I realize that’s like a – that’s a very official sounding thing but it’s one thing to say in abstract, you should have times where you don’t judge, right? But in terms of implementing that in the context of say, a team or an organization, are there any practices that you have seen successful and actually carving out the space for practice?
[0:15:36.3] KK: Yeah, you kind of immediately constrained the thing to talk about teams because a lot of creativity is one person or maybe even a partnership. People who are writing, you are trying to write a letter, you’re trying to work on a logo, whatever it is and so that comes down to, you know, your kind of personal habits and your whatnot.
Doing it as a team is another ability and what they have recently kind of found is that brainstorming is not necessarily the most efficient way to generate creativity. They have done testing in terms of people evaluating the novelty, whatever it is, and founding that just having a bunch of people throwing up ideas is not necessarily the most efficient or productive way.
That actually having individuals work on their ideas alone at first and then bringing them together, actually produced more valuable outcomes. So that’s maybe part of the practice of working but the one thing I am sure about and I think it would apply to this is this has to be an iterative process.
It’s you do it and then you do it again and then you do it again and then you do it again and it’s not something you do once. It’s something –
[0:16:51.0] JU: It’s not like a blue pill, like, “I know kung-fu. I know creativity, I took the pill.”
[0:16:56.0] KK: Exactly. It’s built into the process. I really like the way, I mean, the way that Hollywood makes movies these days is that it’s a very collaborative thing. There are a lot of people but they kind of prototyping and making it in an iterative way the entire way. You know, they start off with scripts, which go round and round and round, with many layers, many people involved, with trying to be creative and imaginative.
Maybe working at home in the passing month for feedback and then moving into where they make storyboards, which is another level of things and another kind of imagination required for that, and then they make animatronics and things where they actually make versions of the film based on like the storyboards and then they make a version of it where they have scratch sound and voices.
So by the time that they’ve come up to the end, they’ve made the movie in many different ways and thrown out many versions of the movie along the way.
[0:17:56.8] JU: Right.
[0:17:58.2] KK: And that idea of again and again, round and round coming, having the genesis, protecting it, having the critics come back and say, “This is terrible, we’re going throw this all away” and then going again, that process is the only way you get to the great things.
[0:18:11.7] JU: You know, it reminds me what you're describing, it reminds me of Ed Catmull’s Creativity Inc. and describing the process at Pixar. I love the way he says, he says, “Our job is to take movies from suck to not suck.”
[0:18:25.3] KK: Right, right.
[0:18:25.8] JU: And we can think something like Finding Nemo, it’s like the Immaculate Conception, right? It’s like, “This has always been an amazing — ” rather than like the thought that there was a point in time in which Finding Nemo, to use his words, “Sucked” is hard for us to do, right? As audience members, like it’s always been amazing.
[0:18:41.5] KK: Well, even their first one, Toy Story, this sucked so bad that they were like going to completely abandon it because Woody was not the Woody he was now. He was a mean guy. People hated him. I mean, the people, the viewers who saw it, they hated Woody and so they had done some level of making the movie this way and it just wasn’t working and that’s something else, by the way, that’s Pixar and other places but WIRED, it was my own experience.
WIRED was a monthly magazine and the thing about it was, you would think, after seven years, whatever it was of doing this every month that it will become like a formula and you would come to Friday night, right before deadline, everybody would go home. That never happened. Every single time, it was a near-death experience, it was just a miracle. The issue was put out on time and it’s true of every Pixar movie, every movie.
If you hear the full story, it’s a miracle that it ever was finished, let alone, great and it’s because basically, it’s because we were always upping the game. We were always trying to make it a little better and you’re just always pressed to the wire. There’s always some drama about something that doesn’t work. It’s the real world, it’s real life. Having done 10 Oscar-winning movies, the next one will still be the same thing. It will be a total miracle that it ever gets done and that is good.
[0:20:14.2] JU: Right. You know, one of my favorite examples of that is, I don’t know if you’ve seen the comedian, the documentary about Jerry Seinfeld. You know, you think if anybody could just go on Letterman tomorrow, it’s Seinfeld, right? Because – especially after the height of his success but seeing – no, it takes – you got to get in the comedy club night after night and bomb night after night and develop the material.
But that’s so – it’s so counter to – I think a lot of us could see somebody like, see Finding Nemo and think, “Those Pixar guys are geniuses” or see Seinfeld on Letterman and go, “Well, it’s just because he’s funny” and we don’t appreciate all of the iteration that goes into him looking like a genius in those 15 minutes, right? One thing you speak about several times in the book actually is deadlines and the power of deadlines.
[0:20:59.7] KK: Yeah.
[0:21:00.1] JU: And I know that’s true. You’re saying it’s like a near-death experience, it’s hysterical every time, and at the same time, we also just talked about the good idea comes relatively early but if you want to go for the great idea, you’ve got to kind of persist or endure an ideation longer.
[0:21:16.8] KK: Yeah
[0:21:17.1] JU: How does one balance the value of a deadline with the somewhat – it’s also arbitrary, like is it possible that a better idea is beyond the deadline? It may be possible. How do you balance those two forces?
[0:21:28.5] KK: Yeah, so the deadline, the chief attribute of a deadline is that it takes something that’s abstract and perfect and it makes it real, which means it has to become imperfect and so you’re going to have to shift something that’s imperfect but that means it’s real and so that’s good.
The deadline and the thing about it in order to get there, you have to give up perfection and you have to be different, different than what you ever thought you were going to be. That’s again if you go to the story of anything well done and creative and great, it didn’t begin that way. There were some way that was made different in order to ship it and so there is an art.
A balance between wanting the very, very best you could be and the fact that we’re limited in our time and that’s the only scarcity there is in the entire universe as far as I can tell, is our own time and having a billion dollars doesn’t give you any more of it. So it’s the thing that there’s a bit of a wisdom.
I think it’s – I put it in a book that came from a book called A Writer’s Time, which says that basically, any kind of worthy project, whether it’s a film or a book or a podcast series, whatever it is, the work involved in making it great is infinite, it’s bottomless, there is no end.
[0:22:44.5] JU: Right.
[0:22:45.7] KK: And so you can’t really manage the work because it’s infinite. You can only manage the time you have for it. It’s like, “No, I’m going to do the very best thing possible to do in six weeks.”
[0:22:59.1] JU: I found that and I’d be curious actually to hear your experience of this book. I only have one book, you have many. The thing that was astounding to me is, a book is in some sense, it’s like a snapshot of your knowledge at a point in time.
[0:23:10.2] KK: Yup
[0:23:10.5] JU: And I’ve noticed like people ask me to talk about me but if I do a keynote or something.
[0:23:14.1] KK: Yeah.
[0:23:14.5] JU: I noticed the other day, I tell basically three key stories in my standard keynote now.
[0:23:19.2] KK: Right, right.
[0:23:19.7] JU: Two of them are not in the book because I learned them after it came out.
[0:23:24.4] KK: Sure, sure, sure.
[0:23:25.1] JU: It’s astounding to me. I mean, there is actually research I think is seminal and central. The research, I actually looked in the index of my own book, I didn’t even reference the guy. It’s like you since with – you underestimate how much one continues learning, be it the editors like, “Jeremy, we cannot put another story in this book. It doesn’t matter that it’s the perfect illustration of…”
[0:23:47.3] KK: Yeah, yeah, yeah, exactly.
[0:23:49.3] JU: Are there things that you feel, you know you hit publish so to speak, I realized –
[0:23:54.1] KK: Yeah, yeah.
[0:23:54.8] JU: Is there stuff that you go, “If I could immediately put out the second edition today, I can’t believe I didn’t say this.” Are there things that have –
[0:24:03.5] KK: I have a list of about a hundred more bits of advice that I’ll probably try to squeeze into the paperback. Yeah, I joke that the best time to write a book is a year after you’ve written a book and given a talk for a year and then you know it’s about and that is the best time to write the book is when you’ve figured out what it’s about because even by the time we finished the book, you still don’t really know what the book is about.
But you could give talk after talk after talk about it and your people’s reactions will get a much better idea and that’s the time you really want to write it.
[0:24:37.4] JU: We can’t tell the publishing industry this, it would send them in a death spiral. So one of the things you mentioned is the daily practice.
[0:24:45.2] KK: Yeah.
[0:24:45.9] JU: I know that you’ve been making a piece of art every day for a long time and lately, within the last year it sounds like with generative AI. Before you get into the specifics of it, can you talk about why that is a touchstone of your life and in terms of habits that you have maintained?
[0:25:01.3] KK: I always liked drawing as a kid, I have always been painting and drawing, and later on, photography took up some of that visual creation aspect and I was photographing constantly. Of course, even things like WIRED was a nice convergence to me because it was very visual and so not only was it the ideas and the text bit was a very visual thing, which I really, really embraced.
Doing it one a day came about because I discovered a tool that I’d always wished I had and that was Procreate on the iPad. So the thing, the one challenge for me as a visual artist painting and drawing was control of color. Color mixing is an incredibly fine high-art skill.
[0:25:52.4] JU: Yeah, it’s art in of itself, yeah.
[0:25:54.0] KK: That people don’t realize how amazing it is, somebody can do it well. So there’s like this, you can say, “I want to make this particular color orange exactly like that” and you have whatever it is four colors you’re starting with, it’s really, really hard to do and to do consistently and again and again and so with the computer, the iPad and the Procreate, I could get those colors without having to mix them.
I could find those colors, I could slide through and suddenly, I had this sort of ability that I didn’t have before and that’s kind of spurred me to want to try and do something every day as a means of looking. So the thing about photography was, for me, photography was a very compelling excuse to get out of the house and look, see things, and painting something every day was forcing me to look to try to do anything realistic.
The only way you can paint a realistic thing is to really look at what it is. Look at a mountaintop and see how that snow fits in there in order to be able to paint it. The same thing with tree bark, you have to kind of really, really study it in some ways. So drawing and painting every day was an exercise a lot in trying to see.
[0:27:20.4] JU: To ask the obvious question, sorry to interrupt.
[0:27:22.6] KK: Yeah.
[0:27:23.0] JU: Why is seeing important?
[0:27:24.6] KK: Seeing, so I talk about like I love the thing about talking about the future is really what I call predicting the present. It’s really like can you see what’s actually going on right now?
[0:27:36.5] JU: Yeah.
[0:27:37.0] KK: That is like if we can see, we can understand what’s actually going on right now, that’s half of the way to understand what might be happening tomorrow or the next day and this necessity to see what’s going on and I don’t mean just even visually see. I mean like to see, to understand, to reckon with, to comprehend, to absorb. So I think observation is a key skill for anything that we’re doing.
The best writers are incredible observationalists, they’re able to write this because they’re observing things in a way that’s almost extraordinary and so I find that painting and photography helps me become more observant.
[0:28:21.7] JU: So you shifted it sounds like from Procreate to what’s your tool of choice now?
[0:28:28.9] KK: So I’m using the AIs, Midjourney, which has just released a new version of DALL-E Stable Diffusion but even Photoshop now has built-in AI they called generative fill, which is amazing and I haven’t released some stuff but I am fooling around with my photography to try and do something. I think of these things as code generation, it’s not just the AI, it’s not just me, it’s a partnership.
It’s teamwork, sometimes I am feeding into the AI artwork that I’ve made as a seed. So I am really – it’s really a collaboration in many ways.
[0:29:11.5] JU: Can you talk about the collaborative dynamic?
[0:29:13.6] KK: Yeah, yeah, yeah. So when I was doing my own painting every day, I had one goal because I would sit down and I would have usually zero idea about what I was going to paint and I had only one goal, which was I wanted to surprise myself. It’s like I have no idea but I want to do something that I had no idea that I –
[0:29:35.2] JU: That was in, yeah, kind of in the moment.
[0:29:37.4] KK: Yeah, I did that well. So where did that come from? So it’s like you know I say, “I have no idea what this means, I have no idea where it comes from but here it is” and so AI is the same kind of surprise but the collaboration is really striking because it’s very easy to get these current versions of AI to produce something that may surprise you. It’s very, very hard to get them to do something that obeys you.
So like yeah, they’ll make something, it won’t be great but it will be kind of interesting, and then to try and move it in a certain direction is really difficult. There is a constant nudging, it’s like trying to move a mule or a donkey. It’s like there’s some – it doesn’t really kind of want to go in your direction and you have to use all these tricks and skills and prompts and whispering to try and move it in certain directions and the epiphany that I had recently is that there are lots of things I was trying to get it to do.
I just couldn’t get it to make this kind of art and I realized the reason was is because this was art that transcended language and the Large Language Models (LLMs), don’t know where you get it. It is using language and if you can’t – if it can’t be expressed in words, you can’t get the AI to make it.
[0:30:54.0] JU: Right.
[0:30:54.5] KK: So there are limits right now, there are limits to it. They can’t do language-transcendent art, it can only do seem to produce art that’s tethered to language.
[0:31:05.9] JU: Well, you talk about the collaboration and the interplay. How many back and forths with the AI to make a single piece of art? The way I kind of think about it is almost like photography. It’s like if you go on vacation I just had, recently I had a family vacation, how many photos do I favorite? Maybe three, right? How many did I take? 300, okay, so I got 300, I got a hundred to one ratio.
Call it input to output so to speak. You know if you think about your collaboration, what’s the interplay ratio like? Like to get a photo you’re satisfied with or an image you’re satisfied with, how many nudges to the mule are you giving?
[0:31:42.0] KK: There is certainly at least 30 generations I would say, 30 back and forths of, “This thing here, not this” going back and forth and it can be hours. It can be hours to get something that’s worth sharing let’s say. So it is like photography and that you’re kind of stalking through a landscape and you have these hunches that there is something promising here and you’re going to kind of be patient.
And you’re going to be kind of explore this little corner of interestingness and you’re stalking through it and you’re taking little steps and you are moving through and occasionally, you’re kind of walking over the hill to see what’s there. So it is in many ways similar to the – what’s the word I want? The topography of doing photography as a concept, where you are hunting for things.
[0:32:31.4] JU: Right. You are helping me kind of crystalize this is really cool and it helps explain something. I am undertaking a modest research study right now about generative AI and its impact on collaboration and team problem-solving. I have survived 14 years at Stanford having never conducted proper research but somehow, I got wrote, and a friend from Harvard said, “Hey, let’s do this together.” So I have been working on it.
What we are doing is basically doing kind of control group and research group side-by-side studies. You could say, “Okay” at the Long Now Foundation you go, “Hey, a problem we’re trying to solve right now is how do we get more people asking long-term questions.” I don’t know, whatever it could be, right? Well, what we do is we take a control group and we have them, we give them a scaffold for a problem-solving activity and we’re measuring kind of quantity of output.
We have the problem owner measure the quality of the ideas, you know grade them A, B, C, D, whatever, and then we have a research group. We give them the same structure, they get the same grading rubric and everything but they get access to generative AI as kind of an ideation partner and then we are comparing the quantity and quality and the feeling, “How does it feel to be a part of a problem-solving group?”
They are doing this Long Now problem, what’s your attitude toward collaboration, towards ideation, and all these things before and after and they were comparing that with the group that has the same structure, same team kind of composition but then also has access to AI and we’ve been finding some interesting things. One of them is, granted it’s small dataset at this point but one of them is the AI groups actually are generating less volume of material despite the fact that it has infinite capacity to generate more, so that’s interesting.
There’s also AI is generating less D-caliber work and less A-caliber work. It produces a lot of Bs and Cs and not a lot of As and Ds, and as you’re describing the stock, the topography, what I’m realizing is I think a human problem solver is kind of looking for a good enough solution and when you think about photography, somebody really wants a spectacular photo, they’re going to stay up late, they’re going trek in another mile deeper, right?
[0:34:32.5] KK: Yeah, right.
[0:34:32.7] JU: All these stuff and there’s not maybe by virtue of the actual structure of the task, we aren’t giving people time for stalking. I’m not sure they’re interested in it by the way because it seems like, “Wow, we got five pages of documentation on a B-plus idea in like five minutes, we’re done” right? So I don’t know if we gave them a hundred more hours they would spend but I think they’d probably just refine that, right? I don’t know.
[0:34:55.4] KK: Yeah, no. So these are all trained on the worst and best of human creation. So they’re basically will produce the average, they’re basically tweaked to produce plausible human responses or creations. That’s by the nature of these Large Language Models and these transformer models, they’re going to produce the average and what a lot of the nudging happens, and in particular I’d like to say ChatGPT, is nudging them not to be average.
It’s like, “Okay, you’ve got to pretend you’re an expert. Don’t give me the average thing, then you’re an expert. Try here.” You’re constantly the default for all these AIs is to produce plausible human average human generation and you have to really push them and work hard to get them out of the average.
[0:35:44.6] JU: You know what’s funny is going back to the beginning of our conversation, we need them to do a bad brainstorm. We an AI to do a bad brainstorm and then give AI the challenge to flip the bad brainstorm into good stuff.
[0:35:54.4] KK: Right, right, and so there are people who are much better at this than others, maybe by a factor of a hundred, and part of that is these AI whisperers is because they spent the 10,000 hours or the thousands of hours doing and they have some idea of what works and what doesn’t work and just yesterday, Midjourney released a version, a new version that has this really kind of cool thing called shorter prompt.
What it does is, it actually gives you the weights on the prompts that you have, the weights of like how important that prompt is or not and it turns out that a lot of the words that people are putting in did nothing, really.
[0:36:39.7] JU: They don’t advance the prompt though.
[0:36:41.5] KK: They don’t change anything.
[0:36:42.5] JU: Wow.
[0:36:43.3] KK: And so a lot of the really good prompters knew that from just experience.
[0:36:50.2] JU: Right.
[0:36:50.9] KK: They would kind of converge on those kinds of words that they knew would actually make a change versus just things that other people were using like putting in high rez, Nikon, you know whatever, all these stuff, which made no difference whatsoever.
[0:37:05.5] JU: Right, right.
[0:37:06.4] KK: So part of what we’re discovering again in the masters is that some people are much better at getting them to produce great stuff because they’ve put in a thousand at it and they’ve come to realize or see or understand what works and what doesn’t work, what moves and what can get them nudged out of the ordinary into something great, so it’s like any other tool.
[0:37:29.7] JU: Okay, I have one more question and another piece of excellent advice before we wrap because I am dying to know. You say ‘you need to make, explain the problem a part of your process.’
[0:37:41.2] KK: Yeah.
[0:37:41.7] JU: Can you just give us two clicks more on that?
[0:37:44.5] KK: That’s actually went up on both, some studies, and then I think there is a book that’s going to the details and also my own experience, which is that when I get stuck on something, I have an assistant researcher, begin to explain the problem and then in the middle of explaining it, it was like, “Oh, wait, wait, wait, I get it.” Just having to explain step by step what it is there’s something magical about that that will provide the answer.
So people who teach troubleshooting make that part of the process of troubleshooting that when something is not working and you can’t figure it out and you try things, you have to explain it to somebody what the problem is in a very step by way – step process of explaining what’s not working and it’s amazing how often the answer will come from yourself as you’re explaining it and I don’t know why that works, other than it works.
[0:38:44.0] JU: No, it’s great. When you say make that a part of your process.
[0:38:48.1] KK: Yeah.
[0:38:48.7] JU: I wonder, are there any other steps that you would encourage people to make a part of their process so to speak?
[0:38:55.0] KK: I mean, with troubleshooting?
[0:38:57.2] JU: Yeah.
[0:38:58.2] KK: Part of the process and that’s like part of the creative process? Well, it’s a good question, so I’ll just talk about my own process. So when I am embarking on something new, not necessarily troubleshooting but new process, I’d like to get like a survey of like kind of what’s out there, what’s a picture of what I call the landscape. I just like to know a bit about the landscape and in particular, I’m interested if somebody’s already done what I’ve wanted to do or whether it exists or something similar.
So that’s to me a part of the process, a part of the process is the snapshot, the landscape, and then part of the process also I find is to articulate a little bit about what success would look like. I do –
[0:39:42.0] JU: Almost your outcome variable. You are talking about experimentation language, it’s like, “What are we measuring at the end of this thing?”
[0:39:49.0] KK: Well, like – okay, like a book because I do a lot of talking with people writing books and I always start with the end. It’s like, “Okay, let’s say this book was successful. How would you define success? What does it look like to you? What it will accomplish for you, what does that?” If it was a raging success and I am not talking about bestseller but maybe that’s part of it, what does that is?
I mean, what does it look like and feel like and do? So whatever project it is if it is any significance, I don’t mean like sitting down to do a daily drawing but if there’s a longer bigger thing, it’s like, “Well, what, how do we know whether this is successful or not or how will I know what would it look like?” So that’s starting at the end is another part of the process.
[0:40:32.3] JU: No, it’s useful, right? I mean, again, I was talking to a leader just earlier today who was saying, “If you don’t define success ahead of time, then your capacity to rationalize, you’ll declare anything a success later.”
[0:40:43.1] KK: Right.
[0:40:43.6] JU: Right?
[0:40:44.0] KK: Or find too, you can do that. That’s a Brian Eno thing, is you paint the arrow, and then you paint the target around the arrow, you know? I think this idea of inventing your own version of success is very, very important but you should know what it is that you’re doing. If you are happy with declaring anything then success is fine but I don’t think success has to be someone else’s metric.
You can make it up and declare successful but you do want to have some sense of that before you start is all I’m saying.
[0:41:15.9] JU: And so maybe just to bring things full circle, as it regards to the book, what do you see as success, and if folks who are listening and I see a bunch of folks who’ve been dropping comments in the chat, thank you, Jonathan and Francis and Michelle and the whole and others, folks who are engaged, who are here now, who are going to be listening, what do you hope they do with this knowledge or how can they support your definition of success for this book?
[0:41:41.4] KK: Yeah, I generate things primarily with the audience of one. WIRED was invented, created primarily to be the magazine that we the founders wanted to read because it didn’t then exist. It’s all we were doing, we were making a magazine that we wanted to read. We were the primary readers. When I gave assignments to the writers I said, “Your audience is very, very simple. It’s not some 11th grader or somewhere, it’s me.”
[0:42:10.4] JU: Yeah.
[0:42:11.1] KK: “You’ve got to amaze me because I’m bored and I’ve read a lot” and so when I did the photo book, the audience was me. I wanted to make one copy. If I had one copy of my Vanishing Asia, that’s all I needed but since I’m making one, it’s printing, I’ll just make a bunch of others but if other people wanted to join me, it’s fine. This book of advice was written basically for my kids, for me and the kids, and if other people – so it’s already successful.
[0:42:41.0] JU: Is there anything listeners could do to influence your children to really take it seriously?
[0:42:45.7] KK: So I’m a big believer in having this audience of one and really trying to create things from a personal expression, from a personal drive, from making it autocratic and you want to protect it as much as possible from the bureaucracies, from thought police, and everything else going on. You want it as much as possible and make it true to yourself. And so I think for me at this point, the book is out and if other people find it useful, I am delighted by it.
I know, I’ve heard many people say, “I have kids, my kids don’t listen to me. I try to give them advice, they’re bored by that but they’ll listen to someone else. So I give them your book and they’re paying attention.” Okay, that’s great, that works, that works for me. So I’m happy when that happens.
[0:43:36.4] JU: The one thing I might suggest if I could offer one piece of guidance for folks who are listening is don’t try to read it in one sitting. There’s always like this sense of accomplishment in our minds. “You know, I want to get through this book.” It’s more important that this book get through you than that you get through the book, right? So to me, I would – and Seth Godin and I were talking about this when he was on the show a few weeks ago but it’s that kind of book, it’s limit yourself to like, two pages a day.
[0:44:00.7] KK: Yeah, yeah.
[0:44:01.5] JU: Just because what you want to do is find opportunities to apply these truths because the proof and the value is actually in the application.
[0:44:09.3] KK: Right.
[0:44:10.3] JU: It’s really easy to skim through and go, “Oh yeah, that was a good book” but if you want to get the most value out of it, by the way, just the other day I was doing a talk at Instacart in San Francisco with a bunch of people and they’re super loud and I started going, “May I have your attention please?”
[0:44:23.2] KK: Whisper.
[0:44:23.9] JU: Because of your line to get the attention of a crowd or a drunk, just whisper and I told them, I actually gave you credit. You know, there is a hundred people in there and I said, “You know why I was whispering? Because Kevin Kelly said that to ask a crowd to be quiet” right? If I had read that amidst a hundred other pieces of advice, I mean, I’d have remembered it because it happened being called my daily reading, it actually had the opportunity to get them.
[0:44:46.2] KK: Well, thank you, and then for the benefit of those who may not know, the book has only aphorisms and proverbs, that’s the entire book. It’s 450 of these little adages and lessons and actually, the original subtitle was called Seeds of Contemplation. The idea was that you would take these little seeds and kind of unpack them yourself, which is what my hope would be, and as you said, they range from the cosmic to the very practical.
So, I’m really delighted that other people find it that way. One cowl – you said it was like the Bible without stories and I took it as a compliment because I’m not a very good storyteller. So I went to my strengths, which is making aphorisms and little telegraphic tweets and so that’s what they are. They’re little zip files that you can unpack at your leisure.
[0:45:43.0] JU: Yeah, and I would say, take it slowly, you’ll find, I found lots of wisdom there myself. I thank you for putting it in the world. I still remember the day that you dropped the 68 bits of advice, which I think was kind of the starting point. I devoured that, we shared it around the d.school community then, and now to see it, it much expanded in book form and it is a real gift to us into the world. So thank you for putting it out there.
[0:46:04.4] KK: You're very, very welcome. Thank you for having me.
[0:46:06.4] JU: Yes, thank you for joining us today, and folks, thanks for listening, and until next time, have a great weekend, see you all soon.
[0:46:14.9] KK: See you.
[END]