Transcript of Currents 064: Michael Garfield and J.F. Martel on Art x AI

The following is a rough transcript which has not been revised by The Jim Rutt Show, Michael Garfield, or J.F. Martel. Please check with us before using any quotations from this transcript. Thank you.

Jim: Today, we’ve got some interesting guests. We have J.F. Martel, who is a writer and podcaster. He writes on art, culture, religion, and philosophy. He’s got a website and he’s got a book called Reclaiming Art in the Age of Artifice. And we’re going to be talking about that concept today. He has a podcast in which he co-hosts with Phil Ford called Weird Studies. That’s an interesting name.

Jim: Our second guest is Michael Garfield. He’s a transmedia artist and performance philosopher. Now, what the fuck a performance philosopher is? I don’t know, but I’m sure we’ll find out here today. I do know he’s a musician, an illustrator, and a painter, and he has a truly fascinating podcast. Lots of cool guests on it, including yours truly called FUTURE FOSSILS. Check them out. Welcome, guys.

J.F.: Well, thank you. I’m glad to be here.

Jim: Yeah. The topic we’re going to talk about today is something at the intersection of technology and art. What is that interaction about? And I think the original thing that spurred me to reach out to Michael and say, “Hey, let’s do a podcast,” was him using about what effect would things like Dolly have on art. What is NFTs doing to art, et cetera? And, well, we don’t need to stick to that. It might be a place to start the intersection of technology and art.

Jim: Who wants to go first?

Michael: Well, I mean, we might as well start with, I think, the post in question, the Facebook post that you replied to and invited me onto your show was one of several conversations I’ve been prompting on social media lately just to invite more conversation about the double-edged sword here of all of these sophisticated new tools. You mentioned Dolly.

Michael: I was just playing over the weekend with an invite to Midjourney, which is another one of these really powerful AI art programs. But of course, AI art is the term social distancing or something. Depending on how you define those words, it does or does not make sense. And I think we’re early enough in the conversation, I want to pay tribute to a piece that I know I sent you all by Eric Houle, who’s a fascinating author and scientist and interesting boundary-crossing person that I vibe with, who recently wrote a piece citing Walter Benjamin.

Michael: And that essay, the Art in the Age of Mechanical Reproduction, as well as Houle’s story and others talking about how art that is not created by a conscious agent isn’t really art, that this new emergent field of promptism, where you’re giving text-based prompts to these generative adversarial networks that have learned how to style transfer from enormous data sets of different artists that they’ve been fed, that these are not really … They may be in his most optimistic scenario, complex paint brush that can be used then as raw material or a substrate.

Michael: It may never get good enough to effectively become indistinguishable from the work of master artists, but that seems to, Eric and myself and lots of other people, naive assumption that it’s never going to … And so I was thinking a lot about this relatedly in 2017 when Adobe first announced that they were capable of cloning the voice print from someone with only 20 seconds of audio recording, and that you could then write text and then have the computer speak it in the voice of anyone you wanted.

Michael: And so actually, while I was in Montreal that summer, I started writing a short story called an Oral History of the End of Reality about the epistemic shock of complete failure, the loss of the arms race against digital counterfeiting technologies and what that would mean to us psychologically and culturally. And so I’ve been thinking about this stuff for a very long time and its effects on the labor market and so on. And as someone who has already, at 38, lived through a profound amount of technological change and the impact that it’s made that I’ve observed on my own career and the careers of many, many other artist friends, it just seems like a really important topic now to try and get a … We can’t really get ahead of this, but we can try to stay on top of the wave as best we can and not get sucked under.

Michael: And the question of how these technologies are not only changing the way that we do business with one another, the way that we express ourselves, but also the way that we understand the self, the way that we understand what it is that humans are good at. What we uniquely do if we do to anything uniquely? And that itself is an interesting question. Yeah, that’s the bundle of ideas that I’ve been trying to juggle and explore and discourse with people.

Jim: Yeah. Well, very obviously interesting and timely. J.F., what do you have to say about all this?

J.F.: Well, I actually joined the Midjourney discord last night so I got to experiment with this newfangled AI that’s able to generate what, to me, looked like genuine works of art, of varying levels of quality. But the point is some of them were actually quite striking and haunting. So I did get the sense that this thing was generating art. But then, of course, the question is, what is art? And how we define the term will have a huge part to play in whether we, ultimately, decide that what computer software and neural network can generate is art or not.

J.F.: And I’m really reading this whole historical moment, this emergence of the AI artist as an indicator of where we’re at epistemologically in our culture, how we think of ourselves and our relationship to reality, what we think consciousness is and where its locus is. All these questions that are metaphysical and extremely, of course, extremely complex and very ancient questions play into this discussion, I think.

J.F.: And ultimately, also, I think that one of the most urgent aspects of the problem have to do with economics, politics, and power. To use a word from Jim’s generation, how do these systems co-opt certain processes which were, at one point, owned by the human that they came from the human, and suddenly we’re seeing machines do what we thought only we could do. It’s a bit reminiscent to the moment where Deep Blue beat Garry Kasparov at chess, right? It occasions an existential crisis, where we have to rethink certain assumptions we had about what constitutes the human and what constitutes human creative activity. Yeah.

Jim: Of course, this is not the first time we’ve confronted this question. If you look back in the history of art, there was a great kerfuffle about photography, right?

J.F.: Exactly.

Jim: You don’t say, “What the hell is this,” right? And then there was another gigantic kerfuffle about recorded music.

J.F.: Right.

Jim: And if we go back further, famously, Socrates was highly dubious about writing, right? He thought that this was a corruption of the art of philosophy, the art of literature, the art of drama. And of course, it was about 400 years after the adaption of the greatest innovation of the Greeks, which was the Venetian alphabetic language model. And there was huge amounts of controversy. So this is not, by any means, a new question.

Jim: And when I was thinking about it a little bit, there’s one distinction that might be useful. And that’s the distinction between artificings or a word for that must be making an artifact.

Michael: Artifice, yeah.

Jim: Artificing and discernment, right? So in some sense, one is artificing, making something, but a lot of stuff artists make, they throw out, right? How many sketches does a visual artist make for each one he turns into a painting? Depends on the artist. Some make a lot, some make a few.

Jim: And so there’s that intersection between making and discernment, and then there’s discernment itself, right? Actually, my wife asked me, “What the hell are you going to be talking about today on your show?”, right. I said, “Oh, blah, blah, art and technology.” And, “Oh, that’s a really interesting question.” She’s a professional art photographer amongst other things. And so we had a little back and forth and I said, “Yeah, for instance, suppose I programmed one of my drones and sent it on a 30-minute mission and randomly had it take a thousand photographs of stuff in the mountains near where we live. And I gave it to you to go through and pick the two best images. Is that art?”

Jim: And she goes, “Hmm, that’s a damn interesting question.” And she came down with the answer that, yes, the result would be art because the discernment was what … And again, she had never thought about this question in quite this way. So I’m going to turn that question back to you guys, your relationship between discernment and making, artificing. J.F.

J.F.: It’s a fascinating question. I actually just wrote an essay about this for a photography book that’s coming out in a few months by Shannon Taggart. She’s an amazing photographer. She’s revived the old genre of spirit photography. So she concentrates on taking photographs of things like spiritualist seances and that sort of thing. If you’ve seen the old Victorian photos of dubious photography of the spiritualist movement of the late 19th, early 20th century, she’s exploring that as an artistic project.

J.F.: And in this essay, I came up with this thought experiment where you take a great painting, for instance, van Gogh’s last painting is the one that I use, the one with the crows flying over the wheat field, which is obviously a beautiful and powerful piece of art. And I was imagining instead of that painting coming upon a photo that was taken either by a camera working automatically or maybe by a child or maybe a photo taken by mistake, but which coincidentally was identical to the painting, had the crows, the wheat field, the path, all the elements of the painting would be present in this random photo.

J.F.: So the question is all the meaning we extract from the painting, would we be able to extract that meaning from this random photo, which is identical to the painting? Do you understand what I’m saying?

Jim: Yeah, absolutely.

J.F.: And the answer is obviously, yes, one could extract all that meaning, the symbolism of the crows, of the wheat, of the light of day, the falling light. You could come up with a symbol intimating death or transients or whatever it is that people have read in van Gogh’s painting. But of course, the one missing element is the discernment part.

J.F.: But to me, that’s hardly an argument against the photo, the accidental photo being art. To me, that’s an argument that the world somehow is constantly producing artful moments. That’s what Jung called synchronicity. Whether it’s intended or not, we’re constantly living through moments where divergent and disparate causal series that have nothing to do with one another suddenly converged to produce a moment of meaning. In a sense, you could see the human being itself as a synchronistic event in the history of the universe if you want to be a materialist because we’re meaning-making and meaning-discerning creatures, I’m putting on my secular filter right now, in a world of blind material interactions.

J.F.: So that the AI might be able to generate art and that the AI might even be able to discern to a certain extent based on art theory, as we know it, because no doubt, these algorithms include commands that ensure that what comes out of the AI will respect certain traditional rules of composition and stuff. Well, then, yes, all that is possible not because robots are natural artists, but because the world as such has a weird tendency to produce meaning-making and meaningful events, strangely.

J.F.: So to me, I was looking at this Midjourney, this AI, generate all this imagery, and what I was seeing, the closest analog to what I was seeing, was dreaming this constant generation of meaningful, if sometimes absurd imagery out of this imaginal depth that contains perhaps, in a sense, all possible imagery. We seem to have given material form to the dreaming, to the act of dreaming. And that’s something important, I think. It’s something meaningful, but when it comes to art, we really have to have a careful discussion about what we mean by art in that sense.

Jim: So does that put you down on the side of discernment as, in other words, the computer might generate a bunch of stuff that adheres to artistic traditions and statistical similarities about color, balance, and things of that sort, but it’s not art until somebody says this is meaningful and tags it.

J.F.: Yeah. Yeah. Well, the real question … And then I’ll throw the ball to Michael here. The real question ultimately is we can create an AI artist. Could we create an AI art critic? Could we create an AI art appreciator? That’s the real question.

Jim: And of course, in the AI world, we have this idea of gender adversarial networks, the idea where you have one that creates and the other that criticizes or attacks, right?

J.F.: Yeah. And that’s built into the process, right?

Jim: Yeah, exactly. And they get better and, of course, you need a scorer. So for instance, you’d still have to get some humans in the loop to say, “This is good. This is bad.” But eventually, you could train up that critic to be better probably than a crowdsource Mechanical Turk evaluation at least. And does that mean he’s an actual critic? That particular neural net, it’s, again, very interesting question.

Michael: Well, I mean, it’s a recursive endless regression because then you’d say, “Okay, we’ve trained these GANs to create and to critique one another. And again, that speaks to the arms race between counterfeiting and forensics analysis. And then you have, “Well, we need a human on the loop to evaluate the performance of those opposing networks.” And then when you say, “Okay, but it may be better than a human critic,” I mean, it’s a constant return to this question of, “Well, who’s determining that it’s better?”

Michael: And J.F., it’s funny that you bring up the AI art critic because, in fact, I recently did see a project that was trying to train AI to create automated product reviews for products that, of course, the machine has never actually directly experienced. I mean, that’s a very interesting and surreal path that we’re on.

Michael: I just want to note … Because I was thinking quite a bit about the process of art criticism when you brought up discernment and I was reminded of this passage from the power of myth where Joseph Campbell says, I’ll just read this quote real quick, “In India, I have seen a red ring put around to stone and then the stone becomes regarded as an incarnation of the mystery. Usually, you think of things in practical terms, but you could think of anything in terms of its mystery. For example, this is a watch, but it is also a thing in being. You could put it down, draw a ring around it, and regard it in that dimension. That is the point of what is called consecration.” That’s the Power of Myth, page 74.

Michael: And of course, the ritualizing, the consecration, I know that, J.F., you and Phil have talked about this in terms of the magic circle created by a boxing ring or by the suspension of disbelief that one performatively and ritually engages in when you enter a cinema house that there are these spaces and we organize our lives episodically by moving through space from one place into a different place. And that is how we decide to treat one situation with a separate set of rules, that we enter this conversation with an implicit understanding of what it means to be in the ring with one another as conversants here.

Michael: And so when it comes to this question of discernment, I think this whole thing is very … I’m glad you said a human on the loop, Jim, because the whole thing really is quite synonymous with the question of automated weaponry and drone strikes. And then is there someone in a sophisticated military arcade somewhere in the United States that’s approving and denying requests by the drone to actually fire its weaponry? Or is the drone acting autonomously? And if it is, then who’s responsible?

Michael: One of the most interesting art projects I’ve ever seen with AI is … I’d have to take a moment and look it up. But it was a machine, it was like a robot that was installed in an art museum and was using cryptocurrency to order illegal drugs over the dark web and …

Jim: All right, I like it.

Michael: … have the drugs mailed to the museum. And the project was all about who’s accountable for this illicit drug purchase, right? My robot bought drugs. And this is the problem that we think we’re generally facing with society now, where as it happens, the order of the first and second search result, when you ask Google for information on a political candidate, can swing an election up to 25%.

Michael: And so these search companies are at least, in theory, capable of rigging an election in such a way that when they’re found out for having this influence, they can basically punt the blame to a rogue employee who can be washed out of the company. And the question of who actually was responsible for that algorithm becomes this vertiginous mystery.

Michael: Yeah, again, I just think there’s this relationship between the question of the modern self. I am choosing to author my own value systems. I am making decisions and this postmodern algorithmic self that people like Yuval Harari have said is the new religion of Silicon Valley, which is that I’m just a collection of automata. I’m just a set of algorithms interacting with one another. And that there really is no free will through that lens. There is no agency.

Michael: And so in that sense, we have to take a much more nuanced stance on, what do we even mean by discernment? I don’t know, there’s a lot of different threads there.

Jim: Yeah. Yep. Let me throw back a couple of items. One, I just like to point out this because I think it’s so funny and scary at the same time. We talk about the AIs and they’re coupling with humanity. What’s the strongest current coupling with AI? I think it’s one people don’t even think about, it’s the algorithms in the dating apps, right? It’s literally controlling evolution at this point. If that is now the number one way people meet, which I believe the statistics say that it is, some AI trained up on something for some purpose that probably has nothing to do with anything other than making money is now driving the evolution of the human race. Isn’t that interesting, right? And it’s probably a really stupid AI, but nonetheless, it’s deeply coupled to who our civilization is going to be.

Jim: Now, back to … And people never think about that, but the AIs are already controlling our evolution. Goddamn it, motherfucker, right?

J.F.: The singularity has come and gone.

Jim: It’s come and gone. Exactly. I’ve seen that argument.

Jim: Now, the next one is about this idea of discernment. And I like your example of the red ribbon wrapped around rock or something. And this is the idea of found art classically described in art history books, Duchamp and his toilet, right, as a good example. There’s been many of them since then. And, again, it’s getting very meta, right? Or what’s the one, this is not a hat or something. Or this is not a pipe or I don’t know. Another one of these just fucking around. Art as fucking around at some level, at least those Philistines like myself would say that. It’s like I tell people every time I go to New York. My wife and I always like to go to the Metropolitan Museum of Art and wallow in the 4,000 years for the greatest art ever created.

Jim: And then I walk into the goddamn modern art wing and I want to pull out a machine gun and just start shooting. I go, “What the fuck is wrong with the society and call that shit art,” right?

Michael: Well, the CIA is responsible for American modern art. I mean, they’re-

J.F.: Abstract expressionism, yeah.

Michael: Yeah. I mean, they promoted this as part of the info warfare against … They wanted to make America look like a creative superpower and that our culture was emblematic of our superiority. And so they were the ones that took Jackson Pollock and gave him a career and put all these shows together across Europe.

Michael: I mean, when you look at a Pollock and you say, “There’s no signal here, that’s just noise,” it’s actually that the signal is the man behind the curtain of Pollock’s success, as far as I’m concerned. What we’re actually looking at is a noise attack against the reasoning of the Russian art scene. And I don’t know, it’s … Yeah. Anyway.

Jim: Well, that’s interesting. I could tell you a little story here because I’d love to tell tales. I had lunch with the leading critic or evaluator of Pollock, right? And we chatted quite a bit about Pollock and his art and how did you authenticate a Pollock. And he told me it took two months to authenticate a Pollock that was worth a hundred million dollars from a fake that was worth a hundred.

Jim: And of course, I’m cynically thinking, “Well, fuck,” right? There’s something seriously wrong with this picture. If it takes the world’s leading expert two months to tell whether it’s a hundred dollar fake or a hundred million dollar real, why would I not just get a hundred dollar fake, right, and have whatever effect comes from that artifact? And as Michael knows, I lived in Santa Fe for many years, one of the leading art markets in the world. And my wife and I did collect art, but we did it in a very field esteem fashion.

Jim: I told people, “Yeah, I’m willing to pay about $500 a square foot,” right? And they’d say, “Well, what about really great art?” And I’d go, “Yeah, I’ll pay 50 grand for a Vermeer,” the original, right? And they go, “What?” And to my mind, I don’t really give a shit if it’s an original or not. I know that in the philosophy of art, there’s this big question of, is it the original, it’s somehow important. I think that’s utter horseshit. And that if an identical Vermeer that could be had, I’d happily hang it on my wall and would think it just as nice as having the actual Vermeer.

Jim: And this idea of the actual artifact in my mind is bizarre and fetishistic. And the whole art scene guarding itself all whipped up into a frenzy about this stuff strikes me as crazy. It’s like $160,000 for Elvis’s comb or some goddamn thing, right? So anyway, Michael and I have been rattling on. J.F., your turn.

J.F.: Yeah. Well, I’m just going to stay with what you’ve been saying because I think it’s really interesting and important part of the problem. The problem of the authentic artwork, right, as opposed to the counterfeit.

J.F.: Now, you said you’d be just as happy to have a fake Vermeer that’s identical to the original as you would to have the original. And I understand what you mean by that because you would be getting all of the art. It would all be there already. So why would you have to buy the original when you can get a perfect replica? And that’s a valid point when you consider the art from the symbolic end of the access of what art is. At one end, art is imaginal, but using that word, I mean that it’s an image that resonates with meaning in a multivalent and ambiguous way.

J.F.: I think it’s true of all artworks. It invites exploration, interpretation, questions. And yet at the same time, it seems to be expressing something positive in some way. So if we look at that end of the art axis, then, of course, the fake for me is just as good as the original. But if you look at the materiality of art, that’s the other thing. And that’s what Walter Benjamin is writing about in his famous essay on The Work of Art in the Age of Mechanical Reproduction.

J.F.: The materiality of art or more specifically the historicity of art, the fact that art came at a specific time, the fact that that image, whether it’s the fake or the original you’re looking at by Vermeer, that that image emerged at a specific time from a specific subjectivity experiencing specific confluence of reality at that moment and such that he felt or he was driven by this need to express this particular image. That’s also an important aspect of what art is, I think. And that’s what gets lost in the idea of AI art. AI art comes out of a null time. And I think that it comes out of a pure logical time.

J.F.: For instance, I’ll give you an analogy that might be useful, and a famous Jorge Luis Borges story, The Library of Babel, right?

Jim: That’s one of my favorites. That’s a good one.

J.F.: Yeah. Wonderful mind-blowing story about an infinite library in which there’s an infinite number of volumes, and each volume contains a completely random combination of the 22 letters of this fictional alphabet, plus a few other syntactical or orthographic symbols, like a comma, period, and that sort of thing.

J.F.: So if you consider a finite number of letters, a finite alphabet, and an infinite library, that means that library will contain not only every text that’s ever been written, but every text that could possibly be written. So everything has already been written in this infinite library, which since it’s infinite, does not exist in the space time we know, it needs to be some eternal space time that Borges is imagining here. So it’s in the space of pure logic, all books have already been written. That’s a fact.

J.F.: In fact, somebody actually came up with a computer version of The Library of Babel, and I found entire paragraphs from my book in that software or whatever, on that website, I can’t remember the web address, but perhaps someone on your crew can locate it. That’s pretty amazing. So in other words, every book has already been written. It’s true. In the space of pure, atemporal logic, all these things already exist. And AI art is just another version of that. You have a finite number of pixels, a finite color spectrum. Therefore, every painting could potentially emerge out of this AI and, in a sense, already exists potentially in its makeup.

J.F.: So at that point, the question of historicity, that other access of art, the part where art is about what was said, what was shown at this one time in history becomes super important. And the abolition of history, which we’re all living through now, the abolition of time under the techno capital everything all at once thing that’s going on right now is making us, I think, forget that important dimension of human subjectivity, which is that we exist in time and in history. And that art gained all its meaning from its historical or its finite nature that it emerges and dissolves in time.

J.F.: And so if you let go of that access, that pole, then, yeah, everything becomes art and also all art has already exists. It becomes, in a sense, meaningless. Yeah. I don’t know.

Jim: Yeah, yeah. I’m going to push back on because this is one of my favorite topics that I hate is fucking infinity, right? You can go down a rat hole of fuckery once you bring infinity in. And fortunately, as far as we can tell, we do not live in an infinite universe. And if I could put on a wager that would be paid off in a billion years on the fact that we can prove we don’t live in an infinite universe, I’d make the bet, right, because you get into all kinds of fucked-up situations once you allow an infinite universe. My favorite being Boltzmann brains, right?

Jim: A Boltzmann brain is a brain that spontaneously emerges from quantum fluctuations that’s sufficiently powerful to simulate anything you want, including, say, our total visual universe, right? Our 13 billion light-year radius universe. And once you entered, not only do we have Boltzmann brains, but we have an infinite number of identical Boltzmann brains and an infinite number of identical Boltzmann brains that differ by one bit, right? So get rid of infinity. It just leads to absurdities.

Jim: And then I would also say, very importantly, that even within the finite world, the number of possibilities is so large that you have to bring in things like historicity and discernment, for instance. And this is in a very interesting and basic idea from integrated information theory, a very simple artistic frame, a thousand by a thousand black and white screen has two to the million possible patterns on it. Two to the million is way more than the number of fundamental particles in the universe. Way more, right? And that’s just on this very simplest artifact.

Jim: And of course, if you just start generating static, most of it would look like static on the screen. Almost all of it would be static. And this goes to, basically, the Boltzmann’s laws of thermodynamics, right, which is disorder is vastly more common than order. However, if you let that random number generate long enough, you’d eventually get a picture of Abraham Lincoln. And eventually, you get a picture of me. Eventually, you get a picture of somebody signing I love you in cursive.

Jim: And so the argument from competent torics, I think, is very misleading. And then I think that brings us-

J.F.: Well, I agree with you. I agree with you.

Jim: It brings us back in a very strong way to your argument for context and historicity and if it’s going to be human art, then tied to something human.

Michael: I love this and it’s a really fertile place. And I’m glad that the two of you brought the conversation here because I wanted to bring up this piece by W.J.T. Mitchell called The Work of Art in the Age of Biocybernetic Reproduction, which is rehashing of Benjamin’s concerns in an age of computation and big data and bioengineering. He talks a lot about Jurassic Park and The Matrix. And consequently, I love this essay.

Michael: And he says in this essay, reproduction means something quite different now when the central issues of technology are no longer mass production of commodities or mass reproduction of images, but the reproductive processes of the biological sciences. What does it mean when the object on the assembly line is no longer a mechanism but an engineered organism? And for me, this is where the rubber really hits the road because the question for me of this reproducibility is a question of The Matrix. It’s a question of Leeloo Dallas being 3D printed in a machine in the Fifth Element.

Michael: It’s this question of … A lot of my writing is motivated by the prompt of what is it going to mean when you 3D print your babies. And already, people are prophesying that we will start, basically, raising digital offspring in the metaverse. And that eventually out of the thousands of parallel iterations of possible children, you’ll have selected one that you actually want to 3D print with these. And then they’re … I don’t know what to make of this futurism. It seems relatively appropriated by really strange venture interests.

Michael: But at any rate, Mitchell goes on to say, and this is echoed in people like Douglas Rushkoff’s book, Present Shock, when everything happens now, talking about this tension between, as J.F., you’re talking about, the historicity, the situatedness of these living sapient people of a given work created by them. And then this other thing, which is, Jim, the apparent order, this pareidolia that we find in the random pattern that looks like somebody’s written I love you in cursive.

Michael: And Mitchell says later in this essay, “My own view is that the present is in a very real sense,” even more remote from our understanding than it was at the time of Benjamin and Picasso and Lennon, and that we need a paleontology of the present to rethinking of our condition and the perspective of deep time in order to produce a synthesis of the arts and sciences adequate to the challenges that we face.

Michael: So earlier you asked, Jim, what is a performance philosopher? I’m really inspired by and riffing on MIT historian, William Irwin Thompson, who I know we discussed briefly when I had you on FUTURE FOSSILS. And Bill pioneered this mode of communication that he said was appropriate to an age of jazz and complex systems, and the global technocratic processes that he called Wissenskunst or knowledge-art, which was a way of cutting across disciplines and looking for patterns in the apparent noise of all of these vast and indeterminate systems.

Michael: And so when I think about what’s required here, Bill, also said that novelty emerges into culture first through the crazies, then the artists, then the savant, then the pedants. So he’s putting art upstream as others do, who have noted the dependence of modern physics on abstract art and on cubism and the multiperspectivalism that’s performed through cubism, that something must first be conceived in order then to be quantified and formalized in a scientific theory.

Michael: And so if we are confusing this driverless process of pattern formation with, again, the process of a pattern recognition and a sense-making to use an abused term, then I think one of the things in jeopardy here is our ability to navigate the avalanche of collapsed space time that is set upon us by our digital surround by all of these technologies that are extending the horizon of prediction and of memory so far into the distance on either side, that we lose the ability to make sense of these patterns.

Michael: And so you have this tension that I know, Jim, you and I have both have spoken about a lot on our shows respectively between prediction and understanding. And the better we get at predicting things, the worse we get at understanding them. And we can understand things very well, but not be able to leverage that understanding for prediction. And so, again, I feel like the machine is a persistent companion with us through all of these historical and “post-historical moments”. But that machine is not just the external machine. It’s also the way that the algorithms are embodied in us.

Michael: And so I think that there has to be a reciprocity there where, on some level I want the machine to be able to make discriminations, but it gets back … Last thing I’ll just stack on this insane yarn ball of ideas is that I think about, I love this paper, the physical limits of communication, where Michael Lockman, Mark Newman, and Chris Moore write in 1999, that a message transmitted with optimal efficiency over a channel of limited bandwidth is indistinguishable from random noise to a receiver who is unfamiliar with the language in which the message is written.

Michael: I mean, they’re basically saying, as Edward Snowden did in his conversation with Neil deGrasse Tyson, that we may be, in fact, surrounded by optimally encrypted alien communication that we can’t recognize. And so, again, to this question of, what does it mean to live in a world that it may, in fact, be more intentional, more artful, more meaningful than we’re even capable of understanding? And we recognize it as noise. So I think epistemic humility is required here. But anyway, yeah, that’s a lot of different ideas.

Jim: Lots of cool ideas. I’m just going to hand that one over to J.F. and see what he has to say about all of that.

J.F.: Well, I certainly like the emphasis on time and on our experience of time because I think that that’s central here because it’s the one thing that the machine doesn’t have that we get is a finitude and time. And I agree with your reservations about infinity, Jim, but at the same time, if the concept exists, it’ll act in the world and it’ll make a difference. Thinking in terms of infinites makes a difference. For instance, like Borges’s story.

J.F.: I mean, I have to say, I typed out a paragraph of my own work and found it on this website. So I know that it wouldn’t have come off if I hadn’t typed it in. That’s an important point. It generated it because it’s possible. So in a sense. But in another way, in another sense, if we project our minds outside of time, then we lose the finite, right? So we have to bring this whole discussion back into, I think, the finite.

J.F.: I’ll just kind of move the goal post a bit here. Let’s say that Midjourney had been somehow invented in 1930 instead of 2021 or ’22. And in 1930, they got it going. And then it started to produce like Pre-Raphaelite paintings and impressionist paintings, and people were like, “Wow.” But then all of a sudden, if it started to bring forth Jackson Pollock’s and abstract expressionism, or if you dial it back a bit, go even further back like cubism, all of a sudden, we’d be like, “Ah, this AI doesn’t work.”

J.F.: But regardless of your personal feelings about modern art, I think that some, at least modern art, is actually quite striking and quite meaningful in some ways and engages in a dialogue that goes back and it gazes with the tradition that goes all the way back to all the beautiful figurative painting that came before it. The people at that time, the discernment would not be available. We would think that the program had failed to produce art.

J.F.: In other words, art, to me, seems to be really about what the new might be, what might come next, where things might go. The classical art of Rome, one could not have looked at that art and predicted the iconography of a Christianized Eastern Rome, right? That was unpredictable. It was completely unpredictable. No one could have gone from the Roman sculpture to the Roman icon that would emerge within 50 years or 100 years.

J.F.: There’s a kind of absolute newness that we see in art, but it’s everywhere, I think, in reality, but we really see it in art because art is about touching that. I think touching what Nietzsche called the untimely, that if we fully go fully on board with this AI art stuff, well, these algorithms, they’re just looking at the past, right? They’re like the angel of history in Benjamin. They’re looking at what was. They’re reconfiguring and rearranging preexisting materials while you could say, “That’s what every artist has always done.” And yet there’s always a moment where our work emerges that people say, “Well, that’s not art.” That’s not okay.

J.F.: And then eventually, we see that, yes, it is. There’s a novelty there that seems to become very difficult to imagine in this new art as pure technology context. It seems to lack an element that’s essential.

Jim: Let me throw something out and see how you react to this, of what may be missing. And that is an evolutionary dynamic, right, which is Darwinian evolution throws out lots of random shit called mutations and recombination. A lot of it doesn’t seem to matter. You make a small mutation, so-called neutral mutation, has no effect one way or the other. A lot of mutations are deleterious, makes the thing less likely to reproduce successfully. And then the rare mutation or recombination makes it more likely to succeed.

Jim: And the analogy I’ve started to use lately is the garage band, right? A lot of garage bands really suck, but if it wasn’t for garage bands, popular music would never evolve. It would just be rehashing the same old shit. We’d be hearing the Jurassic rockers, we’d be hearing somebody once more at a bar playing Amy. Oh my God, I’ll put a bullet in my head next time I hear Amy in a bar, right?

Jim: So we need garage bands with their assort of randomness, but they’re interestingly caught in an evolutionary context, which is if the friends and family of the garage band actually like that band a lot and say, “Hey, this is as good or better than the bands we hear on the radio,” then they’ll probably get a bar gig. And if the bar gig goes well, then they’ll go to a little local showcase and maybe an A&R guy discovers them and on they go.

Jim: And so there’s an evolutionary context that sorts the randomness. We could, at some level, called garage bands almost random creation, right? It just happens to be whoever happens to get together and start banging away on their instruments.

J.F.: Yeah.

Jim: It’s the dynamic, it’s the evolutionary context with sorts and curates in some community sense, which allows popular music to move forward. And I suspect something similar. I don’t understand the dynamics as well in the art scene, but something similar must happen in the art scene that steers in some collective co-evolutionary context towards a future.

J.F.: I think you’re right. But I think that there’s obviously a difference because … I mean, you can model it that way, but I think there is a difference. Let’s look at the emergence of punk, right?

J.F.: So the punk is not just a type of music. It’s a whole aesthetic, it’s a whole world, there’s a punk world. You can make a world that’s punk. We have all these subgenres in steampunk and clockwork punk and all this. There’s something about the punk that it has a essence to it. And it emerged suddenly in the early ’70s. Well, it emerged out of materials that were definitively and definitely not punk. In fact, it emerges partly as a reaction to those things, but it brings in a whole aesthetic and it also takes in, it takes up, and transforms that which came before it.

J.F.: So you’ll have the Ramones writing songs that sound a lot like Beach Boys songs or the early ’60s songs, but in this new idiom of punk. But a whole new set of possibilities become possible in punk, right? So the sudden emergence of the new in art and culture is, I don’t know, it’s more Laplacian. Is that the word I’m looking for? Then Darwinian? Is it Laplace? I keep forgetting the other guy’s name. It’s more Berksonian if you’re speaking philosophically, Berkson of the sudden emergence of a qualitatively new situation that simply could not have been predicted before it emerged.

Jim: And of course, that happens in evolution, too, though, very infrequently.

J.F.: Exactly, yeah.

Jim: And in fact, the work at a certain research institute, which we’re not supposed to name, is … I guess I can on my show, the Santa Fe Institute, some of the great thinkers, Walter Fontana, Jim Crutchfield, amongst them, have modeled the fact that often we would expect evolution to make teeny little changes or no change at all or even drift backwards, and then suddenly discover a portal and go into a whole new thing. I mean, the famous example is a Cambrian explosion, where a certain small number of simultaneous mutations created the neuron, few other things.

Jim: And in a remarkably short period of time, every file, every basic group of living things that exist today came into existence. And maybe as short as 5 million years was ridiculously a short time. And it was one of these punctuated equilibriums as it’s talked about.

Jim: And certainly, this happened, of course, in memetic. Social evolution has the advantage that it’s not bound by generations, by chemistry, or anything else. And so big moves can happen in culture very, very quickly.

J.F.: Exactly.

Jim: But the dynamics are, in some ways, analogous. And of course, it’s very important not to overmetaphorized memetics from Darwinian biology. And I was probably a little bit guilty there in my analogy. And I should have said metaphorically, they’re very, very similar that their contexts, which guide movement forward through some filtering process.

Jim: And I was there at the very beginning of punk or damn close to it. In fact, my little gang of minor criminals supported one particular punk band. And for a while, we were like the only people that went to the bar where they played, and they wrote a song about us and the whole deal. Yeah. I remember that epoch well. I mean, it was definitely a counterculture. We were that and we were not this and it emerged almost instantaneously. So that’s an interesting example.

J.F.: Yeah. So there’s a punk reading of Darwin that’s completely different from a ’60s pop reading of Darwin is what I mean. It doesn’t just change the way … It’s a seismic change, a genre. I mean, I’m a Gen Xer so I remember being in high school. It was very important which crew you ran with, right? I was a hippie, but that meant that I read Darwin very differently from how the jocks read Darwin. So it’s like-

Jim: Or the Ayn Randians, right? The Ayn Randians would had a very different read on Darwin.

J.F.: Exactly, exactly. So it’s how we interpret, how we discern. And so an AI, that would be capable of true discernment, would have to be able to change the criterion of its discernment, change the very foundation on which it performs a discerning act. It would … This was something that humans seem to be able to do that perhaps machines can do it, but in that case, machines would become very much like humans.

Jim: Yeah. Let me scare you a little bit. I had a podcast, one of my favorite thinkers, a guy named Ken Stanley. He’s a leader in neuroevolution, but he also wrote a very cool book on open-ended discovery, algorithmically, called Why Greatness Cannot Be Planned. And the book is basically a broad attack on optimization. And he and his collaborators who’ve come up with a number of technological ways to do things like explore for novelty, it makes a big point. We’re not optimizing on novelty, we’re exploring on novelty.

Jim: He calls an open-ended search and he gives lots of very clever examples of people working in this field. So there are people now who are now thinking outside the box and not using the traditional, “Okay, let’s take the world’s history of art and compress it into one giant neural net and have it generate art on demand,” right, kind of the Dolly or the Midjourney approach, but rather inform itself and work with itself and generate new and novel things.

Jim: And so there are definitely people working on and they make the point that these transitions can occur very rapidly. Once it gets some new attractor, it goes from state A, let’s say, generating Dixieland jazz to generating early ’50s hot jazz, which are, yes, they’re the same instruments, but they’re so entirely different from an aesthetic sense. They’re hard to even see how they’re related, and you can have those kinds of transitions if you use this computer-based, open-ended searches, Ken Stanley calls it. I really recommend that book, Why Greatness Cannot Be Planned that I based the podcast on.

J.F.: It doesn’t really scare me because that’s exactly what you were just saying about evolution, right? You’ll have these spontaneous punctual jumps and leaps, and nature is capable of that. And if we just see these algorithms as extensions of nature, respecting the same fundamental laws, we would expect the same spontaneity, given a flexible, malleable and, sorry, creative enough system, you would expect that thing to happen.

J.F.: And I think that if we feel threatened by that, we’re doomed as humans. We have to give nature a lot of things that we used to reserve for ourselves, right? Descartes had a killer theory for how to make sense of the world, so long as we regarded animals as automatons.

Jim: Well, he was the worst. He took us down so many wrong roads. We could have a whole podcast on goddamn Descartes, right?

J.F.: He did, but he served a very important purpose at his time, I think. And it was very important for the sanity of Europe that Descartes come around and provide the means of preserving some locus of human agency in what was then an increasingly mechanical vision, not just of the cosmos but of the human body. So he was coming in as a savior, I think. And of course, he created all these problems.

J.F.: But the point being that it required him to reserve for the human things that actually exist out in nature that are out there already. And we’re seeing that now through our technology. Our technology’s becoming naturalized. It’s joining nature or it always was in a sense, but now we’re seeing it. We can’t deny it anymore.

Jim: Yeah, very good. Very true, actually. Yeah, I could ran about Descartes for hours. But if you really want to hear me mention Descartes and Russo in the same sentence, then my head will explode. Well, we don’t have-

Michael: That’s keep you alive. Yeah.

Jim: It keeps me rolling. Ah, damn it, right? We don’t have a lot of time left, but let’s turn into a final topic. One of the essays that Michael sent me as a primer for this show was about the importance of noise. And again, with this issue of evolution, whether it’s Darwinian or whether it’s other forms of open-ended novelty search or whatever, noise plays a big part. Put it over to Michael to give us a gloss on how noise and artistic creativity somehow relate.

Michael: Yeah, I’m glad you took it here because I was just about to bring this up. This is an essay called The Future is Noisy that was just published at and is a draft of what I hope is the last chapter in this book I’ve been drafting in public for years now called How to Live in the Future.

Michael: And this piece, again, it’s not a huge jump in topics really from what we’ve been discussing here because so much of what the two of you have just been talking about has to do with this question of the ways that machines do or do not utilize noise the way that biological systems utilize noise, at least yet. And some people working in machine learning now do. I mean, there’s an entire discipline called reservoir computing, where for Michelle Girvan gave a great talk about this a few years ago that’s available to look up on YouTube. I think it’s called Harnessing the Unpredictable. It’s a community lecture you can find.

Michael: But at any rate, the talk was about how predicting these famously chaotic and indeterminate systems like weather, that process is improved by not just training the neural network on the past behavior of that system, but injecting noise into the algorithm to keep it from overfitting to its training data. So the world is always going to defy. When, J.F., you and Phil on Weird Studies, talked about the capital are real and how that’s different from the lower case are real, that the real in that sense is constantly challenging our models.

Michael: I know, Jim, you’ve had Nora Bateson on the show and she likes to talk about warm data and she likes to philosophically attack the premise that we can ever really adequately model the world, this distinction between map and territory. And we double the predictive, horizon of these weather prediction algorithms by training a camera on a bucket of water, kicking the bucket and letting the pattern of waves enter into the machine learning process. So the machine is looking at something that behaves a lot more like convection cells in the atmosphere.

Michael: And this is something that is done a lot. I talk in the essay about Andreas Wagner and his work on evolution and the way that the adaptive fitness landscape across which organisms are always moving where you’re trying to optimize. But of course, the landscape itself is so turbulent and what is easy and what is hard is always changing, and it’s an uncountable mini dimensional kind of thing that we’re describing here. But that when you talk about challenging optimization, I think about … And in this essay, we talk about play and the way that play is the way that organisms who can only imperfectly inherit information about the environment through their genetic lineage are then able to fine-tune their instincts and their expectations and so on for an environment that’s different from the environment of their parents.

Michael: I think that there is this sense in which if we really want to talk about machines that are creative machines that issue some discernment, machines that are capable of true novelty, then, in a weird way, we come back to the question of the role of the oracle of oracular thinking in the process of science. J.F., when you and Phil had Joshua Ramey on Weird Studies to talk about how there’s something of the oracle in the moment of the decision of a scientific hypothesis, the rest of the thing is all very rational. But choosing what question to pursue or what question not to pursue is a moment that revolutionary scientists, down the table, always attribute to a flash of inspiration that they got after taking a walk or a shower or a nap.

Michael: And it’s not, in some way, the way that we navigate the fitness landscape of possible fruitful scientific questions or investment decisions or whatever invites the irrational, invites noise back into the rational process. And so this is … Yeah, that’s where I see this really figuring here.

Jim: Very good. Yeah. Then, obviously, in the scientific process, that leap of a hypothesis formation is a classic example. In fact, we’re over our time. That’s all right. It’s been a fun conversation. I’m going to turn it over to J.F. for some final thoughts on these themes or whatever else.

J.F.: Well, geez, I feel like we just touched the surface, obviously, of all the issues involved here, and we didn’t talk about NFTs and the other business, yeah.

Jim: Fucking stupid ass topic, in my opinion, but that’s all right.

Michael: We didn’t talk about technological unemployment, unemployability.

J.F.: Right, that’s another thing-

Michael: Which I think is a key piece here.

J.F.: Yeah, the way that this pushes the human again to the periphery and seems to … So there’s all kinds of dangers involved here, but we have to remember … I think Heidegger was right. The great danger of technology contains what he called a saving power. We have to see what the affordances are here and renegotiate our situation.

J.F.: The problem is that in our culture of absolute distraction, and I think that distraction is more than just a moment-to-moment problem, if there’s a metaphysical issue here going on with how time is running away with us and almost a severing of some tether that tied us to time that’s happened. We have to be able to return, I think, to a kind of embodied existence, which is less and less obvious to us in the day to day because of our dependence on digital tools of all sorts.

J.F.: So I’m really a big proponent of those books and thinkers who are arguing for a return to the organic. I don’t like that word, a return to the analog, let’s say.

Jim: Yeah, there you go. That’s the right word.

J.F.: And a … But not in a Luddite rejection of the digital, but if only to find our bearings again in the horizon world of embodied finite existence, such that we can then discern what is going on in the digital oracles that we’ve created. And I think that this type of discussion is what is going to allow us to do that.

Jim: Yeah, I absolutely agree. That’s one of the reasons I live in the mountains of Appalachia, not in Silicon Valley. I don’t want to be surrounded by all that goddamn conventional wisdom. It’s also why I used to take six months of sabbatical from social media each year. This year, I ratched it up to nine months, right?

J.F.: Wow.

Jim: Listening to people, started April 1st. We’ll run through January 1st, 2023. Listening to just that shit blah, blah, blah, blah, blah, blah, blah. The war for our attention is just not good for doing serious thinking, goddamn it, right?

J.F.: It’s not.

Jim: So anyway, gentlemen, this has been a wonderful conversation. And as you both indicated, we just tipped our little pinky into a very deep pool here. And that might be fun to have you guys back and continue this discussion at some point. So J.F. Martel and Michael Garfield, thank you very much for an appearance on the Jim Rutt Show.

J.F.: Thank you, Jim.