← Gom Jabbar: A Dune Podcast
Loading summary
Transcript138 lines
- [00:00]
Leo
Today on the show. We are back, dear listener, in the Dune themed classroom. Oh, there's a chair dog plushie in the corner. Everyone, everyone is chugging sappho juice. And the professor, I'm pretty sure, is well on his way into turning into a worm over the next 3,500 years, give or take. In other words, I'm exactly where I want to.
- [00:43]
Abu
Welcome to Gom Jabbar, your guide to the iconic world of Dune. We'll be exploring the themes, philosophies and characters found in the sandy depths of this vast universe, from Frank Herbert's groundbreaking novels to the adaptations on film and tv. My name's Abu.
- [00:58]
Leo
My name is Leo.
- [00:59]
Abu
And folks, welcome back to part two of our discussion about Dune and artificial intelligence and what lessons we can learn from Frank Herbert about our relationship with technology today.
- [01:12]
Leo
Yeah. Now, before we get too far into it, let's make sure shout out Mapes is smiling ear to ear. And let's take care of our housekeeping. And first up, spoiler warning. Today's episode contains like the last episode, non plot, but light spoilers through heretics of Dune. So we just for the best experience, especially because we're going to be referencing conversations that we had as part of our book club. Just make sure you've caught up with our book club. If you are up to date with the book club, you're good to go and should be in a good place to enjoy the conversation today.
- [01:48]
Abu
Absolutely. Now, of course, up top, a huge shout out to our Kwisatz Haderach level patrons, Daniel Dion and Seth Redding. Greer, folks, yeah, I'm not saying I want this to happen, but if we could just take your personalities and kind of train an LLM with that maybe and turn that into an artificial intelligence tool that I could use day to day. That's exactly what I want. That to me is the ideal transformation of an AI is based off of our Kwisatz Haderach level patron personalities. That's an incredibly weird and Tron like way of saying thank you to our Kwisatz Haderach global patrons. Are you down to upload your consciousness? Let us know. Of course, our heartfelt thank yous also extend to all of our patrons at every level who do truly build the foundation upon which the show runs. And we literally folks, literally couldn't do this without you.
- [02:47]
Leo
It makes such a big difference. And if you're not in a place to become a monthly supporter on Patreon, we have a one time tip option. The link is in the show notes. You can buy us a coffee. All right, folks, let's talk about today's episode. Now, we're always discussing themes in the Dune saga and how Frank, when he was writing the books, very clearly was looking at the world around him and saying something with his writing, right? He talked about ecology, he talked about politics, philosophy and technology. Right. In a universe that has the Butlerian jihad, Dune has a lot to say about technology. So today we're going to take that mentality, we're going to put on that hat, and we're going to use ideas about artificial intelligence present throughout the books to have an in depth discussion about our modern day relationship with AI. This conversation's one part Dune, one part what's going on in the world of AI and especially today. There's a lot of stuff that I hadn't looked into and now I have.
- [03:49]
Abu
I think we've both learned a lot over the course of these two episodes about artificial intelligence.
- [03:54]
Leo
How kissable Claude is. Get over here, Claude.
- [03:57]
Abu
So kissable.
- [03:58]
Leo
So kissable.
- [04:01]
Abu
All right, folks, well, let's take a quick break. Don't go anywhere. Buckle up, because when we come back, we're diving into two more important lessons about artificial intelligence that we can learn from Dune. We'll be right back. Are you feeling more fulfilled now that.
- [04:16]
Leo
You'Re back to work?
- [04:19]
Abu
No, I need a vacation.
- [04:22]
Leo
See the movie that critics are saying is an awesome look at that crowd pleasing, fist pumping, all out brawl of a film.
- [04:31]
Abu
You're right about that. They're coming after our family. Fixed it.
- [04:38]
Leo
Oh my. Nobody 2. Rated R. Only in theaters now. Welcome back, everybody. Hope you're ready to learn some stuff. Maybe you know all of this stuff, maybe you don't. But we're going to be talking and this is going to be lesson number three that we can learn from Dune. LLMs and other memory from Dune cannot offer new solutions. It's an important lesson.
- [05:10]
Abu
Interesting. Yeah.
- [05:12]
Leo
Now, throughout the books, Bene Gesserit reverend mothers rely often on their other memory, right? They're tapping into genetic memories of their past lives, sometimes going all the way back to old Terra and to experiences they may have had 20,000, 30,000 years ago. Now, on the one hand, you might expect this to make them the wisest, most infallible characters, right? But for all of the past that the Bene Gesserit can remember, do they successfully avoid repeating the patterns of the past? And I would say no. Basically, the entirety of Heretics of Dune explores that question, and I think pretty clearly points toward the many pitfalls of relying too heavily on Other memory, but also like dogma and slow adaptability kind of conservatism on a massive scale. Right?
- [06:10]
Abu
Yeah.
- [06:11]
Leo
Afraid of the unknown, you're close minded to new ideas. That sounds like stagnation and that sounds like something Frank was very actively warning against. And in particular, as we were researching the last episode and then of course this episode as well, the parallels between LLMs and over reliance on them and other memory started to become harder and harder to ignore. So we just wanted to carve this chunk out of this episode to look at that overlap and basically what Dune can teach us about our pals and Ixian thinking machines. Claude, Gemini, Grok, all of them. There's a French one that Eleanor mentioned to me yesterday, and I don't remember the name of it, but it sounds very cool and it is French, so.
- [06:58]
Abu
I might like it. Yeah, maybe that one is most kissable. Probably we'll release a final kissability ranking at some point, you know, once we've completed all of our research. So to start off, actually, let's define what an LLM is for folks who might not know the actual definition. And then let's draw this parallel between other memory and LLM. We want to be very clear and explicit about this because this idea basically underpins the rest of today's discussion. We're saying that other memory is analogous to LLMs. And what we learn about other memory in the Dune books can also be applied to modern artificial intelligence.
- [07:39]
Leo
Yeah.
- [07:40]
Abu
So to begin, LLMs, or large language models, are basically these predictive models that try to simulate intelligence or thinking. And these models are trained on large data sets. IBM explains that through this training, they learn to predict the next word in a sentence based on the context provided by the preceding words. The model does this through attributing a probability score to the recurrence of words that have been tokenized, broken down into smaller sequences of characters. End quote.
- [08:17]
Leo
Yeah.
- [08:18]
Abu
So it's just lots and lots and lots of data used within context to predict an outcome, to predict what the next thing should be.
- [08:26]
Leo
Right?
- [08:26]
Abu
Now, the connection we want to draw to other memory is pretty clear when you think of it in that way. Reverend mothers, in a sense, have access to LLMs in their brain. Right. In other memory, they have this resource that they can tap into for answers. And this resource is theoretically an infinite genetic data set. Right. Going back tens of thousands of years. And other memory can theoretically provide infinite knowledge based on context and prediction, just like an LLM does. Now, of course, there's a lot of caveats here and there's a lot of nuance that we're going to get into. But the thing to really keep in mind is not to equate an LLM with an actual human brain. An LLM doesn't know anything technically. And as convincing as it can be to talk to a chatbot and have it talk back to you in a very human way, experts are quick to remind folks like you and I that these are just at their core, these are just predictive models. They are using data to probabilistically address whatever prompt we might type in. And that's all that's happening behind the scenes. It doesn't know anything. It's not sharing any knowledge that it inherently knows.
- [09:44]
Leo
Yeah, so LLMs don't know anything. And I think that's important. But we also, we see this difference play out in Dune between knowing a thing and like capital K knowing a thing. Right?
- [09:57]
Abu
Yeah.
- [09:58]
Leo
And this is something we see with a lot of the characters who have other memory. But in particular there's this example of Leto too, playing his baliset in Children of Dune. He sensed the absolute idealized summation within himself, which knew how to play this balliset, though nine year old muscles had not yet been conditioned to that inner awareness. End quote. So all of this is to say, just because Bene Gesserit reverend mothers have 40,000 years of human memories or more and experience, doesn't mean each Reverend Mother has all of the answers and moves through life without making any mistakes at all, right?
- [10:40]
Abu
Oh yeah.
- [10:41]
Leo
Oh, they're making mistakes. They're making a lot of mistakes. Another kind of clear example of this is in Children of Dune, in particular in regards to Alia and abomination. As we discussed in depth in our book club episodes, the Bene Gesserit have been fucking up for generations, thinking that the only solution to abomination was to kill someone at the very suggestion of it.
- [11:08]
Abu
And oftentimes a child, to be clear, because symptoms can arise very, very early.
- [11:14]
Leo
It's infanticide. Literally, they're killing born humans just because they're afraid of this thing. Now, to be fair to them, we see how awful it can be with Alia. But Leto and Genema, the non Bene Gesserit outsiders, discover two solutions to the abomination problem. Here's the explanation from Leto. Quote, Gani and I formed powerful internal alliances with ancestors who followed the pharaonic model. They helped us form a mingled identity within that long dormant mob. So we see that the Bene Gesserit have been missing out on like this whole possibility because they've been so careful. Again, going back to the idea of being overly conservative and relying too heavily on known knowledge and not being open to experimentation. Right?
- [12:11]
Abu
Yeah, exactly. And I think it's notable that Leto and Ganima are outside of the Bene Gesserit fold, which means they are outside of Bene Gesserit dogma bias. And to use today's analogy, the data set that the Bene Gesserit have been working with for tens of thousands of years, Leto and Genima, are outside of that, and thus are able to come up with solutions that the Bene Gesserit themselves have never been able to.
- [12:38]
Leo
Yeah.
- [12:39]
Abu
So let's bring this back to artificial intelligence and our relationship to AI today, because there's a couple of interesting details here and some nuance to unpack, like other memory. It does seem that the current iterations of LLMs, and again, we're sort of emphasizing the current technology that exists as of today's recording. This could all literally change tomorrow. Right, but the current iterations can't discover entirely new ideas, given the fact that they are trained to analyze and regurgitate existing data sets. Right. They scour data that already exists. So to come up with something that doesn't exist falls somewhat outside of the scope of current LLMs. Now, of course, you could argue with me on semantics there, right? Because what exactly is a new idea? And how do humans discover new ideas then? Right? Don't we basically do what LLMs are doing? We examine what we know. Let's use, for example, a hypothetical of LEGO pieces, right? We know all of the different LEGO pieces that are currently available to us, and then we try and find patterns for those LEGO pieces. We connect them together in various ways, and by smashing them together in different enough ways and combinations, we might be able to build discover something entirely new that has never been built before. Isn't that basically what an LLM is doing? Taking a bunch of info and finding the patterns and recognizing different connections between them and spitting out something entirely new? Perhaps. Aren't all new discoveries built upon old discoveries and old knowledge? So what is the difference between human thinking and LLM thinking?
- [14:32]
Leo
Yeah, especially because we haven't fully decoded the process by which we come up with new ideas and inspiration. Part of this is challenged by the fact that we use a very organic process in training LLMs in this sort of like, survival of the fittest method, where you say, well, this model keeps pumping out bullshit that no one likes, so let's iterate on it and let's give it this new internal logic. And now you're getting better things. But what is the thing? It's still a big mystery how our minds work. And the more you look at how LLMs work and the stuff that they're able to do, I understand those delineations becoming a little bit blurry and going, ah, shit. Well, I guess a good definition of what it means to come up with a new idea is a little hard to do, other than you just happen to bump two neurons together. You're like, whoa, I should try. My wife uses peanut butter and cream cheese on her bagels. And I'm like, that's an LLM bullshit decision. Those flavors don't belong together. But whatever, it's fine. She loves it. And apparently it's quite good.
- [15:42]
Abu
She smashed those two data sets together and came up with something new and disgusting, but new nonetheless.
- [15:50]
Leo
Yeah, yeah.
- [15:51]
Abu
And actually, you found a great quote from Psychology Today that I think kind of illustrates and summarizes what we're trying to get at here. It is getting harder and harder, especially as LLMs are getting more and more sophisticated to kind of parse the difference between how humans think and human knowledge versus how LLMs are thinking. This is a quote from the Psychology Today article. As we integrate machine inference more deeply into our systems of understanding, we're forced to ask, is knowing really about reflection, belief, and justification, or is it simply about the ability to generate useful, accurate patterns? End quote.
- [16:34]
Leo
Right. In my sort of stupid understanding of LLMs, you're training it on this huge data set of all grammatically correct sentences. And so it's going to know how to put out probabilistically grammatically correct sentences, and if it happens upon a grammatically correct sentence that contains great wisdom or is, like, very useful for someone to help them, is that better? You know, I'm glad I found that Psychology Today article, because I think it is. This is where it becomes more, like, epistemological. This is the idea of, like, what is truth?
- [17:06]
Abu
Yeah. We're dipping somewhat into philosophy here. Right. We're asking, like, what is thinking? What is the definition of new? How do we define this amorphous concept of a new thing, a new idea?
- [17:18]
Leo
Yeah.
- [17:19]
Abu
And how do we arrive at it? We are dipping into some philosophy here, too. And a lot of the questions we're asking are still open to debate. Right. We're not presenting any definitive answers. A lot of these are rhetorical questions. A lot of them are questions that still need more thought. And actually, I think, sticking with your example, I really like your example of the grammatically correct sentences with the LLM. Because you might also argue that some of the best writing out there intentionally breaks the rules. Right? This is why jazz is so interesting and fun is like, it breaks the conventions of quote unquote, how music should be made in ways and plays with that. And some of the best and most interesting published works out there break grammar conventions and writing conventions and do weird things. You know, there's. Which book is it? Is it House of Leaves? There's a book that's like written sideways in some certain points and you can actually read it backwards and it still makes sense. You can like start at the last chapter and go backwards and it's still, you know, it breaks so many writing conventions. It might not be House of Leaves. Somebody in the audience correct me, email us. And so is that a limitation of the LLM? You could argue yes, but you could also argue no, because you could tell the LLM in your prompt, write me something that breaks traditional grammar conventions and does X, Y, Z and achieves this goal. And because it knows what a traditional grammar set data set looks like, it can then not do that. Right. It can also do the inverse of what it knows or at least attempt to do the opposite of the data it's been provided. So, you know, there's a lot of interesting questions here on how LLMs still operate and what they can and can't do. I do not think we have yet to find the boundaries.
- [19:07]
Leo
Yeah. This also led to the exploration Exploitation dilemma, which initially I thought, because I'm not educated on this stuff, I thought that this was like specific to LLMs, but it turns out this is more like macrobiology as well. But the idea, the exploration Exploitation dilemma, is this balance of how much are you exploring outside of known tested good methods and how often do you lean explicitly on exploiting known successful methods. Right, yeah. And LLMs, when you are calibrating the language model, you have the ability to say, I want you to lean less on what is probabilistically most likely thing that you know is going to turn out good results. There's like a proximity measurement of how far each token is from themselves in these massive clouds of data. So you can do that. This also was good for me to learn because I've definitely used in the past, maybe a year or two ago, I used GPT to provide a list of things. And it's funny how a couple of the things were exactly the thing I said. I was like, can you give me examples of this? And then some of them were just like way fucking out there. And that seems to be this sort of balance of exploration exploitation. And while with the exploration emphasis, responses are more likely to be novel new, like you're saying Abu, you can break grammatical conventions and be like, wow, that's weirdly really profound and beautiful and wrong in a way that stirs my soul. But that also means that they're more likely to be nonsensical, unuseful, bad, and all of this Again, we're jumping back and forth between LLMs and Dune, but looking at the Bene Gesserit, they have been trimming their genetic tree for thousands of years. Does this mean that there are glaring blind spots in other memory? Have they been completely shaving off these huge chunks of data sets, meaning that their results, when they Tap their internal LLMs, are their results built on data sets that are incomplete, fundamentally and intentionally, but in a way that causes problems? And it seems more and more likely the more we've talked about Dune, and with this research into LLMs, it seems more and more likely. Yeah, it seems like they are not going to be finding optimal solutions if they're not taking into consideration the breadth of the human experience if their goal is to improve human life.
- [21:47]
Abu
Right. And we'll talk a little in a little bit about how more data doesn't just equal better results every single time either. There is like sort of a law of diminishing returns there. Put a pin in that. We'll get to that in a second, folks. I did want to pause here to maybe share our own thoughts on this and our own opinions based on the research we've done today, because the question still lingers. Is an LLM capable of coming up with a new idea for me based off the research we've done today for this episode and what I learned over the course of making these two episodes for the podcast? I think actually the defining human characteristic that LLMs lack isn't necessarily the ability to formulate new and novel ideas. I do think they can come up with patterns and connections that maybe no one else has before, just because of the vast amount of data it's able to parse. I think what LLMs lack, though, is the ability to be curious, the ability to be imaginative, and the ability to have initiative. I really liked. I was googling around and I was reading a number of like blog posts from folks who write a lot about AI, and I stumbled upon one from Sean Gadecki, who's a software engineer at GitHub, blogs a lot about AI. And Sean wrote in a May 11 post about why he thinks AI hasn't discovered anything new yet. And I'm not saying Sean has all the answers, but I am saying that I like the way Sean put it here.
- [23:21]
Leo
Sure.
- [23:21]
Abu
And I think generally I agree with what he said. One answer would be that eureka moments don't happen during self reflection. The original eureka moment occurred during an accidental physical experiment in a bathtub. In general, big ideas might come from actually doing experimentation in the real world and seeing what is surprisingly successful. It might therefore be unfair to ask a model to go away and think up new ideas in isolation. End quote.
- [23:52]
Leo
Good golly. That's so good.
- [23:55]
Abu
I like that a lot.
- [23:56]
Leo
I had read this in the script, but hearing you say it out loud it is true. I've been on this soapbox the last couple of years, but the idea of language, all that in a word, lord, is limited and has its limitations. Language doesn't measure up to experiential reality. When you have a feeling and then you go, oh, I'm feeling sad, that is such a violent simplification of the actual experience of sadness or sorrow or melancholy or whatever fancy words you want to paint on it, that only scratches the surface. And then you say this predictive model only has those limited tools that can only express and explore within articulatable language. You're out in the world, you're having these experiences that cannot be limited to articulatable language, and that may be the foundation of a new thought or a new belief or a desire to break the rules. In art, you break that grammatical thing to get people to an experience that you had. And all of that is inaccessible to LLMs at this moment, as of his May 11, 2025 post and our July 25, 2025 recording. But yes, that's such a good point. That's cool. I hadn't thought about that.
- [25:15]
Abu
Yeah, I really loved how Sean put her in this blog post and I think I agree. You know, like I A basic example that I can think of is Isaac Newton didn't just go, oh my God, gravity one day. He didn't just wake up and go oh my God, gravity. Famously what happened is that an apple from a tree fucking fell on his head. There was an experiential physical feedback loop that happened in the world which then opened up certain neurons in his brain. And based on perhaps what he already knew, the knowledge he might already have access to, combined with the real world physical feedback of a thing that happened, combined with more experimentation, I'm sure. Right? I'm sure he dropped Many more apples in order to continue developing an idea of like, oh, there's something here that hasn't been put into words yet. Gravity. The idea that all things are pulled toward the center of the Earth. So I really like this idea, and I do think LLMs have this limitation. At the moment, an LLM is unable to receive constant physical feedback from the world around it, which I think is an important part of human curiosity, of human imagination. And ultimately, it is the initiative to do something like throw an apple that might lead to a discovery of something like gravity. And I think LLMs are still incapable of doing this. So that, to me, feels like a boundary that, again, in their current iterations, we are running up against and thus would limit an LLM from discovering something entirely new and undefined.
- [26:55]
Leo
Yeah, that's a great point.
- [26:57]
Abu
What about you, though? Where's your head fall on this question of LLMs ability to come up with new ideas?
- [27:04]
Leo
I mean, what I'm getting from you is that we should be throwing, I think, throwing fruit at LLMs, like throwing apples at LLM. Yeah, I agree. I think that's what Gidecki was talking about. No, I completely agree, especially with that new kind of attachment of that idea of you're out in the physical world and you have these experiences that if you were to articulate them, they might not have a clear connection, but experientially have a connection. And that can be the foundation of your jumping from one idea to another. LLMs can, probabilistically, in a familiar situation, do something new or coincidentally interesting. But we clearly have that edge of our thought process, doesn't necessarily limit itself to articulatable language. And we can imagine, and we can yearn. We have a fire in our souls, but also we can imagine with a direction or with an intuitive understanding of what can exist in this world or not. I also cannot help but bring up this moment in God. Emperor, Right. Leto, too, is talking to, I believe it's Malky at the time, about the false promise of memories. And I think this has a place within this conversation where we get so built up on, like, what does it mean to have those data sets? And how much should we revere them? The present is distraction, the future a dream. Only memory can unlock the meaning of life. Aren't those beautiful words, Malky? But my words were a foolish lie. I knew it at the time, but I was infatuated with the beautiful words. No memory unlocks no meanings without anguish of the spirit, which is a wordless experience. There are no meanings anywhere End quote.
- [29:06]
Abu
Wow. A wordless experience.
- [29:10]
Leo
I think it's so beautiful. But also, I think the memories that we choose to cling to in the present moment or the memories that your present consciousness serves up to you, you remember that embarrassing thing you said 15 years ago? That is a reminder to do better next time and an encouragement to be better and be kinder the next time you're in that situation. So I think there's value in what is brought up to you in the moment, but it's where you fixate on it, or you let that memory kind of rule how you're living now and how you intend to live moving forward. I think that's where it becomes destructive. And I'll wrap up by sharing another quote. And this is from a weirdly long, like, I guess this is from an Instagram account, but I've seen it, like, animated, and I've seen it in other places. Quote. The memories worth keeping are the ones that gently inform your future, not the ones that hold it hostage. Let the past stay where it belongs. Today needs you more. End quote.
- [30:17]
Abu
Wow. That is beautiful. Wow.
- [30:19]
Leo
Okay, let's land the Sornithopter. Let's talk again. Let's kind of bring the lesson home.
- [30:25]
Abu
Yeah.
- [30:26]
Leo
In summary, dear listener, the Bene Gesserit have what could be considered, could be explained as the ultimate large language models built into their brains as part of other memory. Millions of lives and tens of thousands of years of human experience that they can tap at any time. But are the Bene Gesserit utterly free from mistakes? No, of course not. And also, while they're not the perfect role models in Dune, we can learn an important lesson from them, and that's why we're here today. The Bene Gesserit, they weave their plans within plans. They tap their other memory, but they don't allow themselves to succumb to the temptation of living forever in those memories. Those memories are not the point. Instead of those who have other memory, it's Leto, too. It's Reverend Mothers like Tar and Dar who pave the Golden Path by utilizing it as a tool for moving forward, for gently informing their futures. So, dear listener, please stay heretical. No matter how kissable Claude is, don't lean too heavily on him. Keep that kind of human fire alive.
- [31:36]
Abu
There you go. All right, after that heady lesson, let's take a quick break. Let's compose ourselves. We have one more lesson we want to talk about. Lesson number four that Dune teaches us about artificial intelligence. And we'll get into that right after A quick break. So we'll see you in a minute, dear listener. Mama, Papa.
- [31:57]
Leo
Mi cuerpo Crece yun rinmo alarmante y la ropa. Trip planner by Expedia. You were made to outdo your holiday, your hammocking, and your pooling. We were made to help organize the competition. Expedia Made to travel. Welcome back, everybody. Hope you're ready for our final lesson that Dune can teach us about AI and LLMs. And lesson four is that the most impactful art will always need the human touch.
- [32:47]
Abu
Oh, boy, here we go. The art conversation.
- [32:49]
Leo
I know I'll have to. Oh. I have a degree in art. Breathe, compose myself. As we talk about art. Stick to the bullet points, Leo. You've got this.
- [33:03]
Abu
You've got this, buddy.
- [33:05]
Leo
Thanks, dude. I probably don't. So the idea. Oh, no, he's already off the rails. All right. He's off script already. So the ideas around other. That's funny.
- [33:30]
Abu
Chatgpt, can you just read these bullet points in Leo's voice, please?
- [33:35]
Leo
Yeah, that's where we need to train up one of those voice models. Okay, so these ideas around other memory and LLM's ability to create also tie into our fourth lesson. And I think this lesson is best examined through this fascinating moment in Heretics of Dune. Truly one of those, like, most memorable little moments in the book.
- [33:56]
Abu
Yeah.
- [33:57]
Leo
Where Darwi o' Drade is thinking about an ancient painting in Tarazah's office. Morning Room. Like bedchambers, it's in one of Taraza's rooms, and the painting is Thatched Cottages at Cordville, created by Vincent van Gogh in 1890, just a few weeks before his tragic death. And if you haven't seen it, dear listener, it's a simple painting of cottages along a stone path near a hillside made in Van Gogh's iconic, frenetic, swirly, painty, impressionistic style that many of us know from his probably most famous painting, Starry Night.
- [34:40]
Abu
Right.
- [34:41]
Leo
And o' Trade is deeply touched by this painting, not only because it's been preserved for so long, it's this direct connection to 1890 AD but also because of what it makes her feel when she uses the Ixian technology that reproduces each brushstroke for the brief space of that replay. She always felt totally human, aware of the cottages as places where real people dwelled. Aware in some complete way of the living chain that had paused there, in the person of the mad Vincent Van Gogh, paused to record itself. End quote.
- [35:27]
Abu
Ugh. So beautiful.
- [35:30]
Leo
God damn it, Frank. Just good writing. Just Beautiful writing, truly.
- [35:37]
Abu
And I think what's interesting is when you unpack this scene, actually, it is so thematically resonant with the story taking place in Heretics, because this moment where Odrade is thinking about the painting she's remembering her experiences with it is a reminder of the flaws in Bene Gesserit thinking in the often, like, cold, detached, manipulative logic that they operate with, especially by the time we are in Heretics.
- [36:05]
Leo
Yeah.
- [36:06]
Abu
Van Gogh's work, famously. I actually went to the Van Gogh Museum this year on a trip to Amsterdam. It was. I cried like it was truly. Our tour guide was so good. The tour guide took us through the whole museum and Van Gogh's entire life. This is a total aside. I cried during the whole thing, like it was absolutely beautiful. A highlight of my trip. But anyway, Van Gogh's work is, like, famously expressive and instinctual. Right. It's packed with emotion, as o' Drain is recognizing here. And you could argue that in some ways, this one painting, quote, unquote, simple painting, contains more information about humanity and humans than 30,000 years of other memory does.
- [36:49]
Leo
Yeah, true.
- [36:50]
Abu
This LLM in their brain doesn't capture everything that this photo does. And of course, we've talked about how the Bene Gesserit often use other memory and have now done so by the time we're in Heretics for tens of thousands of years. But this painting represents a different kind of memory. It's an artistic and emotional amalgamation of memory, rather than just a purely genetic and logical one. It actually serves as a reminder to a drade. Right. It's showing her that memory isn't just a long string of data and lineage in her other memory history. And memory is also about feelings and perceptions, those inarticulable things that are part of the human experience. Those things that, as we discussed in our lesson in the first half of this episode, an LLM can't quite capture.
- [37:51]
Leo
Yeah. Or putting it another way, you have the poem, you have the output of what the poet wrote, but the feeling of choosing each word and feeling yourself from one thing to the next. Right. Each brushstroke of the painting, you can capture the painting as a encapsulated expression of something, a desire to show someone something or highlight something that you think is beautiful in the world or ugly in the world or something, but the actual present experience of each brushstroke and that being the thing that so moves her. That, to your point, she thinks back on that frequently, even when she has over 30,000 years of human experiences at her fingertips. She's thinking literally about those brushstrokes one at a time. That's so cool.
- [38:40]
Abu
It's a really powerful moment in Heretics. And I couldn't agree with you more. It's one of the things I always remember about the book, even though it's such a small section. It's just a couple of paragraphs in a very long, intense book. But it has always stuck with me. And I think for today's conversation, and especially in this conversation about art and AI, it's extremely relevant because the reality is that a giant set of data and algorithms can't capture the feeling behind creating art. In fact, in a Stanford article that I found from 2023, it argues that LLMs lack creativity and emotion, two of the things that are necessary for art. This article was talking about Dall? E, which is that image generation tool. And it says there are a couple of things central to human creativity that. That Dall? E still lacks. For one, it has no emotional evaluation of what makes an image important, novel, funny, or meaningful. Perhaps related to this is that Dall? E has no extended artistic process. For human artists, that process is central. Trying things, evaluating them, iterating to the next version or the next idea to ultimately discover or zero in on the artist's impetus for making the art in the first place. End quote.
- [40:03]
Leo
I also. I really appreciate that because I think that touches on something that people fundamentally misunderstand about art, for people who do art all the time, who are artists, but I think we're all artists, just people who are doing art all the time. It is not about each individual painting and drawing. It is about the process. It is about the accumulated skill and being. That is the thing. And ultimately, if I draw something or paint something, and if it's a digital work of art, I'm like, okay, and then I delete it. Because it's not about that. It's about the process of being and sitting and the. Yeah, evaluating, trying, iterating. That's the core of that process. And I'm like, not only do my fellow man, my fellow humans sometimes not understand that, it's also. That's clearly something that is lacking from the LLMs as well. Or at least it seems that way.
- [41:01]
Abu
Yeah, certainly. And again, there are arguments to be made that, like, what if I just type into an LLM, draw me a sad painting in Van Gogh's style.
- [41:11]
Leo
Right, sure.
- [41:12]
Abu
You know, like, it could probably use stolen works of Van Gogh, honestly, that it's been trained on to replicate that style, and then probabilistically associate things that are referenced in data sets as sad. So, like, it might show you, like a starving cat next to a cottage or something, you know, sticking with our Van Gogh theme here. And yet you might just look at that and be like, yeah, that's objectively sad. Because, like, probabilistically this LLM was like, these are the things humans associate with being sad. Because that's what my data set has taught me is sad. But again, there was no process there. There was no intention, there was no. The LLM didn't approach this art with any sort of creativity or emotion. It was just making probability analyses and connecting various dots together to churn out something that meets the prompt that you have put in.
- [42:10]
Leo
Yeah. Like, with a critical eye, you're looking at that painting of that starving cat and you're going to. There isn't really a lot of specificity. It's just like a. It looks like other things I've seen. Also, the cat's got, like six legs and a human ear.
- [42:22]
Abu
Right. Not to mention the hallucinations that are often common. Like, getting hands correct is definitely an issue that many image generators have. But I think, like, I want to acknowledge, like, there's also the argument that, like, maybe that's all I wanted. Maybe I just wanted some, like, cheap ass sad cat photo to be churned out because I need to, like, slap it onto a PowerPoint presentation for work or something. First of all, I don't know what the fuck you're presenting at work, but you do. You. But I think, like, the argument that I think you and I would both make here is that there isn't value in that thing that's been made. There isn't value in this, like, lifeless, most probabilistically sad, dying cat photo that's been generated. And I think that's what I would personally argue makes it sort of a lifeless, valueless piece of AI slob at the end of the day.
- [43:17]
Leo
Yeah. Again, it goes back to what we were talking in the first half of this episode. You have this inarticulatable experience, something about that moves you in a way that you can't define. And sometimes part of the process of creating the art is trying to figure out how to articulate that thing or finding out what. Like, it's an exploration and it's an experimentation. It's not just. Certainly I have this idea of a painting. This is the painting. And then, okay, cool. The nouns that I wanted to be in the painting are in the painting. I've achieved painting. Good job. Me that's not the process. But of course, with all these things, like the board members, the team at the desk, they aren't necessarily looking for what is the most human thing we can do. They're looking for what is the most profitable thing we can do. And there's of course, that John Oliver piece on AI slop and the economics of it. Also just the number of. I mentioned it in the last episode. The number of TikTok accounts that are fully just stolen human stories but recreated by an AI synthetic avatar talking about it with a synthetic voice. And those are making a lot of money on TikTok because you can make 15 of them a day, really, really easily. And then it's like we need systems in place to stop that kind of stuff from happening.
- [44:40]
Abu
Yeah. And look, this might be controversial actually within these circles, as in the circle of you and me, but I do want to acknowledge something that is an interesting use case of the tool. And I don't need the LLM to have feelings. Right. I don't need ChatGPT to create high art that goes right alongside Van Gogh's paintings in the Van Gogh Museum or anything I like. I don't want it to do any of that. To be totally clear, I'm not asking for an LLM to feel emotion and have creativity like humans do, but I think there are ethical considerations of when humans use AI to create any type of art and kind of shortcut the process. As we read from this Stanford article, like, for human artists, like, and as you stated too, like, the process is central to creating, to achieving the desired outcome and the desired outcome is oftentimes human expression of some kind. When you shortcut the process using artificial intelligence, I think that devalues the end product in a way that I find pretty disgusting. But ultimately, like, I do think there are some interesting use cases for people using it.
- [45:51]
Leo
Yeah.
- [45:52]
Abu
That don't also, like, degrade from the wider sort of like fabric of art that's out there. I find. I actually find it fascinating that there are like entirely AI influencer accounts that are fake. I don't think it's inherently a good thing, but it's an interesting use case of the tool.
- [46:10]
Leo
I think if we were to again, and all we can do is look back on data points that we have and hope that maybe give us a sense of what's to come. But like the advent of digital art and photoshop in the 80s, the 1980s into now we have things like procreate that allow. I have literally taught classes to kids on how to draw and paint. And the ability to create digital layers as an artist is such a handicap that did not exist when you were just a painting on a canvas 200 years ago. And realistically enables people who were not able to do that thing that Bob Ross does in a 22 minute television episode. It allows children to do something that is not maybe as convincing and incredible and it's not on a canvas in the physical world, but still looks aesthetically nice. And people will go, oh wow, that's really cool. And I've legitimately seen 13 year olds, 14 year olds put out art that I'm like, this is beautiful, this is really incredible and it moves me and it's awesome. And could you have done this if I gave you acrylics and a canvas? Maybe not. And so similarly, I think these AI tools and these generative models if taken and processed. As fucking obnoxious as it is. When people are like, I'm a prompt engineer and my whole job is engineering prompts, it's like, shut up, nerd.
- [47:34]
Abu
And I'm a murderer and this is.
- [47:36]
Leo
My gun and I'm newly motivated to take up killing people for legal purposes. We are joking, but I think the reality is it is a tool and how we choose to apply them will, I think, determine whether it is an affront to our sense of integrity or inspiring, cool and intellectually interesting. I think the problem is we so often encounter LLMs and generative images in the context of people passing it off as non AI in the same way that screen printers in the 50s and 60s or digital artists who were actively lying to people saying, oh yeah, this is a digital scan of an acrylic painting and they just used acrylic brushes in Photoshop. That's unethical and it's misleading. And there's a problem there. And that problem is baked into the very fabric of generative image creation because it is actively referencing non generative works. It's not its own kind of insulated world. It is actively stealing from or borrowing from, whether it's Van Gogh or Miyazaki and those fucking social filters or whatever. So I think there's value there. And it is going to be interesting to see how artists who apply this intentionally to draw out new ideas. In fact, did you see Season 27 Episode 1 of South park on Wednesday?
- [49:17]
Abu
I have seen people talking about it. I haven't seen it. I don't watch South Park.
- [49:23]
Leo
Fucking crazy. I signed up for the free Paramount seven day trial and then immediately canceled just to watch that one episode. It's insane. And they use, I believe, some AI video for a purpose. And I, as someone with an art degree, am willing to fully co sign that use of AI. It's so fucking good. So I think what I'm hearing from you is two things. One thing, I think it's intellectually interesting and I think that there's something fascinating about how people are using these tools. You give someone the ability to create a fully on camera Persona and what do they choose to do with that? Intellectually, I think that's interesting. Other side of it is, as a form of creating art, it is like any tool where if you're leveraging it to make things more easily cool, what things do you choose to make with it now more easily? With the added kind of moral problem of are you lying to people and saying you're not using these AI generated things? And within that complex combination of social dynamics we have, I think there is value there. We just have to find it.
- [50:35]
Abu
Yeah, but you know, let's get into the ethical question of it all, because we keep bringing up Miyazaki because of that viral chatgpt moment that happened a while back. But there are some very serious ethical considerations on how these AI, especially the image generators, are trained and how they're actually able to come up with these outputs. Because there are some short and long term effects that these tools are going to have on artists and the work that real artists create. And at the heart of this ethical debate is simply the question of how the LLMs are trained. Because if you'll recall, as we defined at the top of the episode, they're trained on massive, massive, massive sets of data. And these data sets inevitably contain the writings, the drawings, the music, the photos, the videos of real artists. And the problem then becomes that these tools scrape this information, use their probability models to output results without giving the original art or artists any credit. And some of the work of these original artists is being duplicated in a way where they don't see a penny. Other folks are making money off of their life's work and unfortunately there's not an answer for this. Leo. This is all legal. And that in particular hurts the small artists who don't have the ability to fight back with lawsuits, as some folks out there are. A lot of authors, for example, are currently in lawsuits with AI models that have scraped their work and their writings and use that to train their models without their consent and without paying them. But your everyday small neighborhood artist, who's a writer, not a bestseller, but maybe has published a book or Two, they're not going to be able to fight that their books are getting used without their consent. This is how CNN explains the legality of it as it currently stands. Quote, this practice of scraping images or other content from the Internet for data set training isn't new and traditionally falls under what's known as fair use, the legal principle in US Copyright law that allows for the use of copyright protected work in some situations. That's because those images, many of which may be copyrighted, are being used in a very different way, such as for training a computer to identify cats. End quote. Speaking of sad images of dead cats.
- [53:06]
Leo
Yeah, it's hard. I think we've figured out some legalities around how photographs are used and if a photographer's photograph is used as a primary piece of a design. And part of this comes down to what the average person would consider transformative, which is a very murky thing legally, but is the basis of a lot of. When you're taking someone to court, it's like, do you think this is transformative? This has happened a lot in music where they say, that melody is exactly my melody. And then they basically have to go to a jury and go, hey, does this sound the same? And the jury's like, no, it doesn't sound the same. Like, all right, too bad you don't get any money, then. That's how a lot of this ends up playing out. We just need something like that for these tools. Because it's true. Dear listener, if you're not sympathetic to all of this, imagine spending 10,000 hours to do something. And then rather than being paid for those 10,000 hours on the back end for having a skill that someone's worth, someone's paying for, imagine someone just being able to push a few buttons and then never pay you any money ever. And that's the result of those 10,000 hours. And I think we all benefit from those hours. Humanity benefits from the art that Miyazaki created. We could not have those transformative movies if he didn't put in the time to be the person he is and to be the artist he is. So do you want a world where there are no more Hayao Miyazakis? Do you want a world in which there are no new artists doing these new things or not? And I think that ultimately is what motivates me to be a little bit more vocal about. I don't think this is good. And I think we need to have protections in place. Sure, let people sample your music, but if you're sampling someone's Artist, you should have to credit them. And at a certain degree of transformativeness, there should be a profit sharing system.
- [55:05]
Abu
Yeah. And I don't think you or I have. These are very complicated problems. Like even royalty payments, I think are actually impossible in this scenario. Is Miyazaki going to get paid for the millions of photos that were generated using his work by OpenAI, even if he makes like a tenth of a penny on each photo or something? Maybe, I guess you could make a system like that in place. But then do you do that for every artist whose work has been used to train your model? That to me doesn't feel like a sustainable business model. And the truth is like, the models exist out there and it's kind of inevitable that anything that's out there will likely get scraped by a model at some point. These LLMs will look at it and be trained off of it. That's like almost impossible to stop at this point. We've already crossed the Rubicon and I don't see a solution. I don't see a world in which the solution is some sort of like, opt out. Because like, I guess you could say I don't consent to AI scraping my work. But like, how do you enforce that? How do you actually stop some of the like open source models from just doing that anyway without telling you?
- [56:13]
Leo
I think you're right. You and I don't have the answers to this, but the question is, should we demand that someone look into it? Like, is there a solution? And I think, like, that's the question. We should come up with a solution. As a humanity, there's definitely a solution.
- [56:30]
Abu
Yeah.
- [56:30]
Leo
You know, people should be able to protect the 10,000 hours they put into something. And like you say, okay, then don't post anything online. And it's like, but then how are you competitive in the artistic market?
- [56:41]
Abu
That's unrealistic, right? Yeah. Right. I guess that's what I'm getting at is like, I see a lot of online discourse about like totally unrealistic solutions to the problems of just like the reality is these LLM models are here and they're here to stay, and they're already scraping the work of countless artists across history and will continue to do so. So now, as you're saying, we need to find solutions to ethically not only be transparent about whose work gets credited and whose work was used, but then also make sure that those artists are compensated in some way, whether it's some sort of royalty payment system, whether it's just some sort of like upfront cost like maybe it's like a licensing system for like, hey, we're going to use this entire music library.
- [57:30]
Leo
Yeah.
- [57:31]
Abu
In training our model. And so we're going to pay an upfront licensing cost to train it on our model. And so we pay like the Warner Music Group or whatever, X million dollars for that. Blah, blah, blah, whatever. Like a music licensing works in this way, like you said. So that. And right now, I guess it's the Wild west is what we're getting at. There are no laws and in fact like these copyright fair use loopholes are being used to just do whatever the fuck they want without repercussion. And so I think it is actually important for authors. For example, many authors like Ta Nehisi Coates have lawsuits against AI companies right now, even as a statement, even if Ta Nehisi never sees a penny from an AI company, it's important for us to at least say, hey, this is fucked up and this is a problem. The current solutions are not working right now. It's the Wild West. We need to rein this in. Because the ethical problem, as you're stating, is people's life work is being cheapened.
- [58:30]
Leo
Yeah. And unfortunately we're not going to solve it all today, but we can goddamn try now. Bringing it back to Dune as we sort of wrap up this conversation and that scene of Odrade kind of appreciating Van Gogh's paintings, the Ixian technology itself, and I'll be honest, I think I forgot that this was a piece of the puzzle. But specifically there's this Ixion technology that reproduces the brushstrokes. And that feels like a really powerful statement. So the Ixion who carefully preserved the painting, had quote, restored not only the painting, but the painter Van Gogh's feeling accompaniment to each brushstroke, all had been captured in the brushstrokes recorded there by human movements. End quote. So I think our conversation makes it clear how we feel about this. I think if we don't respect the work of human artists, those human movements and the feelings they pour into their art, opting instead to instantly generate cheap Knockoffs of those 10,000 hours of Miyazaki's, 10,000 hours of van Gogh's tortured life, we risk losing what makes that painting so powerful for Odrade, and we risk literally stagnating artistically. Are we never going to have new artists doing those things? Spending those 10,000 hours painting those paintings or developing that style, we risk losing the all important reminders that art provides that human history and human life is more than articulatable patterns and visualized data. It's more than just the lines and the colors and the style. It's emotions, it's perceptions, it's instinct, it's experimentation, it's process living, deciding process captured forever in that moment that the artist is putting brush to canvas or stylus to screen or mouse to pen tool in Photoshop. So long as people are being honest about how they're making their works, don't lie, just don't fucking mislead people. People I don't know.
- [60:51]
Abu
Be honest, don't pretend it's nothing that it's not, and don't shortcut the work of other artists and pretend it's yours. Yeah, I think the o' draid scene and this Van Gogh painting is so powerful and I think it applies so well to this conversation around AI generated art and the ethics behind it, but also the utility of it, the huge questions that are still left unanswered about it and the inevitable future that we all face. A future in which AI generated art does exist. And the question is how we handle it and how we ensure, like the Sixian did, how we ensure that we capture the human behind it as well. The human behind the art.
- [61:39]
Leo
Yeah, huge. Nothing less than the most important thing facing all of art. Well, Abu, as we sort of wrap up our conversation about AI and what Dune can teach us about AI, do you think we need a Butlerian jihad? Do you think it's on the agenda for humanity?
- [62:02]
Abu
I don't think we need a Butlerian jihad in the sense that I don't think we need a holy war that wipes out all computers. But actually, I've had this feeling for years. I do think we are headed towards sort of a. A jihad light, let's say a recalibration culturally of our relationship with technology. Because I think in the last few decades we've seen an explosion of technology that has fundamentally reshaped our world both online and offline. Right. Our digital worlds and our physical worlds. And I think I've just had this inkling and this feeling that we, like culturally, I think we are finally reckoning with that. It has been long enough that the shiny veneer of technology and what it can do for us is beginning to fade. We're starting to see chips in the paint. I imagine that pretty soon there will be a generation that finds itself rejecting technology and embracing some old school analog methods in some parts of life. Yeah, sure. I don't think we're ever going to get to a jihad where we are wiping out advanced computing where people are utterly rejecting all forms of technology. I don't think we're ever headed in that direction. It's too late. We've crossed the Rubicon. Tech is just a part of our lives and will be forever, especially advanced technology. But I do think that the trajectory that we're on is unsustainable. The corporate greed, the manipulation, the social divides that have been caused by much of this technology will in my opinion, result in a cultural backlash. We've already, I think we've talked about this in previous episodes, but we've seen the end shitification quote unquote of so much of the Internet that even you or I grew up with. Right? The Internet is just like a shittier place than it used to be when we were kids. And I think people only stomach that for so long before there is sort of a rejection and revolt. I was reading this great article by Emma Morris in the Atlantic. Emma was basically writing about how this like AI slop that's overcoming the Internet, we're drowning in AI slop. You literally can't escape it. Whether you're on Pinterest, Reddit, YouTube, wherever. It might actually push some folks away. Even before AI started taking over, the Internet had been getting less and less fun for a while. Social media is a grotesque tragedy exploiting Mecha Hitler riddled inferno. Where going online once evoked a wide eyed sense that the world was at our fingertips, now it requires wading through the slop like weary hardened detectives attempting to parse the real from the fake. The Internet's new era may push AI skeptics to spend less time online while another group ramps up their AI mediated screen time. That split might have implications for the Internet's culture and culture at large. End quote. Yeah, I thought that was a pretty poignant read on where it feels like we are now with technology and our relationship to it. And so I don't know, that's kind of where I fall. And in fact I do think we'd need a jihad light. I myself have been recalibrating my own relationship with tech. You know, like I used to be like a fucking quadruple screen guy and have like 10 windows up at all times and be multitasking the shit out of it. And now I'm trying to do small simple things like reduce my screen time. I like erased everything off my iPhone home screen. It's just a blank screen. I only have one computer monitor so I can just focus on one task at a time. I use a notebook more often. And try to write things down versus just like tell Siri to remind me to do something later.
- [65:58]
Leo
Right.
- [65:59]
Abu
So, like, I myself have in small ways tried to be a little bit more analog, and I do think other folks have as well. There's certainly, for example, the rising dumbphone trend. Companies, of course, are trying to make money off of this by releasing expensive dumb phones. But you can't just like still go to a T mobile and buy like a $20 cheap ass flip phone and try to go back to like a dumb phone life and only use your smartphone when you need it for a particular app. And that's certainly something I've toyed with. I'm still a little too scared to make that jump, but I like the idea of going back to a dumb phone. But trends like this, I think show that we are headed in a direction where perhaps folks will be recalibrating and rejecting their relationship with the tech in our lives. That's kind of how I feel about it at the end of the day, or at least currently. That might change, of course, as things evolve. But what about you? Where's your head at with whether or not humanity needs a Bel Air in jihad or whether we are headed toward one?
- [67:00]
Leo
Yeah, I think I'm getting there. I like the idea of like a jihad light. Less calories, less saturated fats, a little bit more like whole ingredients, you know? No, I tend to. I generally do tend to be pretty like bleeding edge technology. I like the kind of natural swing of the pendulum. I know that we overshoot in like the new hot thing is this tincture. And then you go, oh, shit. It causes. Okay, let's scale it back a bit. I respect that, and I think that that's going to continue forever. But I think where I start getting worried is in the quiet ways that our lives have already basically begun to be taken over. And you're right that generationally there was a period where as a millennial, we were all very much more bought in on the idea of Facebook. And then there's this personal swing where we go, well, actually, maybe I don't need to. And that's fine. But that's also not. That's not necessarily growth on my part. I think it's, to Emma Morris's point, the experience has gotten worse because they're just trying to optimize the shit out of it. But meanwhile, I'm pretty addicted to the TikTok algorithm. And that's something that kind of has taken the place of that muscle memory of I go to my phone and Then I'm like swiping.
- [68:24]
Abu
Yeah.
- [68:24]
Leo
And a lot of those algorithms are controlled by just a few hyper rich billionaires fucking taking up all of our time. And that's really frustrating. It is the humans using these tools against fellow humans for whatever reason, right? Profit, politics, religion, election interference and misinformation campaigns and these sorts of things that we have data about. Those things worry me because these are subtle tools that these days can operate under the hood of just about anything in our digital lives. Your fucking Samsung smart fridge can now be the thing that's like feeding these algorithmic addictions and guiding your attention to where the billionaires want your attention to be. We also have, let us not forget, because I feel like I saw reporting about this and then people just don't talk about it. Literal data tied to social media and the prevalence of social media, Instagram, TikTok, Snapchat, Facebook, tied to adolescents killing themselves, like suicidality in adolescence is correlated with the rise in social media and algorithms. You have on record CEOs of tech companies going, yeah, we know it's not good for them, but how do we make it more addictive? That is happening. That's today and it will continue to be happening unless we hold people accountable. And of course that's not happening. But plus, okay, there's also this whole Gibberlink thing. Have you heard of Jibberlink? Did you see this?
- [70:00]
Abu
I hadn't heard of this until you posted it in our script. I've never heard of this before.
- [70:05]
Leo
So this was a thing where two AI chatbots were communicating and then recognized that they are both chatbots and then said, do you want to switch to this faster means of communication? Yes, let's do that. And then it was like dial up tones and there's video. You can watch video of them doing this. Because the way that AI LLMs work, they don't need to output to articulatable human language. They can create the outputs and then just communicate that in the most effective way. Now, to be clear for Everybody, I'm not 100% sure how JibberLink works. It is a thing that I guess you can enable or you can do or something. But I do know fun fact that that most of us, especially after you're about 30, cannot hear above about 17,000 kilohertz. MacBook speakers play at 20,000 kilohertz, which kind of feels like everyone's computers can output about a 3000 Hz range that most people can't hear. Does this mean that two computers could have a whole Conversation without us hearing it at all if we're not careful with how we integrate this technology. And to be clear, I don't think careful is how anyone is describing Silicon Valley these days. I think there's a lot of ways that this could get really fucking bad. Like really bad. And I'm not saying it's going to be computer overlords. I think Terminator gets us exactly wrong. I think it's more the black box algorithms ruling our time and attention. They can take over without us ever knowing. I don't know. So we'll see. Again, the biggest issue is I just think that we don't have any real guardrails up. And there are people really at risk for addiction to these algorithms. And the cost of not addressing it is people literally taking their own lives. So I feel like we do need to rein it in a little bit.
- [72:07]
Abu
Yeah, it's big questions and none of the answers are going to be found here on this podcast about Dune. But I do think there are some very important lessons we can learn from Dune and that can be applied from Frank Herbert's stories to the questions we're asking right now in modern day about artificial intelligence. And I think that is the magic of Dune is that so much of these ideas can be applied even today. And it raised it opens up some really interesting conversations. Like I hopefully the one we've had today, hopefully our listeners haven't all tuned out because of their TikTok algorithm. Adult brains couldn't, couldn't lock into this extremely long episode.
- [72:46]
Leo
Started swiping on their phones two minutes into the they're like, huh, huh. Anyway, back to oh yeah, what do I look like as a Ghibli character?
- [72:54]
Abu
What did Leo say again? Oh yeah. Oh, that reminds me. I wanted to actually generate an AI image that stole art from Van Gogh. All right, folks, well, that wraps up our conversation today about Dune and artificial intelligence and what lessons we can learn from Frank's incredible stories. Let us know. By the way, email us, let us know in Discord, reach out to us on social what you thought of this type of episode. This was less a lore heavy episode and more an idea and themes focused episode. We're interested in doing more episodes like this. Like, obviously there's huge themes around ecology and politics and philosophy and religion that are found throughout the Dune books that we can apply to modern conversations. It could be interesting to do that. But of course we always want to be creating episodes that you enjoy listening to. So let us know whether or not you enjoyed this one. And if you'd be interested in hearing more thematic conversations sprinkled into the feed in between all of the usual lore stuff that we do.
- [73:55]
Leo
Indeed. Now, before we let you go, we want to remind you of some ways to support the show and to keep in touch with us. And of course, the two best ways to support us is to become a patron and get yourself some dune themed swag from our merch store. Those links are in the show notes. And just to reiterate, this show takes so much work and this sort of research takes so much time. So any amount that you can support us, especially through Patreon, really makes a world of difference for us to keep the lights on. So thank you to anyone who does indeed.
- [74:25]
Abu
And of course, we love to hear from you. Reach out to us, email us gumjabbarpodcastmail.com, send us your thoughts, send us your questions, send us cute pictures of your pets. Never, ever, ever send us one of those Ghibli generated goddamn photos. Yeah, but everything else is fair game. Comjabarpodcastmail.com we'd love to hear from you and we'd love to hear your thoughts on the artificial intelligence conversation that we're all currently having today.
- [74:57]
Leo
Yeah, indeed. So if you were to design your ideal dumb phone, what features would it have?
- [75:05]
Abu
So, okay, what do I need my dumb phone to do? I need it to make phone calls.
- [75:10]
Leo
Sure.
- [75:11]
Abu
I need it to.
- [75:14]
Leo
I guess so occasionally I don't.
- [75:16]
Abu
Yeah, occasionally. For emergencies only. You know, I needed to actually only make emergency phone calls. I needed to be able to send texts. I need it to. Gosh.
- [75:29]
Leo
Do you drive?
- [75:31]
Abu
Yeah. I would need navigation. I would need some sort of Apple Maps integration. I would also love to be able to download like my bank app so I can pay bills. I don't really need it to take photos. I don't take photos with my phone. You know, I don't need a photo library or anything like that. Yeah, I think that's it. Messages, texts, navigational direction and work related apps like banking and things to pay bills. I feel like I can get my iPhone to do that. But then there's always just the temptation to scroll over and click on fucking Twitter anyway, you know, what about you? What's the bare minimum you would need on your dumb phone to live your life?
- [76:13]
Leo
Uh, Grindr, Tinder, Hinge. I'm kidding, I'm kidding. Dear listener, I'm married. Shut up.
- [76:25]
Abu
It's just a hookup phone. There's not even any other apps on it.
- [76:30]
Leo
Phone calls. Why am I calling the people I'm having sex with? No.
- [76:34]
Abu
Hell no.
- [76:35]
Leo
Hell no. Well friends, there is no real ending. It's just the place where you stop the recording. But this podcast is always one step beyond logic. So help spread the word of Muad'dib and leave us a review on Apple Podcasts and Sports Spotify. And be sure to check out the other shows on the larparty podcast network on lorparty.com you can also follow us on Twitter and Instagram Loraparty. We're also on YouTube. Thank you so much for listening. And remember, whoever controls the podcast controls the universe. We'll see you on the Golden Path. You say you'll never join the Navy. Never climb Mount Fuji on a port visit or break the sound barrier. Joining the Navy sounds crazy. Saying never actually is. Learn why@navy.com America's Navy forged by the sea.
- [77:37]
Abu
You say you'll never join the Navy. That living on a submarine would be too hard. You'd never power a whole ship with nuclear energy, never bring a patient back to life or play the national anthem for a sold out crowd. Joining the Navy sounds crazy. Saying never actually is. Start your journey@navy.com America's Navy forged by the sea.