Kevin Kelly on AI, Art, and Staying Sane

An exclusive conversation for Sublime Premium members

Happy Saturday everyone.

This conversation with the one and only Kevin Kelly is the final bonus interview for Whoa, Vol II: Conversations on AI x Creativity.

Kevin Kelly is the founder of Wired and my personal role model for what a well-adjusted, intellectually, spiritually, and creatively nourishing life looks like. Brie Wolfson wrote an exceptional profile of him, which is worth reading in full here.

In this conversation, Kevin shares:

• How he’s using AI in his creative process (what it’s good at and what it’s not)

• Whether 1,000 True Fans still works for creators today

• What he thinks people will pay for when AI can make “good enough” anything

• How he would build Wired if he were starting today

• The platform that is becoming the center of culture (and that most creators are underutilizing)

• His new book Colors of Asia, which he describes as “cool and useless”

• And what he’s most excited about right now

This conversation is an exclusive perk for Sublime Premium+ and Lifetime members (and paid newsletter subs, if you prefer our relationship be mediated by Substack). To read the full conversation, become a paying subscriber.

Alex Dobrenko: There’s a lot of conversation right now around AI. And what we’re not seeing as much is, how are people actually using it? Or what are the concerns people have about it taking away their creative livelihood? Let’s start there.

Kevin Kelly: I use it a fair amount and I’m not very worried about it. I will say that the more powerful the technologies are, the more powerful the problems it will create. So I am expecting AI to be one of the most problematic technologies we’ve ever made while being the most powerful one that we’ve ever made.

And my formula is that if we can create just 2% more than we destroy every year, that 2% compounded over time is what progress is. So there will be a very, very long and big list of all the problems with AI, but the list of what’s good is a little bit longer, not by much. So I’m not going to argue that there’s no problems, there’s plenty of problems, but as long as the solutions and the benefits outweigh it by 2%, which is not very much, it’s hardly even visible.

And the other thing I would say is there’s something I call thinkism, which is this idea that we can solve things by thinking about them or that we can even understand things by thinking about them, that what the world needs is more intelligence. We’re not going to solve or even identify the right problems by thinking about them. We only figure it out through use.

So your question of how people actually use it is the most important question because that’s how we’re going to figure out what’s good and bad. Not by thinking about it, not by imagining what could happen. That’s important, but it doesn’t go far enough. It’s just the first step and we’re now at the second step. And the second step is that you have to actually look at how we use them and what the actual harms are versus the imaginary harms. The actual harms are pretty minimal so far.

Very few people have lost their job to AI. Very, very, very few people. There might be a lot of people who may not be hired because of AI. That’s a second thing. But very few people who’ve been fired because of it. I think it’s really interesting that if you ask people about their own jobs, whether their own jobs could be replaced by AI, nobody thinks their job can be replaced by it. But my friend of a friend... their job is going to be replaced. It’s like, no, it doesn’t work that way. So, anyway, I can’t say I have no worries about, there’s lots of things to be concerned about, but there’s more to be excited by.

AD: What excites you about it lately?

KK: Well, let’s talk about what we have right now versus what we could imagine. It’s very clear that an AI that reads everything is really, really good at answering questions. I’ve been trying to get it into the habit of asking it every question I have. It’s stunning, its ability to kind of answer any question.

But, how much does that change my life? Well, I mean, that’s been a lifelong dream to have the answer machine, where you can ask anything. So I would say, there’s a 7% satisfaction improvement in my life. As someone who likes to learn, that’s a 7% increase in satisfaction. So that’s one thing. It’s like Google plus in terms of doing the daily things that we do better and sometimes things that I couldn’t even do before. So I don’t know how to rate that improvement. It’s not 10x, but it’s better.

There’s that part of my life, using it like a research intern, and it’s amazing. When I left Wired, I had a little tiny bit of money that I got out of the sale of Wired. And I did not buy a boat. I did not buy a house. What I used my money for was I hired a full-time researcher 30 years ago.

And I’ve had a full-time researcher all this time. Research has always been a really huge, important thing to me. The research capability of AI is really great. However, I still have my researcher and they’re using AI. And so they have been 5x’d. I think their productivity is about 5x because now my researchers are using AI to do research. So that’s good. That’s really good.

In terms of creative stuff, I did a one-year private analog art project, once a day. And then I did a year of AI art with mostly Midjourney. And that was fun. And I learned a lot about some of the limits and the capabilities.

And recently I’ve been trying to take a graphic novel that we had made, a really huge one, and turn it into a movie by feeding the panels into Sora and Vio. That’s been very frustrating because it’s just not quite there yet to actually do something that’s releasable.

I use generative AI a lot. Whenever I do slides, I generate the images for the slides using AI. When I do a post, I generate the images. So these are things that are accelerating or they’re increasing the quality of what I had been doing rather than enabling something new. Generating the movie from the panels, that is new. But that’s been very frustrating. I haven’t gotten very far with that. It’s just not quite there yet in terms of total control that I need.

I am working on a project right now where I did a bunch of generating still images and I am going to be producing three-dimensional artifacts from them. I was struggling for a while with that workflow, and I just finally had someone help me and I’m going to start. I’m going to be printing them out initially just on my 3D printer, but finally I’m going to be using a color, I’ll send them out for a full color, three-dimensional thing.

So I’m doing some art where I produced the final thing as physical analog art that comes through the AI generation workflow. But they’re going to be physical artifacts. And I’m really excited by that. And so I’m about halfway in kind of proving that and getting the workflow done. And it looks like I can do it. I think the tools are now close enough that I can probably make that pretty automated workflow. And so I’m going to be producing these analog physical objects rather than just little pictures. I’m really excited because it’s my year of analog art. I’ve kind of decided I’m going to do as much analog art as possible.

I did a whole saga of a series of 10 books, a counterfactual history that was really cool that I do not intend to show anybody because I realized that the joy of co-writing it and co-creating it was more joy than reading it. I don’t think anybody would enjoy it as much as I enjoyed co-creating it. So I wrote this piece called “The Audience of One,” which is this idea that I think mere co-creation will fuel a lot of the use of AI in creativity, where you like doodling or sketching, diarying or taking a walk. It’s the pleasure of that moment of doing it and creating it. That is the main benefit.

AD: Which is so interesting to me because I write on Substack to an audience, and I don’t know how to write without an audience, you know?

KK: Yeah. Well, that’s interesting. And I understand that, but there may be other things that you might want to do other than write where maybe you’re going to sketch or you’re going to design things or you’re going to make other things where you don’t need the audience.

AD: Totally. I think this veers a little bit into something else I wanted to talk to you about, which is the way that for people in my generation, and it seems like younger too, there’s this sense of: anything I do must be shared and monetized, because that’s sort of my way to my thousand true fans, right? All roads kind of lead there somewhat. And I’m curious to talk to you about your One Thousand True Fans thesis. How do you feel about that almost 20 years later? Do you still believe in that? I think a lot of people want to know from you, what is the magic rule now for success?

KK: I think the mathematics, the arithmetic around Thousand True Fans still works. It’s something that I said pretty early on, I think in the second version of the piece, which I haven’t changed my mind about, and that is it’s not for everybody.

It requires a set of skills of managing, overseeing, interacting with people that some creatives don’t have and don’t want. So it is nearly a full-time job tending your fans. And not only is it even for the people who might work for it, it may only work for a certain period of time. It’s a great place to start. Doesn’t mean you have to end up there, but it’s a great place to start. And many people might start there and then move on to something that’s more mediated where they have other people involved or they have labels and studios or publishers.

So for me, the important thing is that it’s an option, an alternative to the only route that was before, which was a bigger and bigger audience. And I’m saying, bigger is fine. But there’s another alternative. And it does cost something in terms of the energy you need to put into the fans, but it’s another possibility. So I still think it’s a valid alternative way and has been proven now many, many, many, many times, particularly on the YouTube dimension. So the math still works.

The idea of the Thousand True Fans for those who are listening and not understanding is that if you are a creator of any type, using technology you can have a direct contact relationship with your audience, meaning that you have their emails or phone numbers and they can contact you. If you can have direct contact, so they pay you directly, then you don’t need millions of fans to be supported. You could have as few as a thousand true fans, true fans being someone who will buy whatever you produce. They’re the true super fan who is going to go all in with you. And if you get a thousand of them and you could generate a hundred dollars a year from each of them, then a hundred thousand dollars, you can make a living. You’re not going to make a fortune. You can make a living. If you’re a duet, a duel, you have to multiply it by more numbers to make it work.

That’s the mathematics of it. But it also means that the most harebrained, craziest, weirdest idea in the world, the most esoteric passion that you have is probably with eight billion people, probably shared by at least a thousand people, if it was a one in a million. The difficulty is in making that connection. Well, how do they hear about you? How do you find them? And that’s the technology that we don’t have yet. The AI might be able to solve it. We might be able to use AI to make this matchmaking function that looks around everybody and says, you two are going to get along. This guy and you are talking about exactly the same thing. Have the same interest. Here, here are your thousand true fans.

I think it’s possible that we could imagine AI getting to the point of being able to do that kind of matchmaking. And that will again be another lift for the Thousand True Fans enabling the one in a million idea that you have with the thousand true fans that you need to make it work.

AD: What if, I think I feel this way, you have a lot of fans, but they don’t want to pay for your stuff? Are they not true fans?

KK: They’re not true fans. You have the true fans and you have the avid fans and you have casual fans. And by the way, the true fans become the best marketers for your other fans. The whole thing about the thousand true fans is that’s not all your fans you have. You have lots of other fans. But yeah, you need the true fans. That’s about whether you’re useful, beautiful, true. Just because you’re creative doesn’t mean you’re going to have a thousand true fans. You have to be used, you have to be wanted, all those things. If you are intending to make a living out of it, then you have to be appealing to other people. That’s very, very simple. This is not like UBI where we’re going to support you no matter what. There’s a marketplace here. And if you want to be supported, you have to be producing things that people find valuable enough to pay for.

There’s a separate challenge of how you balance that trade-off of serving your fans versus your own inclinations. And that’s the classic Bob Dylan going electric. They want to hear the same old thing, which is what they want to buy, he wants to go in a different direction. And you kind of have to educate them. That’s a whole other challenge. But the point here is that if you can solve that challenge, it’s much more doable than having to have a million people pay you. So it doesn’t change the challenge of getting people to pay you. It’s just saying, you don’t need a million people willing to open up their pockets.

AD: Do you feel like you wrestle with that challenge? Or do you feel like people are kind, you’ve got your thousand and you’re good?

KK: I’ve been lucky in the sense that oftentimes when I’m interested in something, it turns out that other people are as well and are willing to pay enough to do it. But I’m not reliant only on the thousand true fans. That’s what I’m trying to say. I’ve had publishers and I’ve had a magazine. And so, while I have a thousand true fans, I’m not reliant on them fully. So I have a little bit more leeway. And that’s part of the evolution. The Thousand True Fans is only one option. And there are other avenues and sometimes it’s the place you start. I’ve been doing this long enough that I have more than just a thousand true fans.

AD: If you were starting out now though, given everything is the way it is, how would you do it? How would you make Wired? Would you make something like that?

KK: It’s a good question. I don’t know. Starting a magazine these days would be really tough, because we’re in this period of time when it doesn’t make any economic sense to have a magazine. The only magazines that are working are subsidized fully. But we’re in this period where it’s almost going to be feasible. Substack is kind of showing the way. And then I think the next step after that is to have full-blown magazines where you have multiple authors and people are kind of trained again, re-educated to pay for good stuff. For a while, it was not the culture to pay. And Substack has really been leading the way of retraining people to be willing to pay for writing and periodic writing. And so I think the next step is an evolution of these multi-author things, which are all magazines. I don’t think it’s quite there yet. So if I was starting again, it’d be a little rocky to do it. You might have to have funding from somebody who’s willing to fund something to get it going until maybe a couple more years until there’s more of a base. So maybe that’s what I would do. Find somewhere that’s willing to fund and invest into something a little bit longer term.

What else would I be doing? Well, it’s very clear. I would be much more active on YouTube. YouTube is completely underrated, both in terms of its consequences and audiences.

AD: Say more about why you think that? Are you using YouTube a lot?

KK: Oh yeah, yeah. I’ve subscribed to like 400 channels. I watch way too much of it. Moving images is where the center of the culture is. It’s not books. It’s moving images. I have a little media empire, right? We do podcasts, newsletters, websites, books, everything, cool tools, all this stuff. The only place where there’s a growing audience is YouTube.

AD: That’s fascinating.

KK: Yeah, and it’s not really recognized. I’m sure school age kids learn more on YouTube than they do at school. And it’s subterranean in the sense that, in the old days, you would look at the newsstands and you would say, my gosh, there’s a lot of magazines for guns or there’s a lot of interior decorating and makeover magazines. On YouTube, you don’t have that sense. You don’t know that there are these huge, huge communities and whole domains of stuff that you don’t see at all. And it’s so much bigger and deeper than anybody realizes. People pay attention and now they’re doing merch and finding all kinds of other ways to monetize it. Again, the thing about YouTube is you don’t have their emails. So you have to do something else to move them into your thousand true fan domain because you can have millions of fans, but they’re not true fans in the sense that they’re not paying you directly. And so moving them into that true fan thing is another whole process.

AD: What do you think people will be paying for more, given the fact that AI can make really good quality work?

KK: Yeah. Well, I was just saying this the other day, am I paying for anything that’s completely generated? And the answer is no. I’m not. I’m not subscribing to any Substack that I feel is AI generated. I’m sure I’m reading things that are, but I’m not paying for them.

It’s unclear to me how far completely AI generated content of any type will go. On my Instagram, I do follow some AI artists, which I really enjoy, Time Traveller, and there’s Joann. There’s a bunch of them that I love and it’s only AI generated stuff. I haven’t quite got to whether I would pay for them or not. The reason why they’re so good is because there’s a lot of them in it. It’s not just pushing the button kind of thing. It’s a skill. And I think what we have learned so far is that there is a whole bunch of art that I want to make that I could not generate because there were no words for it. That art was always tethered to language. I was trying to make the thing where there is simply no language for it. Even if later on when you could do the reverse prompting and take a prompt out of it, I was still not going to produce what I wanted.

So I think there are still constraints within these models and what they can produce because of what they’ve been trained on. And that’s going to continue for a while, even though we are trying to make them broader and give them some abilities to explore areas. They’re still constrained by what has been produced. Therefore, I think while there are lots of things that can be done that are really going to be fantastic in that space, there’s a lot that they aren’t going to be able to reach for the time being. And so the human alone or human with AI is still going to be a very, very profound source.

In the near term, 10 years, 20 years, I see lots of partnerships with AI as creative. I’m using this thing, a beta version of Google, I call it a writer’s room. It’s like every creative writer could have a writer’s room. We have multiple writers working with you on your project. And I can see more of that. There could be something that’s completely AI generated. And I think that it’ll be less interesting than the one that’s co-generated. And then there are those who won’t use it at all. So I see those two: those who don’t use it at all, and co-generation. But 100% generation, I don’t see that being that interesting to most people. It’ll be out there and there’ll be some uses for it. But for art, not so much. For everyday logo generation, yeah, great. There are different levels of what we demand of creativity. In some cases we don’t demand very much and it’s going to be sufficient.

AD: I feel like the appreciation of that kind of weird AI art will be similar to how people appreciate Gonzo, Outsider because it’s bad.

KK: Yeah. The constraints and limitations of any media later on become that kind of defining character that people are nostalgic for. You can see that in Midjourney already. Some of the early Midjourney stuff is so Midjourney-ish. That will be like a style at some point. So, yeah, I think you’re right. It would be kind of an edgy thing where you have this alien art, but a lot of co-generated art that’s really, really good, and then some human only. More choices.

AD: You strike me as a very chill guy. Are there things that stress you out? What gets you really?

KK: No. I’m never stressed. Why should I get stressed? It’s not good for you.

AD: I didn’t know that was possible. How do you do that? Or have you always done that?

KK: I’ve always done that. I must be a genius. No, but I also engineered my life not to have stress. I married someone with no drama. It was very deliberate. No drama. What you see is what you get. There’s no games, no hidden secrets, that kind of stuff.

AD: What advice would you have for very stressed out people then?

KK: I think you’re asking the wrong person. I have no idea. You should ask someone who’s very stressed. They’ll have answers for you.

AD: I want to know what you said you designed your life around.

KK: I think taking a long view, as much as possible. This too shall pass. The thing about the long view is the little tiny ups and downs. You realize that’s just going to be overwhelming over time. Understanding that setbacks are temporary. That’s the optimistic view. Setbacks are not about you, they’re not you, they’re just temporary things that can be overcome and will be overcome. So when things are happening that I’m not happy about, it’s like, well, okay, we’ll get around this. So what’s the best way to deal with it rather than getting upset over something?

Here’s one thing that I, even as a teenager, realized: when people are really angry or rude or whatever, I just imagined that they have a disease. It’s like, this is a symptom of an illness. And so it turns to sympathy rather than anger. It’s like, I’m sorry that you have that thing that makes you so angry. That must be terrible. So rather than responding with anger, I say, they’re sick. They have an ailment. I try to be polite, which is also disarming. So yeah, on the road when someone cuts you off, I try to think, you’re having a bad day. I’m sorry you have to do that.

AD: Wow. Do you meditate?

KK: No. I don’t feel any different afterwards. It’s the same. I have a natural high, I guess. Never did take any drugs. Never take any. I don’t think I can teach anybody any of this. I do have a book of advice.

AD: Yeah, is there anything you’d add to that?

KK: Yes. When you’re doing heavy manual labor, a change in position is almost as good as a rest. If you’re writing, you want to be as specific as possible. Instead of saying there’s a flower in the vase, you say a pale rose rests on the lip of a water glass. The more specific and concrete you can get, the better your writing will be. Some investing advice, which is don’t buy stock unless you’re willing to hold it for three years. That’s investing rather than just day trading. And that will generally pay off better for you.

I think I also recommend for long-term happiness that you try to do something useless every day, like drawing or art. I make art. I have this huge book and I needed a book stand. So I made this book stand out of scrap wood. It’s really, really weird, but it looks kind of beautiful and funky. There’s a certain funky style that I like of unfinished rough stuff. Anyway, I spent a lot of time on it. It’s completely useless in that sense, but it’s beautiful. It just brings me all kinds of pleasure. So yeah, try to do something useless.

You know, a lot of people talk about trying to optimize their life and they’re doing the minimize things, get through stuff as fast as possible. And I think you want to do the opposite. You want to try to work on things where you want to spend as much time doing it as possible. So you want to look for those things that you want to keep and never stop doing. That’s the sign of success. It’s not how little time you work on things, but how much time you work on things. And so try to steer your life towards the things that you just want to spend as much time doing as possible, whatever that may be. Like talking on a podcast with interesting people, do more of it.

AD: I know you have a new book out. I want to talk about that. Tell us. Bam!

KK: Yeah. So here’s cool and useless. I have this book, Colors of Asia, which was based on about 300,000 images that I took in Asia over 50 years. And the weird whimsical thing is that they’re all arranged by color. So I have pages of these sort of odd, remote, esoteric stuff that I saw over 50 years in remote parts of Asia. And now they’re all arranged by color. And if you like really esoteric things, this is a book for you. It’s for designers and people who like alternative ways to design.

So it’s Colors of Asia. And I have a Shopify where it’s a self-published book. And I only printed a thousand for my thousand true fans. So if you’re a fan, there is a copy for you. Nobody was yelling for this. No one’s saying we need this book. Nobody said that what the world needs is a book of photographs organized by color. I just decided it just tickled me and I haven’t had as much fun making a book in a long time. I just really had fun doing it and hopefully some of that enthusiasm and smiles came across. And so it’s art. You don’t need to read this book! It is cool and useless.

AD: Yes, we need more of that. I love the: nobody asked for this, but here it is.

KK: Nobody asked for it. And by the way, some of these photographs are from the 70s, 80s. Even stuff from the 2000s is not happening anymore. Like the elephant processions in India, animal rights legitimately are concerned. Like, no, they shouldn’t be in processions, elephants. So they’re gone. And so this is also a little time machine of seeing things that used to be, but they can be inspirations for things that could be now.

AD: Well, thank you. I want to close with a quick speed round. First question, how would you describe your relationship with AI?

KK: For me it’s an intern. A useful intern.

AD: What do you wish AI could do but fear that it never will?

KK: I don’t see anything it can’t eventually do.

AD: Everyone’s making predictions about AI. Let’s make a prediction about humans. In 10 years, where will humans be?

KK: Ten years will be just about where we are right now. I don’t see much change in humankind in that short period of time.

AD: How about a hundred years?

KK: A hundred years, we might begin to speciate in terms of there might be people who want to edit their genes and others who will never do that. So I could see some genetic diversion in 100 years.

AD: What question about AI and creativity isn’t being asked enough?

KK: I suspect that we’re going to have multiple kinds of AI. So I don’t think we talk enough about the plurality of AIs and the ways that they’ll be different. And there could even be AIs that are developed and acquired and programmed just to help artists with different biases and different backgrounds. So I think we don’t talk enough about the plurality of AIs.

AD: Last question, what do you wish people talked about more instead of AI?

KK: I think there should be more art in the world. I think there should be more art in corporations. I think artists in residence have proven to be powerful within corporations. I think artists should be more comfortable in their role. And again, what they’re doing is often useless and that’s hard to justify in a corporation. But we know that actually they will be productive in an inadvertent way. They’re inadvertently productive. It’s not a big sell, but it is true. And it does work. So even though they’re doing useless stuff, in the end, it’ll be more productive. That’s kind of a weird thing. But yes, that’s what we need.

AD: Cool. That’s a great place to end. Thank you so much. This was amazing.

KK: Thank you. I love your questions. Best to you and don’t stress. It’s not good for you.

---