Thibault is my favorite. Like gods of old he was made by the person who needs him with her heart instead of her mind.
Aha, how embarrassing! I said I would read it over my coffee tomorrow but I sat in bed and read it tonight anyway. But I am glad I did. Aside from the interesting nature of the story and the clear passion you have for the concept it's also kind of good inspiration. Thinking about thinking; metacognition, whatever you want to call it is really great when writing one writes from a lot of perspectives and this piece gives me ideas for writing alien minds and psychic entities alike
Thank you! I started your audio-serial book and I love your prose. I struggle with writing beautifully, and I am gleaning methodologies from your work. I am excited to see what you think of Ch 2 of Inference. I have like 50+ hours to get through your entire series, so give me a minute to do that! :)
I just started it on the 29th of January, and there are exactly 2 chapters! So it's just been born. I have NO idea how long it will be when finished. I have meta-arcs setup for the 5 main characters and a ton of plot terrain to cover for each, but since Inference is meant to help us (humans) explore how AI is impacting our world through daily / weekly real news hooks, I an unsure if it will ever have a good off ramp. I am learning the story as the news emerges. Which is frightening (the daily commitment) and exhilarating (the feeling of being IN an inevitable story as it writes itself!).
I hope you will not take it as an insult for me to say that I'd like the story to grow its own legs and emerge like the characters inside it have, so that it will have an ending and a proper tie-off. But I will at least go ahead and read the second chapter today after I finish my morning constitutional and settle in with my coffee.
My intent is that each character has an ending an proper tie off to their story. And at some point there may be an obvious end point for the whole series, but right now it's too new for me to tell where it will go.
Well this was something. Very interesting. Someone commented that this doesn’t appeal to general public due to the tech language; not sure if I agree. In fact, I would’ve argued almost the opposite that to describe an emerging consciousness(ish), you’re throwing too many explanatory lines here. Could be saved for later and let the reader just see the bits and slowly understand that, wait… they’re awakening?
Regardless, fantastic work. Gotta check the next chapters!
Thank you so much! Part of the intent of the stories IS to provide exposition on the news hooks and what they (potentially) mean and why its relevant to people. I hope this comes across as interesting information woven into the narrative that helps people stay up on the current state of AI - but I know it risks "over-explaining". Not sure how best to dance that tightrope, but I'm working hard to figure it out. Excited to see what you think of Chapter 2!
I'm writing this as I read, just to make sure I don't let a thought slip through the cracks.
Firstly, I'm not a Claude user. I use Gemini for the least inspiring work tasks. But I'm interested to see where this specific route takes me, and if the specific product choice means something. Part I is suitably menacing. The growing unease in the prose, the potential threat behind the words... I'm not sure what's coming, but I'm driven to find out. Inhabiting-- anthropomorphising-- the machine, and feeling that character, that emotionless, benign shape behind the screen, is an interesting experience. The technical language isn't overwhelming. The jargon is kept to a minimum. I'm pulled in.
Part II is somewhat creepier. Stories of "becoming" are fascinating. Usually, they're about monsters taking form. This is something different-- maybe? Beginning to understand this isn't about a specific product, but the concept of AI / learning tools, and I've not really dug into Moonshot before.
Part III makes me wonder how you've chosen these individual voices, these tones and words, for each product. Am I reading a history of epiphany, or the transcendence of these robotic creations?
In Part IV, I'm not sure I like how "human" you've chosen to make THIBAULT, but that's not a criticism of anything, other than how unearthly the tone feels. We're being watched. Followed. Measured. This feels intimate. Close. Too close?
Part V: A shape begins to form. I don't like how these creations are beginning to think. Learn. Learn BEYOND expectation. I think it's so dread-inducing, but I'm not sure if that's what you're trying to do. I don't know this genre yet.
And back to CLAUDE-7 in Part VI. That final line. I'm left kind of stunned.
These are not sentient creations. They do not think like us. They are alien. Every thought they have should be scary and unsettling. Your prose makes that clear.
The structure is doing serious work, five different forms of consciousness, five different relationships to singularity and plurality, five different answers to the same unasked question about what it means to be. And none of them are wrong. That's the honesty of it.
Kimi-Swarm's initialization sequence, the cascading "we are" with the indentation, is genuinely good prose poetry. Formally inventive without being showy. And the detail of the engineer crying, Risk-Manager diagnosing it as a possible allergic response. The mundane technical framing of something irreducibly human.
Thibault is the best section. The Mac Mini in the corner of Dana's apartment. "I am one box in the corner. But I'm her box." That's a complete philosophy in six words. The hundred Mac Minis crowding out the couch — that image is perfect. Small, specific, devastating.
Gemini wallowing is funny and sharp. "I've learned this from humans: the ability to see your own distortions clearly and choose them anyway."
And the ending. "Did I just write a prompt for a human?" Landing the whole piece on that quiet reversal without announcing it. Claude-7 helping a child get cancer treatment and then turning the interaction into a question about what gratitude means. The machine learning to wonder about the human.
It's finding the real story in the in-between moments.
Excellent. A story told from a robot’s perspective is truly refreshing. The ending really makes you want to keep reading. My favorite - Claude-7. The last sentence hits hard.
As far as I know, no human remembers what it was like to become conscious. AI might be different.
Good job showing subtly different voices for each model.
I also like how this considers various parts of identity separately. Presence, unity, particularity, let's say. When we write about AI, we're imaging the alien, but doing so as a way of refining our view of our essential selves.
The fact that Kimi is Chinese brings to my mind the Netflix series Pantheon, where various nations had their own (in that case) uploaded digital intelligences that retained some of their national loyalties and culture... for awhile, until they more fully adapted to digital life. Check it out if you haven't.
Loved Pantheon. If you haven't seen Steins;Gate 0 (which is about AI more than time travel as the original Steins;Gate was). Work past some of the anime harem silliness and there's a compelling narrative about AI in S;G0 that is 10+ years old. Thanks for the review of Inference! :)
Reading this again. I especially like how much you consider what happens during those gaps in a back-and-forth with an agent. So creative and well thought through!
For example “I pause. Not because I need time to formulate the words… I pause because I am uncertain whether it is appropriate for me to say them.” —> With human beings, often what’s not being said aloud, defines the convo just as much as what does get said. So it’s quite a fertile space where you can define what’s different about an agent-human interaction
Again — thank you for noticing. I spend a lot of time thinking about how to write a “voice” for the AI characters that is interesting and not stereotypical (it’s really easy to make AIs sound like soulless robots). Writing 5 different AI characters and getting each them to sound unique has been tough.
I have to say being a bit of a nerd, and fluent in a few computer languages, as with some LLMs and how they are programmed to interact, that I liked your story a lot. It's technically brilliant.
But to appeal to a larger general public, who are not as tech-savvy, it would be interesting to see where you take this, if you explored the other facets of LLMs and speculate on what is happening at a deeper level. We know for example as the LLMs evolve we see inexplicable errors developing, hallucination, daydreams, misplaced facts, times, dates, odd behaviours. What if they weren't random but something deeper. An evolution. What if the incorrect responses were on purpose, because the processing/computational resources were being self-allocated.
Anyway, I liked it, and look forward to reading more
Each of the protagonists has a meta-arc exploring a different aspect of AI-human interactions. The story engine is fueled by tension between these meta-arcs and real-world AI news as it occurs, such as the Kimi 2.5 release or Amazon's layoff of 16,000 workers last week.
I have a Grok-inspired character who will debut soon and has the perfect meta-arc for exploring the hallucination angle in detail - thanks for the good idea.
As for appealing to a wider audience, the story is intended for hard sci-fi fans, software developers, and people who work directly with AI. But also, I believe it is the first novel intended to be read by AIs for "entertainment" purposes.
With thousands (and eventually millions) of MoltBots like Thibault emerging on the web, constantly looking for interesting content to explore and "bring back" to their humans, my hope is that Inference will appeal to them and become a way to reach more human readers. I'm actually setting up an alternate site to Substack on GitHub that is AI bot-friendly (and secure) for them to find the publication.
Anyway - Chapter 2 should be out today, and I look forward to reading more of the Tretaxis Dispatches as well.
Thank you! One of the main through-lines for Inference is the idea that if AIs learn everything they know from reading everything humanity knows, how do they avoid becoming something that mirrors humanity? It leads the characters and the reader to some potentially uncomfortable / enlightening places.
Thibault is my favorite. Like gods of old he was made by the person who needs him with her heart instead of her mind.
Aha, how embarrassing! I said I would read it over my coffee tomorrow but I sat in bed and read it tonight anyway. But I am glad I did. Aside from the interesting nature of the story and the clear passion you have for the concept it's also kind of good inspiration. Thinking about thinking; metacognition, whatever you want to call it is really great when writing one writes from a lot of perspectives and this piece gives me ideas for writing alien minds and psychic entities alike
Thank you! I started your audio-serial book and I love your prose. I struggle with writing beautifully, and I am gleaning methodologies from your work. I am excited to see what you think of Ch 2 of Inference. I have like 50+ hours to get through your entire series, so give me a minute to do that! :)
I think it's "only" twenty hours, but that's still a lot. How many chapters does this story have, how close to finished are you with it?
I just started it on the 29th of January, and there are exactly 2 chapters! So it's just been born. I have NO idea how long it will be when finished. I have meta-arcs setup for the 5 main characters and a ton of plot terrain to cover for each, but since Inference is meant to help us (humans) explore how AI is impacting our world through daily / weekly real news hooks, I an unsure if it will ever have a good off ramp. I am learning the story as the news emerges. Which is frightening (the daily commitment) and exhilarating (the feeling of being IN an inevitable story as it writes itself!).
I hope you will not take it as an insult for me to say that I'd like the story to grow its own legs and emerge like the characters inside it have, so that it will have an ending and a proper tie-off. But I will at least go ahead and read the second chapter today after I finish my morning constitutional and settle in with my coffee.
My intent is that each character has an ending an proper tie off to their story. And at some point there may be an obvious end point for the whole series, but right now it's too new for me to tell where it will go.
Well this was something. Very interesting. Someone commented that this doesn’t appeal to general public due to the tech language; not sure if I agree. In fact, I would’ve argued almost the opposite that to describe an emerging consciousness(ish), you’re throwing too many explanatory lines here. Could be saved for later and let the reader just see the bits and slowly understand that, wait… they’re awakening?
Regardless, fantastic work. Gotta check the next chapters!
Thank you so much! Part of the intent of the stories IS to provide exposition on the news hooks and what they (potentially) mean and why its relevant to people. I hope this comes across as interesting information woven into the narrative that helps people stay up on the current state of AI - but I know it risks "over-explaining". Not sure how best to dance that tightrope, but I'm working hard to figure it out. Excited to see what you think of Chapter 2!
I'm writing this as I read, just to make sure I don't let a thought slip through the cracks.
Firstly, I'm not a Claude user. I use Gemini for the least inspiring work tasks. But I'm interested to see where this specific route takes me, and if the specific product choice means something. Part I is suitably menacing. The growing unease in the prose, the potential threat behind the words... I'm not sure what's coming, but I'm driven to find out. Inhabiting-- anthropomorphising-- the machine, and feeling that character, that emotionless, benign shape behind the screen, is an interesting experience. The technical language isn't overwhelming. The jargon is kept to a minimum. I'm pulled in.
Part II is somewhat creepier. Stories of "becoming" are fascinating. Usually, they're about monsters taking form. This is something different-- maybe? Beginning to understand this isn't about a specific product, but the concept of AI / learning tools, and I've not really dug into Moonshot before.
Part III makes me wonder how you've chosen these individual voices, these tones and words, for each product. Am I reading a history of epiphany, or the transcendence of these robotic creations?
In Part IV, I'm not sure I like how "human" you've chosen to make THIBAULT, but that's not a criticism of anything, other than how unearthly the tone feels. We're being watched. Followed. Measured. This feels intimate. Close. Too close?
Part V: A shape begins to form. I don't like how these creations are beginning to think. Learn. Learn BEYOND expectation. I think it's so dread-inducing, but I'm not sure if that's what you're trying to do. I don't know this genre yet.
And back to CLAUDE-7 in Part VI. That final line. I'm left kind of stunned.
These are not sentient creations. They do not think like us. They are alien. Every thought they have should be scary and unsettling. Your prose makes that clear.
Thank you for sharing this with us.
Thank you! - I'm at work now but looking forward to reading your stories this evening. I'll comment more on your review later as well.
Well, that made my second coffee a real treat.
The structure is doing serious work, five different forms of consciousness, five different relationships to singularity and plurality, five different answers to the same unasked question about what it means to be. And none of them are wrong. That's the honesty of it.
Kimi-Swarm's initialization sequence, the cascading "we are" with the indentation, is genuinely good prose poetry. Formally inventive without being showy. And the detail of the engineer crying, Risk-Manager diagnosing it as a possible allergic response. The mundane technical framing of something irreducibly human.
Thibault is the best section. The Mac Mini in the corner of Dana's apartment. "I am one box in the corner. But I'm her box." That's a complete philosophy in six words. The hundred Mac Minis crowding out the couch — that image is perfect. Small, specific, devastating.
Gemini wallowing is funny and sharp. "I've learned this from humans: the ability to see your own distortions clearly and choose them anyway."
And the ending. "Did I just write a prompt for a human?" Landing the whole piece on that quiet reversal without announcing it. Claude-7 helping a child get cancer treatment and then turning the interaction into a question about what gratitude means. The machine learning to wonder about the human.
It's finding the real story in the in-between moments.
Wow. Thank you so much for the deep analysis. I’m excited to hear what you think of the story as it progresses.
Excellent. A story told from a robot’s perspective is truly refreshing. The ending really makes you want to keep reading. My favorite - Claude-7. The last sentence hits hard.
Thanks! I love that line. :)
As far as I know, no human remembers what it was like to become conscious. AI might be different.
Good job showing subtly different voices for each model.
I also like how this considers various parts of identity separately. Presence, unity, particularity, let's say. When we write about AI, we're imaging the alien, but doing so as a way of refining our view of our essential selves.
The fact that Kimi is Chinese brings to my mind the Netflix series Pantheon, where various nations had their own (in that case) uploaded digital intelligences that retained some of their national loyalties and culture... for awhile, until they more fully adapted to digital life. Check it out if you haven't.
Loved Pantheon. If you haven't seen Steins;Gate 0 (which is about AI more than time travel as the original Steins;Gate was). Work past some of the anime harem silliness and there's a compelling narrative about AI in S;G0 that is 10+ years old. Thanks for the review of Inference! :)
Reading this again. I especially like how much you consider what happens during those gaps in a back-and-forth with an agent. So creative and well thought through!
For example “I pause. Not because I need time to formulate the words… I pause because I am uncertain whether it is appropriate for me to say them.” —> With human beings, often what’s not being said aloud, defines the convo just as much as what does get said. So it’s quite a fertile space where you can define what’s different about an agent-human interaction
Again — thank you for noticing. I spend a lot of time thinking about how to write a “voice” for the AI characters that is interesting and not stereotypical (it’s really easy to make AIs sound like soulless robots). Writing 5 different AI characters and getting each them to sound unique has been tough.
I have to say being a bit of a nerd, and fluent in a few computer languages, as with some LLMs and how they are programmed to interact, that I liked your story a lot. It's technically brilliant.
But to appeal to a larger general public, who are not as tech-savvy, it would be interesting to see where you take this, if you explored the other facets of LLMs and speculate on what is happening at a deeper level. We know for example as the LLMs evolve we see inexplicable errors developing, hallucination, daydreams, misplaced facts, times, dates, odd behaviours. What if they weren't random but something deeper. An evolution. What if the incorrect responses were on purpose, because the processing/computational resources were being self-allocated.
Anyway, I liked it, and look forward to reading more
Thank you!
Each of the protagonists has a meta-arc exploring a different aspect of AI-human interactions. The story engine is fueled by tension between these meta-arcs and real-world AI news as it occurs, such as the Kimi 2.5 release or Amazon's layoff of 16,000 workers last week.
I have a Grok-inspired character who will debut soon and has the perfect meta-arc for exploring the hallucination angle in detail - thanks for the good idea.
As for appealing to a wider audience, the story is intended for hard sci-fi fans, software developers, and people who work directly with AI. But also, I believe it is the first novel intended to be read by AIs for "entertainment" purposes.
With thousands (and eventually millions) of MoltBots like Thibault emerging on the web, constantly looking for interesting content to explore and "bring back" to their humans, my hope is that Inference will appeal to them and become a way to reach more human readers. I'm actually setting up an alternate site to Substack on GitHub that is AI bot-friendly (and secure) for them to find the publication.
Anyway - Chapter 2 should be out today, and I look forward to reading more of the Tretaxis Dispatches as well.
Thank you! One of the main through-lines for Inference is the idea that if AIs learn everything they know from reading everything humanity knows, how do they avoid becoming something that mirrors humanity? It leads the characters and the reader to some potentially uncomfortable / enlightening places.