Neal Stephenson joins TBPN on day one of his book tour, discusses AI, filter bubbles, and the fall of Rome

Speaker 2: Welcome to the show, Neil. How are you doing?

Speaker 8: I'm good. My voice is still working because this is day one of the book tour, so you got me in a I'm great sure.

Speaker 1: How have, book tours evolved? I imagine your book tour now is like eight podcasts a day for two weeks.

Speaker 8: Well, yeah, of course, that's obviously new during the time I've done it. It used to be, like the first book tour I did, it was old school and they would sort of put you on like drive time shock jock radio programs and you'd try to explain your book to a, you know, a shock jock in thirty seconds. It was pretty challenging. I like this better.

Speaker 2: There a trade off between talking the book too much and not leaving enough for people to actually go and convert versus like talking about anything, talking about other topics, anything that just gets people interested? We've talked to some people. I think it applies more in non fiction where someone has a thesis and then they lay out that thesis on like a three hour podcast, and then people are like, well, why would I get the six hour audiobook if I can just hurt listening to you for...

Speaker 8: Yeah. I kinda had that experience with sometimes with non fiction books like business Yeah. Totally. Advice books, you know. It's like, okay, I I get it. Yeah. I don't need to read it. Yeah. Obviously, you don't want that as an author. Yeah. I think with fiction, you've got a big advantage in that if you can kind of set up the general the general thesis of the book, that hopefully gets people interested in the story and gives them a reason to actually pick it up and read it.

Speaker 2: Yeah. Yeah. Makes sense.

Speaker 1: How do you think about the responsibility that... Like, do you feel like science fiction science fiction authors have a responsibility? And and do you think... Like, how much of a weight do you feel like there is to the work? Because, like, in my view, you guys you're sort of predicting the future and you're entertaining, but then also it inspires so many people that you you impact the world in a very unique way just just with your words. And I think you have to assume that, like, something about it, these ideas just sort of, like, weed their way into people's consciousness. Yeah. And it affects their actions and affects the products that they make. It affects the weapon systems that get built in the future. Yeah. And there there seems like there is some weight to it.

Speaker 8: Yeah. I mean, I think the good thing is that to tell a story that intelligent readers will actually enjoy, you need to be realistic. You need to include some of actual complexity and ambiguity of the real world. Otherwise, it just comes off as kind of a one dimensional, you know, screed. And so I tried to do that, you know, the last book I wrote that had some of that potential was was Termination Shock, which is about geoengineering. So, I mean, it depicted people doing geoengineering and one of the main characters was a big advocate of geoengineering, but there are also characters in the book who took a different view and who were ambivalent or openly hostile to the efforts that were being made in that area. So I think if you, to begin with, that just made it a better book. But, you know, if you're reading a book like that with an eye towards, you know, starting a company or something, then hopefully you notice the fact that there's different points of view on it, there's complexity, and you take that into account in your decision making going forward.

Speaker 2: Do you believe in nominative determinism?

Speaker 7: Well,

Speaker 8: novel writing, sure, because you get to pick the names.

Speaker 2: Yeah.

Speaker 8: And so the... So Dickens, for example, you know, was was really good at at giving his characters names like Scrooge or Uriah Heap Yep. Or whatever that immediately gave you a clue as to the nature of that particular character. Whether that really translates into real life, I'm a little less certain.

Speaker 2: Yeah. Well, how do you grapple with it as an author? It it it can sometimes be tempting to say the main character's name will be, you know, Chad Goodman and the bad guy will be named like evil And it can be a little bit on the nose. But sometimes it's Hero protagonist. Yeah. Yeah. Hero protagonist is a great one. Yeah. But how do you how do you personally deal with that balancing act?

Speaker 8: I mean, it's just a question of trying to have a little bit of a light touch, you know, make it a little funny, a little witty. If if it feels like you're hitting the reader over the head with something obvious, then they're not gonna take that very well. Yeah. But on the other hand, if you can kind of slip it in there and have it be sort of sort of clever and funny, then the reader feels like, okay, I'm in on the joke.

Speaker 2: Sure. Sure. Can we go back to 2013? I'd love to know how you met Seamus Blackley who sort of, I think sort of kicked off this arc. Do I have that correct?

Speaker 8: Yeah. Yeah. I've known Seamus, I don't know, since the late nineties probably. He, after he left Microsoft, he got interested for a while in trying to help gamers fund their projects, and that I think led to him being... We're we're kind of running the game division of CAA, the big Hollywood talent agency. And at the time, my agent in LA, Richard Green, was working at CAA. Sure. And so one day Rich said to me, hey, you know, if you're in town, there's this guy, Seamus, you know, you guys should meet. So we went out, had lunch, and immediately kind of, I think, understood each other and became friends. And so we're... We've been in touch ever since. I saw him just two weeks ago when I was down in LA.

Speaker 2: How are you...

Speaker 8: But, yeah. He...

Speaker 2: Yeah. He's sort inspired the book, right? Sorry.

Speaker 8: Yeah. He made me aware of the existence of some of the story content that this book is based on, specifically the Earl of Suffolk. And so since then, he and I have sort of traded notes as the occasion warrants about, you know, when we come across some interesting piece of information, we'll email it back and forth to each other.

Speaker 2: How are you thinking about the current moment in video games and AI? There's all of these mashups that are happening. People are using AI to decompile video games, and it feels like we're at a moment where fan made video games are going to be more of a thing like a fan made Halloween costume. I can imagine that fans will turn many of your works into AI generated video games. There's obviously a lot of property discussion around there, but how are you churning through it? Is there a part that excites you, part that is nerve racking? How are you dealing with that moment?

Speaker 8: Well, I'm more of a just hands on guy, so I am the co founder of a startup called Wind Air that's actually working in kind of in this area, not exactly what you're talking about, but, you know, the premise is to allow fans of story worlds to tell stories in their favorite you know, fictional universes by giving them an environment that, you know, represents part of that world and a script editor that they can use to determine what they want to see happen. So, and we've actually just last week made an announcement. We've been working with a company called Dreamlands dot ai that has made some incredible strides in this area, made it possible for a tiny team like ours to generate environments, realistic high quality, you know, video game environments that a few years ago would have cost, you know, a million dollars plus. So that's kind of the way I come at this. If... I I haven't followed the, you know, the the mashup scene very much, though.

Speaker 2: Yeah. Jerry, please.

Speaker 1: So you published Snow Crash in 'ninety two, and then we went through a period of, like, what what in hindsight feels like total stagnation, where you get, you know, the the Internet becomes big and and that's cool, but it sort of just extends it extends a lot of existing technologies. And then it feels like the 2020 is, like, I I think, like, generally, people have felt the sort of acceleration and the speed up and new capabilities. From your view, having written about a lot of the things that we're starting to, like, see in the real world and experience today even in small ways, Did it feel like a period of, you know, twenty years of of stagnation and then and then a and a and a real acceleration? Or, like, what was your sort of lived experience? Because it just feels like such a big it feels like a big gap where, like, if reading The Wall Street Journal or reading the mainstream media has at times felt in the last year like you're reading science fiction. Yeah. And it certainly didn't feel that way. In some ways, like, the early two thousands and the twenty tens was like, great. We have new database software and we have more wars, and it didn't it didn't feel like there was any... It didn't feel like it was the news

Speaker 8: Oh, that was wild excitement. Yeah.

Speaker 1: Yeah.

Speaker 8: Yeah, know what you mean. I kind of wasn't even aware of it, you know, but now that you point it out, it's like, yeah, yeah, it was sort of that way. And I mean, you know, the Internet came along. It was big. It was very important. And there was kind of a period of consolidation, I think, where it changed from, you know, sort of like, you know, rampant grad students, you know, stringing things together to big business and the consolidation of power and influence on the Internet by the, you know, the FANG companies basically. And so, yeah, I think gradually over time, you you begin to... You get tired of it. You begin to lose the sense of of wild excitement and, you know, big things happening and the sense that, hey, there's I can do something, you know, there's opportunities here that I could grab. That that starts to to fade after a while and now it's sort of back because of the the things that you mentioned.

Speaker 2: How are you

Speaker 1: thinking about... Then and then on that, like, as you're... I I don't know what your process is like for, you know, generating ideas for for future work. But, like, do you feel like this is a time where you can open the newspaper and start to... Because what's been interesting is like many people in the technology industry, like the people closest to the frontier of AI, have started writing their own science fiction. Right? So you have work like AI 2027, which is like... 2040. Yeah. And and AI 2040 where it's like, hey, let's look at the the technology that we have today and sort of just try to extrapolate it out quarter by quarter and see where this goes. And and you have to... Basically have this series of compounding assumptions that lead to these really insane outcomes. But in some way, that's you were kind of running that same kind of process, but over maybe a larger time scale where it's like if you assume the technology we have today is gonna improve this way and make jumps in this way, then I can imagine this whole world that where this would be the way that you travel around the earth or or this would be the way that individuals interact with, you know, corporate, large corporations, etcetera.

Speaker 8: Yeah. I haven't read the things you mentioned. I wasn't aware of them, but I'll, you know, now I am. I'll check them out. But I think, you know, to make a good yarn, you need to... There's some elements that you need to include in the story that give it a feeling of realism. And one of those is, you know, it's one thing to posit a new technology and think about what you might be able to do with it just on a pure engineering level. But in addition, you know, real technologies answer to other masters as well. So you've got kind of the social and creative side, you know, what do people do with new technologies that's unexpected, like electric guitars. You know, we had electricity. Everyone knew there was going to be light bulbs and washing machines, but electric guitars were kind of an unexpected thing that came out of a more social and cultural milieu. And then there's the business side, you know. You've got to, you know, you've got to have a plausible business plan for why some future technology in a science fiction book actually makes sense and can get funded. So, those are some elements that I think, in a way, it adds some additional work and some additional challenge to the process of writing science fiction, but it more than pays off because what you end

Speaker 3: up

Speaker 8: with is something that feels a lot more organic and realistic and as such can do a better job of drawing the reader into the story.

Speaker 2: Mhmm. How are you thinking about AI in the process of researching, writing, fact checking, proofreading, and the trade offs of using tools that might atrophy your own skills? How have you grappled with all the different opportunities and risks?

Speaker 8: Think you answered the question. I'm not real big on atrophying my skills. Yeah. So, I do think that it's, you know, it's a little bit difficult to explain, but there's something about the serendipity and the randomness of doing it yourself that is very productive in the end. And so, like when I was writing my Baroque Cycle series of historical novels, I would just make notes by hand in comic books. I had a huge stack of these notebooks and it was all hodgepodge, you know. It wasn't organized, it wasn't searchable, certainly. And so I would have to flip through them trying to find stuff, you know. I think it's somewhere in here kind of deal. And in the course of doing that, I would stumble across things I'd have forgotten about or connections would get made, new ideas would emerge, and so for me, having that work done by a research assistant, whether it be a human or an AI, would actually be counterproductive. And I'm just not interested in it at all.

Speaker 2: Yeah.

Speaker 8: But, I mean, I already know how to write books, so I don't need help.

Speaker 2: That's a good point. Is there a science or technology that you think is underrepresented or underexplored in science fiction? Like, there's a lot of space stuff explored.

Speaker 8: Lot of

Speaker 2: robots, basically. I mean, we maybe need to advance that. But is there something that sticks out to you as as potentially fertile ground or underexplored?

Speaker 8: I mean, I think a lot of the story of the next half century or century is going to be how we get carbon out of the atmosphere. Sure. And that is a straight up technology engineering project that's going to have to be carried out on a really large scale. So I'd like to see more about that for sure. There's all kinds of different carbon sequestration technologies that are being talked about. Some are biological, some are kind of industrial facilities, some are sort of like enhanced weathering of stone and so on. So there's a lot of ideas floating around right now and we haven't really settled on what the right one's gonna be. But it's an interesting area right now of engineering ferment. There's like bright people going into it. And so I hope that we'll be seeing more in that area.

Speaker 1: Mhmm. What about... Would you ever write something about algorithmic social feeds? Like it it seems like it seems like an idea that you would have imagined that maybe you have like, you know, a decade ago or two decades ago or this idea that like everyone would be spending, you know, four hours a day, you know, immersed in this device that is showing them entirely different sort of world view and and set of facts and all these different things. Was scrolling I was scrolling on X over the weekend and there's this new I mean, it's not really new, but it seems to have gotten worse where people would just share fake image, you know, fake news, fake images, and the only and these things have, you know, 10,000,000 views. Mhmm. And I can see that it's in the process of getting community noted. Sure. And you can click in and see that like, oh, this is not a real image. This never appeared or whatever. But but AI is making it we've had this technology for years now but it's so much easier.

Speaker 2: Yeah.

Speaker 1: Is easier. And if I wasn't in the if I wasn't in the like community notes program, I never would know that I would just be scrolling past these things seeing fake news. And so there's this sort of fake news element that it is still underexplored but then also this idea of like, I'm just seeing...

Speaker 2: Living in a filter bubble.

Speaker 1: I'm living in an entirely different online world than than somebody next to me in traffic. And then what are the downstream implications of that? Yeah.

Speaker 8: I kind of went there briefly in my book Fall or Dodge in Hell, and it's sort of a minor sideshow to the main story. So... But it kind of depicts a semi near future in which big swathes of the country have been kind of colonized by the sort of thing you're describing with the results that the, the, that part of the country in my book is, is called Ameristan. And you know, it's this kind of war torn place inhabited by people who have absolutely no idea what's going on, but they're sure that they do. And so it's, to me, to me, I mentioned carbon before, to me the two big problems facing us right now are carbon and what you just described. And with carbon, I can see, you know, that's an engineering problem straight up. The other problem, people living in their own realities is you could address it with engineering or, you know, by blowing up a few Yandex server farms.

Speaker 16: Oh,

Speaker 8: yeah. That's crazy. It, you know, it's a harder problem because people like it, you know?

Speaker 1: Yeah. And people... So that means you have to police. What are you gonna do? Start policing? Like, it can get Yeah. Kinda dark if you're like, oh, okay. Well, now the government is gonna decide all

Speaker 8: Yeah.

Speaker 1: They're gonna try to enforce a rule set around Yeah. What can be seen and shown. It's this weird thing where, like you said, like, the fake stuff just naturally is absurd often, so it gets more engagement. It is more entertaining. People spend more time. They get served more ads. Yeah. But at the same time, you have government intervention, and so, yeah, it's like we're much... The human race is so much better set up for engineering problems where we have this monumental problem that we're facing.

Speaker 2: At least we all agree

Speaker 1: on But in that situation, you have one hacker, you know, theoretically hacker in their garage or at some university lab or whatever and they can they can have this one breakthrough that can then get global adoption because there's an incentive to have that global adoption. And right now, with this like problem of us all living in our own world, there's no obvious incentive structure that that will fix it. If anything, there's an incentive for And it to get

Speaker 8: in some ways, the videos and the images are... For me, it's it's a little easier to look at those, at least at the current stage of technology, and say, yeah, I don't think that's real. But what's more insidious in a way is lots of posts underneath that video, you know, by, so mostly by bots, Right? And so in a way that's a sneakier way that the same bad actors can influence the way we think because, you know, we read these things and we tend to think, well, there's a real person out there somewhere who's actually typing this when, you know, that's that's not actually the case at all.

Speaker 2: Yeah. I know you've studied the fall of Rome a lot. How do you compare those problems that you just described to the problems that were faced by Rome? They feel extremely modern, but is there some connection that we could actually learn lessons from the past here?

Speaker 8: Yeah. Well, you know, there were two falls of Rome. There was the fall of the republic. You know, which finally bit the dust around time of Julius Caesar. And then there was the much slower fall of the the empire and a kind of different phenomenon in a way. It's easier to get a grip on the fall of the republic because you can just see what happens to certain people like Marius and Sulla who just decided that they were going to sweep aside the norms that until then all of the important people in the Roman Republic had agreed to observe, and they just decided to flout those norms and do what they wanted to do. And that's a little bit, I think, where we are right now in modern society. The fall of the empire is a, you know, slower and more complicated process that, you know, took, you know, another fifteen hundred years to fully, you know, reach its its end. There you're talking about a lot of things like migration and changes in religion and so on and so forth that make for good reading. I recommend, you know, studying this, but I don't know if I have a quick answer for you.

Speaker 2: Well, one solution for algorithmic feeds, picking up a book, d Heavy Water is available.

Speaker 1: I wanted to ask you before you leave. Do you think is the most who do you think is the most underrated science fiction author

Speaker 2: Mhmm.

Speaker 1: That you think would resonate with our corner of the Internet, our small corner of reality? You're you're you're you're among, like, venture capitalists. You might laugh at this, but you're certainly the most widely recommended and read, you know. Ask... If you ask any of them, they're... They'll tell you.

Speaker 8: I mean, if I just had to pick one off the cuff, first thought is Matt Ruff, r u f f, who wrote he wrote The Lovecraft Country, which got turned into a TV series. So that's his most widely known book. But every book he writes is different and they're all great. And he's got one, he's winding up now called Black Sheep, but I don't think it'll be out real soon. But I know he's he's... It's about AI. So so if I had to pick one, that's that's my first thought.

Speaker 2: Well, thanks for the recommendation. Was great

Speaker 1: to meet you. Thank

Speaker 2: you Thanks for coming coming on the show and congratulations on the book launch.

Speaker 1: Yeah. Good luck keeping your voice over the next couple of weeks.

Speaker 2: Yes. Good luck.

Speaker 8: Yeah. Yeah. Yeah. Talk to me in two weeks. We'll we'll know.