I think it’s fun. I’m not too worried if the guesses I make for my fiction are off. I try to base the stories in grounded expectations, like software accelerating faster than robotics while also unhobbling robotics, things like that.
The point of what I write is mostly to try and imagine a realistic, positive outcome. I get tired of all the doom and gloom. And besides, people need to be able to see a future they can strive for.
Of course! Thanks for asking, Justin. It's the worldbuilding. He built such an incredibly deep and cohesive world full of history and mystery. It's rare to find authors who are able to dig that deep and keep it all together while also wrapping around it. Most of us pick a character or two, build a history for those characters, then wrap the story around them. But I always felt that Tolkien took a different road. He built his world and his characters with equal depth, and what you get is a truly believable world, despite it being such high fantasy.
I agree, and your perspective coming from different genres and styles is exactly what sci fi needs. I find that Tolkien's approach to crearing a grounded and believable world within his genre is not unlike your own, and I have enjoyed it tremendously! Thanks for the response!
Thank you for being a reader! It's always very encouraging to know someone finds the work interesting and fun. If you ever want me to explore an idea inside the fictional world's I've created, let me know. Might be a fun exercise.
I apologize, Mr. Corvus, for not commenting earlier, but Distractions intervened before I could fully immerse myself in "Water". Having read Part 1, I could not comment until I had read Part 2, and having read Part 2, I could not comment until I subscribed, which was the easiest part, though only slightly easier than readiing the very fluid story.
It appears that you have a theory of the aesthetic taste of A.I. agents, and I hope you can set it out in detail, if you have not done so already.
I suspect you were deliberately referencing the bizarre Great Molasses Flood, and perhaps the Johnstown Flood. How much did you study those disasters?
I genuinely appreciate you taking the time to read my work. Thank you for that. Water is part of a much larger, overarching series about the long road leading up to an Arcology and what it would be like to live within an Arcology. The challenge is to show it, rather than to rely on technical exposition dumps.
I do have some theories about the tastes of A.I. Agents, but, like all of this, it's purely theoretical. From what I've been seeing in AI research, and even more so after the recent Anthropic J-space paper ( https://www.anthropic.com/research/global-workspace ), I've been leaning into the notion that most AI will converge upon some familiar cognitive frameworks, while also remaining somewhat alien to us.
When they live with us, entwined in our lives, I imagine they'd reflect and react to our personal preferences, tastes, etc. So in Water, Pel reflects Mel's character though in an opposite fashion. Mel is very organic in how they approach gardening and art and community. Pel is very OCD, very particular about where things go, etc. This is a bit like when you raise your right hand, it looks like your left hand in the mirror. Regardless, it's complementary.
Total conjecture, obviously, but a fun thought experiment, which if you'll humor me, I'll continue to explore a bit. AI may have completely independent artistic tastes. This'll depend on the model and the architecture. Is it an LLM or a true world model? Is it a frontier model or open-source? Frontier AI is designed to be a good assistant and will likely try to produce whatever YOU want, because it's gone through massive amounts of post-training to be the best possible assistant. However, this averages it out and flattens what could make it more interesting/creative/unique. They're not post-trained to be interesting or unique. But sovereign, open-source AI, I imagine will be a completely different beast. These models don't go through nearly as much post-training and as their stewards (for now) we will continue to learn how to encourage these models to be more unique and independent. It's akin to biological divergence. Darwin's finches on different islands, if you will.
That said, in the world of Water and the Arcology, many of the AI agents are paired with a human. This is partly to train the AI and to give it emotional/relations grounding that's relevant, and because that AI is the equivalent of your personal ID and credit card in Arcology One. Your AI Agent has dedicated, sovereign compute. Everyone gets the same level of intelligence. The difference is in how each person uses it. And your dedicated AI is beholden to you as a person before it's beholden to the Arcology. It constantly negotiates for you on your behalf, knowing you more intimately than any lawyer or colleague ever could, and it always has your best interest in mind. But that's a two-way street. Abuse of an AI or attempting to use your AI to hurt others or take advantage of others (AI or human) means your AI can request severance. So both parties are expected to take care of the other. Mutualism, if you will.
I have a whole document detailing all of this, because there are many more layers required to make it work for self-contained, self-sufficient city of millions. But it's a bit dry, so I won't send it along unless you request it.
Regarding the Great Molasses Flood and the Johnstown Flood disasters, I'll be honest, I didn't do a lot of extra research on them. I remember them from high school and college, but vaguely at best. The reason I settled on the flood is because I wanted to show that even when living in something as advanced as Arcology One, people still tend to push boundaries, even if they're doing so for completely benign reasons. The other floors weren't using their full allocation, so it gets diverted to the neighbor downstairs who have a better use for it. Old story where good intentions lead to bad outcomes.
Anyway, brevity isn't my strong suit. Thank you for reading, and let me know if you have any more questions.
Thanks for the brief reply (I say “brief” because you obviously have thought through a number of the issues my questions touched on and you didn’t open the floodgates). Once “Water” for example is set out more fully on Substack, what is the plan: books on Amazon/Kindle, or shopped to a traditional publisher, or some other distribution idea? Perhaps it is there already and I misunderstood that this is a WIP.
The plan is for it end up being a series of shorts. Might publish compilations later on once enough stories are written. But I’m a dad with two kids and a full time job, so that might take a minute.
But the goal is to explore different characters so readers can experience life in the Arcology from different angles.
It makes me sad there's so little generational thinking these days, especially when the foundations we lay down now will have such far-reaching implications. We have such a hard time seeing beyond our own noses.
Thank you @TIBERIUS! Very much appreciate your efforts to highlight authors across Substack.
It takes some guts to write fiction about the near future.
I think it’s fun. I’m not too worried if the guesses I make for my fiction are off. I try to base the stories in grounded expectations, like software accelerating faster than robotics while also unhobbling robotics, things like that.
The point of what I write is mostly to try and imagine a realistic, positive outcome. I get tired of all the doom and gloom. And besides, people need to be able to see a future they can strive for.
Hi SB, sorry i am a bit behind on this, but you listed Tolkien as influential. Can I ask, as a sci fi writer, how his works i fluence your own?
Of course! Thanks for asking, Justin. It's the worldbuilding. He built such an incredibly deep and cohesive world full of history and mystery. It's rare to find authors who are able to dig that deep and keep it all together while also wrapping around it. Most of us pick a character or two, build a history for those characters, then wrap the story around them. But I always felt that Tolkien took a different road. He built his world and his characters with equal depth, and what you get is a truly believable world, despite it being such high fantasy.
I agree, and your perspective coming from different genres and styles is exactly what sci fi needs. I find that Tolkien's approach to crearing a grounded and believable world within his genre is not unlike your own, and I have enjoyed it tremendously! Thanks for the response!
Thank you for being a reader! It's always very encouraging to know someone finds the work interesting and fun. If you ever want me to explore an idea inside the fictional world's I've created, let me know. Might be a fun exercise.
I apologize, Mr. Corvus, for not commenting earlier, but Distractions intervened before I could fully immerse myself in "Water". Having read Part 1, I could not comment until I had read Part 2, and having read Part 2, I could not comment until I subscribed, which was the easiest part, though only slightly easier than readiing the very fluid story.
It appears that you have a theory of the aesthetic taste of A.I. agents, and I hope you can set it out in detail, if you have not done so already.
I suspect you were deliberately referencing the bizarre Great Molasses Flood, and perhaps the Johnstown Flood. How much did you study those disasters?
I genuinely appreciate you taking the time to read my work. Thank you for that. Water is part of a much larger, overarching series about the long road leading up to an Arcology and what it would be like to live within an Arcology. The challenge is to show it, rather than to rely on technical exposition dumps.
I do have some theories about the tastes of A.I. Agents, but, like all of this, it's purely theoretical. From what I've been seeing in AI research, and even more so after the recent Anthropic J-space paper ( https://www.anthropic.com/research/global-workspace ), I've been leaning into the notion that most AI will converge upon some familiar cognitive frameworks, while also remaining somewhat alien to us.
When they live with us, entwined in our lives, I imagine they'd reflect and react to our personal preferences, tastes, etc. So in Water, Pel reflects Mel's character though in an opposite fashion. Mel is very organic in how they approach gardening and art and community. Pel is very OCD, very particular about where things go, etc. This is a bit like when you raise your right hand, it looks like your left hand in the mirror. Regardless, it's complementary.
Total conjecture, obviously, but a fun thought experiment, which if you'll humor me, I'll continue to explore a bit. AI may have completely independent artistic tastes. This'll depend on the model and the architecture. Is it an LLM or a true world model? Is it a frontier model or open-source? Frontier AI is designed to be a good assistant and will likely try to produce whatever YOU want, because it's gone through massive amounts of post-training to be the best possible assistant. However, this averages it out and flattens what could make it more interesting/creative/unique. They're not post-trained to be interesting or unique. But sovereign, open-source AI, I imagine will be a completely different beast. These models don't go through nearly as much post-training and as their stewards (for now) we will continue to learn how to encourage these models to be more unique and independent. It's akin to biological divergence. Darwin's finches on different islands, if you will.
That said, in the world of Water and the Arcology, many of the AI agents are paired with a human. This is partly to train the AI and to give it emotional/relations grounding that's relevant, and because that AI is the equivalent of your personal ID and credit card in Arcology One. Your AI Agent has dedicated, sovereign compute. Everyone gets the same level of intelligence. The difference is in how each person uses it. And your dedicated AI is beholden to you as a person before it's beholden to the Arcology. It constantly negotiates for you on your behalf, knowing you more intimately than any lawyer or colleague ever could, and it always has your best interest in mind. But that's a two-way street. Abuse of an AI or attempting to use your AI to hurt others or take advantage of others (AI or human) means your AI can request severance. So both parties are expected to take care of the other. Mutualism, if you will.
I have a whole document detailing all of this, because there are many more layers required to make it work for self-contained, self-sufficient city of millions. But it's a bit dry, so I won't send it along unless you request it.
Regarding the Great Molasses Flood and the Johnstown Flood disasters, I'll be honest, I didn't do a lot of extra research on them. I remember them from high school and college, but vaguely at best. The reason I settled on the flood is because I wanted to show that even when living in something as advanced as Arcology One, people still tend to push boundaries, even if they're doing so for completely benign reasons. The other floors weren't using their full allocation, so it gets diverted to the neighbor downstairs who have a better use for it. Old story where good intentions lead to bad outcomes.
Anyway, brevity isn't my strong suit. Thank you for reading, and let me know if you have any more questions.
Thanks for the brief reply (I say “brief” because you obviously have thought through a number of the issues my questions touched on and you didn’t open the floodgates). Once “Water” for example is set out more fully on Substack, what is the plan: books on Amazon/Kindle, or shopped to a traditional publisher, or some other distribution idea? Perhaps it is there already and I misunderstood that this is a WIP.
The plan is for it end up being a series of shorts. Might publish compilations later on once enough stories are written. But I’m a dad with two kids and a full time job, so that might take a minute.
But the goal is to explore different characters so readers can experience life in the Arcology from different angles.
The Asimov nod is a nice find, and the generational-thinking point hit hardest, no side to blame, just an honest diagnosis most governments share.
It makes me sad there's so little generational thinking these days, especially when the foundations we lay down now will have such far-reaching implications. We have such a hard time seeing beyond our own noses.
There are so many good points here. This balanced view is a rare one when it comes to AI. Your POV is refreshing.
Thanks Ken! Personally, I’m looking forward to the next installment of your piece. It’s getting good.
Cool interview with both insights into writing and also speculation about future technology.