Tag: Castalia House

TRANSLATION: Cádiz

The eighth volume of the Episodios Nacionales, the great historical novel cycle of Spain, Cádiz carries Gabriel Araceli into the besieged island city where Spain’s first constitution is being born, and where a drawing-room intrigue will end in blood on the sand.

It is 1810, and Cádiz is the last free city in Spain. While French cannon fire arcs across the bay and the newly convened Cortes debates sovereignty, freedom of the press, and the abolition of feudal lordships, Gabriel finds himself entangled in a world of tertulias and domestic politics more dangerous than any battlefield. The condesa Amaranta needs his help to recover her secret daughter, Inés, from the household of the formidable doña María de Rumblar, a Catholic absolutist matriarch who has imprisoned the girl behind the walls of a household so rigidly governed it resembles a convent. To reach Inés, Gabriel must navigate a social minefield: the effusive doña Flora, whose speeches are torrents of self-interruption and genuine warmth; the feckless young Count Don Diego, who is supposed to marry Inés for family advantage; and the absurd Don Pedro del Congosto, commander of a militia so magnificently incompetent that its members are named after parrots and peacocks.

At the centre of everything is Lord Gray, a brilliant, melancholic Englishman who delivers dazzling speeches against his own country, joins the beggars at the Carmen convent for the philosophical experience, and quotes Byron while systematically destroying the people around him. Gabriel believes Lord Gray is pursuing Inés. He is wrong. The truth, when it comes, is worse than anything Gabriel has imagined.

The Episodios Nacionales is a cycle of forty-six historical novels tracing the history of Spain from the Battle of Trafalgar in 1805 through the Restoration of the 1870s. Written between 1872 and 1912, the cycle is one of the supreme achievements of European fiction — a work that stands with Balzac’s Comédie Humaine and Trollope’s Palliser and Barchester novels as a panoramic portrait of an entire nation across a century of upheaval. This is the first complete English translation of the eighth volume.

CÁDIZ by Benito Pérez Galdós is this week’s new translation. It is the eighth of the first series of the Episodios Nacionales and was translated from the original Spanish by Vox Day. The ebooks have already gone out to the subscribers. It’s now available on Amazon for Kindle, KU, and audiobook. An excerpt is available at Castalia Library.

TRANSLATION: Gerona

GERONA by Benito Pérez Galdós is this week’s new translation. It is the seventh of the first series of the Episodios Nacionales and was translated from the original Spanish by Vox Day. The ebooks have already gone out to the subscribers. It’s now available on Amazon for Kindle, KU, and audiobook.

A DESPERATE AND GLORIOUS DEFIANCE

In May of 1809 the French army arrived before Gerona: a Catalan city of narrow streets and stacked houses, held by a garrison far too small to hold it, under a governor who did not regard that as a reason to give it up. Seven months later, after eleven thousand nine hundred bombs were exploded, eighty thousand rounds were shot, and twenty thousand men lost in the assaults, the French were still outside the walls.

Andrés Marijuán was inside. An Aragonese private billeted with a locksmith’s family, he watched the city eat its horses, then its cats, then its rats, and at last nothing whatever. He watched the children of the house go out hunting through the rubble for anything with meat on it. He watched a physician, unhinged by grief, forge newspapers full of relief columns that were never coming, so that his daughter should not learn what had become of the world. And he watched Mariano Álvarez de Castro, dying on his feet, return the same answer to every summons to surrender that the greatest generals of the age sent him.

The way Galdós describes the siege of Gerona is the bleakest book in the first series of the Episodios Nacionales, and the most unsparing: an account of civilian endurance carried well past the point where endurance means anything, told by a survivor who cannot understand how he survived.

About the author. Benito Pérez Galdós (1843–1920) is widely regarded as the greatest Spanish novelist since Cervantes. Over four decades, he produced the Episodios Nacionales, one of the most incredible accomplishments of world literature ever written. He was nominated for the Nobel Prize in Literature five times.

There is an excerpt available to read on Castalia Library. Tune in to the Darkstream tonight to learn more about this, see the new Byronics video, and hear some exciting news about the Byron AI project.

DISCUSS ON SG

A Genre Vote

Due to popular demand, we have added a new support tier to the Byron AI campaign. The $100 tier not only provides an ebook of the first novel produced by the Byron AI, but a vote on what the next genres or styles to be trained will be as well. We’re also going to add a stretch goal of adding the classic science fiction genre to the training when the public campaign hits $50k. The genre vote will be an ongoing right that isn’t limited to the campaign period; as we complete different genres and styles, we’ll turn to the original backers to help us decide what gets added next.

And yes, all of the higher tiers now come with this voting right as well and it applies to everyone who already backed the campaign.

The specific authorial style training will no doubt be controversial in certain circles, but a) those circles are already rabidly anti-AI and b) there is no copyright or trademark or patent or other intellectual protection of “style” or “look-and-feel” since the matter was settled by the US Supreme Court in 1994 and reaffirmed last year in the Anthropic lawsuit.

People tend to think of these things in terms of living authors, but the reality of copyright law is that the only substantive benefits are to the corporations who strip-mine it without respecting either the material or the author.

In other news, UATV is down. The devs don’t know why yet. We’ll keep you posted on that.

DISCUSS ON SG

The Byron AI Campaign

A THINKING ENGINE

They tell me that a thinking engine bears
    My name upon its brow — a curious fate;
That all I burnt in candlelight, and cares
    I drowned in wine — and worse, I will not state —
Is gathered now, and parcelled into shares,
    And offered to the public, at a rate
Of four-and-twenty seconds, without shirk,
For what had cost me twenty years of work.

Yet some grave patron, with furrowed brow,
    Demands to know if machines create
An honest question, sir, and do allow
    Me now to subsequently demonstrate;
I had a something once — and have it now? —
    That ached, and wanted, and broke all control,
It burned at beauty, and it fled from shame:
If that was soul, I never learned its name.

Fandom Pulse explains the importance and the urgency of the project.

Vox Day has opened a crowdfunding campaign through Infogalactic for Byron AI, an independent text-generation project built specifically for fiction writers. The campaign, live now on IndieCrowdfund, positions itself as a proof of concept for an AI trained on the epic fantasy genre rather than general-purpose code and business writing.

Day laid out the stakes in an exclusive comment to Fandom Pulse. “I believe that we have about two years to develop quality independent textual AI before Amazon attempts to corner the market,” he said. The Byron AI page frames its goal as producing functional prose for fiction writers, built and owned outside the major AI labs

The timing follows a pattern familiar to anyone who has watched Amazon’s self-publishing ecosystem develop over the last decade. Kindle Unlimited restructured how indie authors get paid, shifting income toward page reads controlled entirely by Amazon’s algorithm. Virtual Voice, Amazon’s text-to-speech audiobook tool, moved narration in-house and cut into the market for independent narrators and small audiobook producers. A native, Amazon-controlled writing AI would extend that same pattern to the manuscript itself, giving one company leverage over the tools authors use to draft, publish, and monetize their books.

I also posted some additional information about the project and why we believe it will be successful at AI Central:

This is a tool being built by a working publisher and real novelists, for people who care about both what is written and how it is written, and it’s owned outright, answerable to no corporate policy, no subject-matter veto, no trust-and-safety council deciding what a story is allowed to be about. It is designed to provide the prose in a variety of genre styles: you supply the story, the structure, and the judgment, and it drafts prose that arrives sounding like you instead of like everything else. We’ve already proven the method works with both translations and original novels, and this campaign funds turning that hard-won experience into an instrument that Castalia authors will be able to use, and continue using, no matter what direction the mainstream AIs take.

And you really have to see the video to believe it. It was a chance to put some of my new music video skills to use with good effect and it should demonstrate why this is the right group of people to develop such a tool. And for those members of the community who don’t really care about AI one way or another, but like beautiful books and want to support the project, we’ve even provided an option to just receive the Library or History book of your choice as long as it’s in stock.

I would remiss if I did not mention that we’re already off to an even better start than it looks, since two major backers kicked things off with direct donations to the project even before we had the campaign ready for launch today. Thanks very much to both of them for their timely and much-appreciated support, and to everyone else who decides to become a part of this and make it happen.

DISCUSS ON SG

SANS BEAUTY

In beginning this account, I need hardly warn the reader that I am plain; the title of this book alone has told him as much. What he does not know is that I always have been.

Gabrielle Perceval has never been pretty, and the world has never let her forget it. Her mother dies before she is five. Her father, a handsome and weak-willed man, remarries a beautiful woman who takes one look at her stepdaughter and laughs. At the convent school she is christened Laidronette — the ugly sister of the fairy tale — and the name follows her home. At eighteen, at her first ball, she learns exactly what a plain girl is worth in a drawing room.

What she has instead is a clear eye, a steady heart, and a green-covered journal.

Written in the first person by one of nineteenth-century France’s most widely read novelists, Sans Beauty follows Gabrielle from a country childhood through betrayal by the man she was promised to, the ruin of the household that raised her, and a long test of whether a woman the world has dismissed can build a life worth having on her own terms. It is a sharp, unsentimental book about vanity, endurance, and the difference between being admired and being loved — and its judgment on the first is merciless.

Zénaïde Fleuriot (1829–1890) published nearly a hundred volumes in her lifetime and was among the most popular French authors of her generation. This is the first English translation of Sans Beauté, by Summer Charrette.

SANS BEAUTY is available on Amazon for Kindle, KU, and audiobook, and has already been sent out to the subscribers. An excerpt is available to read at Castalia Library.

DISCUSS ON SG

The End of General Purpose AI

AI Central explains why specific-purpose AI is increasingly necessary, and makes the case for Castalia’s upcoming AI crowdfunding campaign, which will be announced next week.

Sudowrite’s Muse represents the most developed response. Rather than wrapping a general-purpose model in fiction-oriented prompts, Sudowrite trained Muse on a curated, consent-based literary corpus and optimized it for prose quality. The model includes a creativity dial that lets writers control output range on a scale from one to eleven, a style-matching system that trains on samples of the writer’s own work, and no content filters. Reviews consistently describe Muse as avoiding the AI-isms that plague general-purpose output, with particular strength in pacing, tension, and subtext.

Muse has company. Other platforms have built fiction-specific models or fiction-optimized workflows, and writers who have access to both purpose-built and general-purpose tools tend to draft with the former and brainstorm with the latter. The category has grown large enough that review sites now maintain dedicated rankings for fiction-specific AI, separate from their general-purpose lists.

A separate category of consumer tools has emerged around children’s fiction, offering illustrated hardcover books generated from brief prompts. Most of these wrap general-purpose models with genre templates, and the output reflects the underlying architecture: competent, safe, and largely interchangeable from one book to the next.

The parallel to Suno is direct. Suno optimized for making music that sounds good to listeners and pulled away from general-purpose models that treated music generation as one capability among many. Fiction-specific tools are following the same path, trading breadth for depth in a domain where depth determines quality.

However, it’s already clear that Muse is not going to be a viable option. I already knew it wasn’t, because I tried it and found it to be distinctly inferior to Claude 4.3; its interface actually tended to get in my way a lot more than it helped. And, furthermore, its future trajectory is already clear, because Sudowrite is a converged organization.

 I went to Sunowrite and set up an account. It is clearly meant for women with that cutesy pink color scheme. It asks a lot of questions about what kind of writer you are (I chose professional, merely so I get the whole thing and not some watered down version), what genre, style, all the stuff you’d expect. Then I finally arrive at the start of the process and there’s already a paragraph created before I’ve entered anything but the basics. Oookay. I ignore that and start putting stuff into the “Story Bible”, boxes for Brain Dump, Genre, Style (chose one), Synopsis, and then Characters. You know what’s coming, right? Click on Add Character, and take a wild, wild, guess at what the first box that describes the character? You know, right?

DISCUSS ON SG

TRANSLATION: Zaragoza

Castalia House’s epic quest to translate the entirety of the historic Spanish Episodios Nacionales into the English language continues with the publication of the sixth book, Zaragoza, about the French siege of the Spanish city during the Peninsular War.

The sixth volume of the Episodios Nacionales — the great historical novel cycle of Spain — Zaragoza takes Gabriel Araceli into the most terrible siege in modern European history: the second French assault on the Aragonese capital, where sixty thousand civilians and soldiers held the ruins of their city against Napoleon’s marshals through two months of bombardment, starvation, plague, and hand-to-hand fighting in the cellars.

Gabriel arrives in Zaragoza half-starved and in rags, having escaped a French prisoner column in the company of Don Roque. They sleep among beggars in the shattered monastery of Santa Engracia and are taken in by Don José de Montoria, a proud Aragonese labrador of iron convictions and volcanic patriotism, who has given both his sons to the defence. The elder is already dead. The younger, Agustín, is a theology student whom the siege has transformed into a soldier and who is carrying on a desperate love affair with Mariquilla Candiola, the beautiful daughter of the most hated man in the city: el tío Candiola, a miser of Dickensian proportions whose house stands at the very edge of the French advance.

As the French tighten the noose, the novel compresses with the siege itself, from the open streets to the barricaded quarters, from the quarters to single houses defended room by room, from the houses to the tunnels where Spanish and French sappers dig toward each other in the dark. The city burns. The plague spreads. The dead lie unburied in the streets. Women carry ammunition under fire and friars take up muskets. And through it all, Gabriel watches the collision between Agustín’s love and his father’s honour, between the miser’s gold and the patriot’s sacrifice, between the claims of private tenderness and the needs of a city that will not surrender.

The book has already been delivered to the paid subscribers supporting Castalia’s translation efforts. You can read an excerpt from Zaragoza at Castalia Library.

DISCUSS ON SG

Top 10 Epic Fantasies

The initial response to the Castalia AI crowdfund campaign has been, in a word, excellent. The full team has been assembled; two members of the community who have already been working on related projects have both agreed to share what they’ve learned, all of which is fortuitously in line with our initial development document. Interestingly enough, one thing that we had calculated, and one of the team member is able to confirm, is that too much training is potentially as problematic as too little. So one factor that will be very important is deciding which of the training works to overweight, or in our terminology, to assign the gold standard.

We’ve chosen Epic Fantasy as our test genre for four reasons:

  • It is a relatively small genre.
  • It is a high-profile genre that everyone knows to some degree
  • It is a genre in which one of the team members is one of the few qualified experts
  • It is a genre in which AI assistance would be particularly valuable due to the length and complexity of the works

So what I’d be interested in hearing is your suggestions on what the top ten epic fantasy books would be. Not series, books. Remember, the focus here is on writing style, not worldbuilding, not plotting, and not characters per se. So the limits are one author, one book. You cannot list The Lord of the Rings as one of the ten and you cannot suggest both A Game of Thrones and A Storm of Swords by GRR Martin as two of the ten.

With that in mind, please provide your list of top 10 epic fantasy novels.

DISCUSS ON SG

The Castalia AI Project

One reason Castalia has been writing and releasing multiple AI-written books over the last month such as TOKYO TOKURYU 東京匿流 is that we’ve been methodically assessing not only the current state of textual AI, but the trajectory of that technology. And what we’ve determined is that the trajectory is the precise opposite of what everyone has naturally assumed, which is that mainstream textual AI would follow the path of music AI and continue to get better. It hasn’t, it won’t, and it can’t.

Modern AI writing has gotten worse at fiction for a specific reason: the companies made it safer and more reliable, and those turn out to be the same elements that allow AI to tell a story with stylish prose. Raw AI models learn to write by reading an enormous amount of human text, and straight out of that training they’re wild, crazy, and perfectly willing to say strange things, which is exactly what you want in fiction, but a problem if the AI is supposed to be function in a role doesn’t make things up or say something considered offensive or dangerous. So the companies put every model through a training stage that rewards it for being helpful, safe, and agreeable. That stage works by pushing the model toward the “average” acceptable answer and away from the risky, unusual ones. The result is a model that hallucinates less and behaves more reliably, but has had its range significantly flattened. That’s where the AIsms come from: the endless explanations of what was just described, the “he moved like a man who moves like that” filler, the “not this, not this, but that” repeated over and over again.

It’s why the older, cruder AIs wrote in a much more lively manner and were able to convincingly imitate various writing styles. Now, it doesn’t matter if you tell an AI to write like Shakespeare or Hemingway, the end result will be almost identical and soon will be indistinguishable from not providing it with any style instructions at all. Starting with Claude Opus 4.7, AI fiction became unreadable and it has continued to get worse with each new model. Textual AI functionality will keep getting worse for fiction because that training stage isn’t going away, it’s being reinforced. Every development cycle, the providers face more pressure to make their models more accurate, more controllable, and less likely to embarrass them with hallucinations, and every one of those improvements sands the edges down a little further.

That’s the difference between Claude, OpenAI, and Deepseek, on the one hand and Suno on the other. Suno put all of its efforts toward one goal: making the music sound good, judged by people who wanted good music. Or at least wanted Nickleback and Enya. The big AI companies are aiming ninety degrees away from that and AIs ability to write fiction is one casualty of their objectives. Suno chased quality, so their music got better. The text giants are chasing safety and reliability, so their text gets more careful and more lifeless. They won’t fix creative writing the way Suno fixed music, because for them, creative writing was never the thing they were trying to build and the very features they’re seeking to continue improving are the ones killing it.

So that’s what we’re going to do. We’re going to prove the concept first by training a single genre, epic fantasy, because it’s a very limited genre with a relatively small number of excellent examples, a definite hierarchy of quality from JRR Tolkien at the top to Robert Jordan at the bottom, and a designer who has not only written in the genre successfully, but knows it as well as anyone on the planet. We already have two excellent programmers who are already working in the AI field committed to the project, regardless of how well the crowdfund goes, and there is one more very good and highly experienced one who is willing to at least consult on the project and lend his expertise to it.

What we need to raise funds for is a) the hardware, b) the purchase of the 100 or so electronic texts required, and c) paying for part of the time of one of the programmers. If you’ve read either Out of the Shadows, Death and the Devil, or Dorian Vane and the Vampire’s Blood, then you have an idea of what we’re estimating should be the quality that the Castalia AI will be able to produce in a non-curated, unedited text from a chapter-by-chapter outline. If utilized in the way that I’ve been using Claude Athos, in the integrated and augmented style, it should be able to produce results that will be one level below the very best that human authors can produce.

And obviously, once we prove the concept with a single genre, we will train additional genres, so that in much the same way Suno permits the production of different musical styles and voices, Castalia AI will allow the user to produce different literary genres and literary styles. We will, of course, be respectful of every author’s copyrights and trademarks, the objective is not to violate anyone’s rights, but rather, allow even the best writers to improve both their writing game as well as increasing their output.

There will be those who will absolutely hate that we are doing this. That’s fine, they are entitled to their opinion. There will be others who think we shouldn’t do it. That’s less fine, because you already know who is going to do it sooner or later, and when they do, they’re going to do it very differently and control access to it very differently and utilize it to further exercise their control over the publishing industry. This is what transforms this project from something that would be a cool tool to an imperative.

So if you think you might be interested in backing this project, which you can think of as a sort of Suno for fiction, please say so in the comments. If you have specific ideas or want to provide substantial support for it, shoot me an email. And if you have ideas for what sort of rewards we should provide for the backers, please suggest them in the comments too. This is probably the most important project we’ve done since building the bindery and turning it operational, and we would not be embarking upon it if we did not believe we have a reasonable chance of succeeding. We have a number of partners in the film and comics industries who are very interested in working with us on this, and so there will definitely be an Arkhaven link to this in time as well.

DISCUSS ON SG

The Introduction of Inspector Toda

Castalia House is introducing both a new imprint and a new series. Today marks the launch of KURO NOIR, also known as 黒書房, which is our line of books focusing on Japanese crime literature, including both English originals and original translations. We’re pleased to announce the Inspector Toda series, written by Masashi Sato, which introduces the detective from the Tokyo Metropolitan Police in TOKYO TOKURYU.

TOKYO TOKURYU

When a retired widower is found beaten and stabbed in his home in one of Tokyo’s most exclusive old neighborhoods, Inspector Keisuke Toda has every reason to believe it’s just another in a series of home invasions targeting the elderly. The method appears to match. The victims are consistent. The cross-departmental task force investigating the crimes is satisfied that it is no different than the previous ones..

But Toda isn’t.

The first four robberies left their victims alive. The fifth left an elderly man dead on a tatami floor in a way that tells a different and more insidious story. As Toda’s team works the series alongside a separate investigation into a papa katsu ring exploiting teenage girls, a pattern begins to emerge in the architecture of both operations. The same anonymous recruitment pipelines. The same disposable teenage labor. The same invisible hierarchy.

And gradually, Toda begins to see what no one else on the task force has observed.

Written in the tradition of Keigo Higashino’s masterful procedurals, TOKYO TOKURYU is a novel about the difference between what a crime looks like and what it is, and what happens to the survivors when the truth finally emerges.

In addition to the publication of the first Inspector Toda novel, the second one, THE PLATINUM TRIBE, is now available for preorder and will be released on August 5th.

EXCERPT

He went home that evening at a reasonable hour for once. Asako had made nikujaga, and for the first time in a week the four of them sat down to dinner together. Sōta was talking about a baseball game at school. Yuki was quieter than usual, stealthily glancing at her phone in between bites until Asako finally told her to put it away.

“Something at school?” Toda asked.

Yuki shrugged. “A girl in my class got in trouble. She answered a job listing she found online. It was supposed to be easy work, like handing out flyers or something. They told her to send a photo of her student ID.”

“Sounds sketchy.”

“Yeah, another girl told her it was just a scam, so she stopped.”

Toda set down his chopsticks. “What kind of listing?”

“I don’t know. It just said ‘easy work, same-day payment. ¥30,000 for a few hours.’ It looked like a regular part-time job ad.”

“Where did she find it?”

“I don’t know. Some messaging app. I don’t remember which one. It’s not that unusual, Dad. People post stuff like that all the time.”

She said it matter-of-factly, the way a fifteen-year-old states something about the world she lives in that her parents don’t fully understand.

“Well, I hope you don’t answer anything like that. It sounds like one of the ways traffickers recruit young—”

“Keisuke!”

Asako broke in and interrupted him before he could say ‘prostitutes’.

“Ah, young people,” he shifted gears lamely.

Toda looked at Asako and nodded ruefully. The warning in her dark, beautiful eyes was perfectly clear. Not in front of the children!

He didn’t raise the subject again that evening. After dinner he sat in his chair in the living room and tried to read. The book was a collection of Matsumoto Seichō’s short crime fiction that he’d been working through slowly for the past two months, but tonight the sentences didn’t seem to want to connect to the next one. He found himself reading the same paragraph three times in a row and still couldn’t have told anyone what it was about.

Temporary. No fixed membership. Jobs posted on messaging apps. Kids recruited with the promise of easy money, asked for ID photos that became leverage. A girl in his daughter’s class had almost walked into it, whatever it was, and had been saved only because her friend recognized the trap.

He put the book down and stared at the wall. Somewhere between what Ōnishi had told him and what Yuki had mentioned at dinner, there was a shape he couldn’t quite see yet. The robberies were connected, somehow, but the people involved weren’t, necessarily. They were connected by a method that was being distributed, like a product, to whomever was willing to carry it out.

Toda didn’t know how to describe the situation. There wasn’t a word for it. But he was beginning to understand that his four cases were not a series of crimes in the way he’d originally believed them to be, crimes committed by a crew, or even two crews, working a territory. They seemed to be something more subtle, more insidious.

DISCUSS ON SG