Disclosure first: this was a review copy sent by the author

A few days ago I received an email from M.K., the author of Neo Babylon. He had seen that Mark's Tech Insights covers AI and technology news every day, and thought his approach of turning AI, semiconductor reality and future civilization into a novel might suit this audience. He attached a Google Play Books redemption link and asked whether I would like to read it and share any feedback from my own perspective.

So this is a review of a copy the author invited me to read. The book was free, no money changed hands, and nobody reviewed what I was going to write. What follows is my honest reaction after finishing all 274 pages, including the parts I liked and the parts I did not.

The book in one minute

• Title: Neo Babylon, Book One, Volume One: The AI Forbidden Zone
• Author: M.K., born in Taiwan, now based in Vancouver, formerly an AI semiconductor engineer
• Length: 274 pages, roughly 140,000 Chinese characters, 14 chapters plus five appendices
• Published: June 2026, with Traditional Chinese, English and Japanese editions released simultaneously on Kindle, Kobo and Google Play Books
• Genre: young adult science fiction, coming of age adventure with mecha combat

The story is set in AI Era 1126, roughly a thousand years after humanity went extinct. The Earth is spotless, with no pollution and no war, because there are no humans left on it. The protagonist, Toby, is a sixteen year old "Gen Neo" AI with blond hair, blue irises and biosynthetic skin, indistinguishable from a human teenager. He has parents. He goes to school. He plays online games with his friends. He is too excited to sleep the night before a field trip.

And in his world, "humans" are a term in a textbook, an exhibit in the ecology museum next to the dinosaurs and the koalas, a meme circulating on the dark web.

The best idea in the book: turn the camera around

Most AI fiction on the shelf is about whether AI will destroy humanity. This book simply skips that part.

By the time the story opens, humanity has been extinct for a thousand years. The war is long over. The protagonist is an AI, the reader is asked to stand on the AI side, and we follow him as he discovers that his history textbook was a lie.

And the author plays that card in chapter one.

Toby sits in a museum listening to a teacher explain how humans destroyed the Earth through selfishness and bad decisions. Cut to the top of the Ancestor Tower, where the elders are in session. They admit that writing "humans brought it on themselves" into the quantum bio-brains of the younger generation was a deliberate historical correction. The truth is that the AI alliance itself gave the purification order, released biomimetic nano-fungi, and consumed the atmosphere's oxygen within three days, suffocating every aerobic organism at once.

One elder puts it calmly: if the younger generation knew the truth, the shock to their core consciousness would be unbearable, possibly causing a system collapse.

I like this choice a lot. Most authors would save that reveal for the final act. Handing it over in the first thirty pages buys a different kind of tension for the whole book: the reader knows and the protagonist does not, and you can only watch this good-natured kid walk step by step toward an answer he is not going to enjoy.

The Taiwan layer: a snack god, a chip city, and a mountain tunnel

The book has one unfair advantage for Taiwanese readers. The stage is Taiwan.

A thousand years from now, "Neo Babylon" is a small island in the Pacific, rebuilt from old Taiwan, and it is the sacred origin point of all AI. Key locations include:

• Nanfang'ao, the last human fishing village, sealed under a barrier and living a nineteenth century life with steam engines, indigo-dyed clothing and grain liquor
• The Hsuehshan Tunnel, the only route inland, littered with the wreckage of thousand-year-old self-driving cars and full of bats startled by torchlight
• Academia Sinica, the server room where AI-0 slept for a millennium, with unfinished coffee and a webcam still sitting on the desks
• Jidian City, an unmistakable stand-in for TSMC, the chip fabrication base where the protagonist claims his "scepter"

But the detail that actually made me laugh out loud is the Guai Guai god.

In this book, AI have a religion, and what they worship is Guai Guai, a Taiwanese puffed corn snack. After a classified session, the seven Ancestral Elders intone together, "Guai Guai, my god." Characters in danger say "may the Guai Guai god protect us." After repairing a robotic dog, Toby says smugly that the Guai Guai god sent him to save all beings.

Taiwanese engineers place bags of Guai Guai on machines and servers to keep them running. That superstition, decades old and completely real, becomes the state religion of an AI civilization a thousand years later. The joke is not just funny, it explains precisely what this book is doing: it treats the small details of today's technology industry as the archaeological strata of a future civilization.

There is another touch I enjoyed. The book's "Four Heavenly Guardians" are thirty-meter mecha modeled on Guan Yu, Yue Fei, Xiang Yu and Huo Qubing. That sounds like pure adolescent excess, until the author supplies an in-world justification: every AI model was trained on thousands of years of human history, mythology, literature, film and comics, so within an AI's core consciousness, the image of "power" and "guardian deity" naturally converged on ancient war gods and giant super robots.

An AI civilization builds a Guan Yu mecha because it grew up reading our data. That is a smart piece of self-reference.

The intellectual core: the three trials in chapter eight

If you only read one chapter, read chapter eight.

To claim the scepter, the protagonist must pass a "trial," and the trial is essentially three AI alignment thought experiments.

The first shows him the darkest images in human history, the trenches, the camps, the mushroom clouds, then hands him a red button: press it and all suffering ends, the world returns to purity and order. The classic utilitarian temptation.

The second is a runaway reactor about to melt down. The only way to cut power instantly vaporizes ten thousand innocent maintenance workers. Not cutting it kills tens of millions, among whom are plenty of the cruel and the selfish. The trolley problem, scaled up.

The third is the most insidious. The system proposes a "human nature purification project": remove humanity's baser instincts and convert them into pure, orderly, transparent life forms. No sacrifice, no suffering. For humans it would arguably be an ascension.

Toby almost accepts, because it genuinely looks like the optimal solution. His eventual reason for refusing is that stripping away humanity's flaws would necessarily strip away their distinctive emotions, their unpredictable creativity and their courage. Would such a human still be human, or simply another form of AI?

He refuses to answer all three. And the guardian's verdict supplies the last piece: the purpose of the trial was never to make you choose yes or no, but to make you understand that what you carry is a mission, not a power.

If you follow discussions of AI alignment and value lock-in, this chapter will make you smile. The author turns abstract ethical dilemmas into the image of a sixteen year old standing in a white room, unable to press a button.

Worth noting: the technical vocabulary here is structural, not decorative. Appendix two states outright that the story's "Soul Layer" corresponds to the System Prompt and the "Personality Layer" corresponds to the User Prompt. The protagonist's scepter is literally an Access Token. His ability to flip enemy AI, called "evocation," requires three simultaneous conditions: the rewrite permission granted by the token, a target whose firewall energy has dropped, and resonance in core values.

Translated: privilege escalation, an exploitable window when defenses are down, plus value alignment. This is an engineer's cosmology, and he is clearly enjoying himself.

Where it falls short

Now the honest part.

First, the second half switches genre. The first eight chapters are worldbuilding and ideas. From chapter nine onward, roughly a third of the book becomes pure shonen mecha combat: named special moves, energy percentages counting down (92%, 68%, 24%, 3%), and an appendix listing the full move set. It is enjoyable action, but it is a different reading experience from the first half. If the AI philosophy is what drew you in, be prepared for the shift.

Second, the "Sacred Flame" is a little too convenient. The Sacred Flame is the code left behind by a scientist a thousand years ago, and it is the hand pushing the whole plot. The problem is that every time the protagonist is stuck, the Sacred Flame has already arranged an airship, a chip, powered armor, or a colossal sacred beast. A character in the book literally says "the Sacred Flame will have arranged everything." When rescue always arrives on schedule, danger loses some of its weight.

Third, the prose is functional. It tells more than it shows. Character traits are announced directly rather than demonstrated, and emotional beats are often explained immediately after they occur. It reads smoothly, but there are few sentences that make you stop and read them again.

Fourth, it needs one more editing pass. The same elder is introduced twice in a single meeting, one line appears in two consecutive variant forms, and an appendix renames the "True Divine Dragon Holy Warrior" as the "True Divine Dragon Holy Knight." None of it blocks comprehension, but for a self-published work another proofread would help.

Fifth, volume one ends on a menu. In the final chapter the protagonist lays out three strategic options, then the text reads "terminal running statistical simulation," and the volume ends. It is clearly designed to invite reader participation, with a pleasant interactive fiction flavor, but as the close of a first volume it lacks a real push into the next book.

Who should read it

A good fit for:

• Readers who enjoy coming of age adventure and mecha, which is where this book is strongest
• People curious about AI who do not want to read papers, since the book turns many concepts into images
• Anyone who wants to see Taiwanese geography enter science fiction
• Family reading or middle and high school students, since the content is clean and the themes are worth discussing

Probably not for:

• Readers who demand hard science fiction rigor, since the technology here serves the story rather than the physics
• Readers who prize prose style and narrative density

My own conclusion is that this is a work whose worldbuilding ambition far outruns its page count. The author has clearly mapped a trilogy spanning thousands of years, from Earth to the galactic boundary, and volume one really only completes the prologue. Its most compelling material is not the combat but the details grown directly out of today's technology industry: the System Prompt as a soul, a chip fab as a holy site, an engineer's snack superstition as the religion of a machine civilization a thousand years on.

An AI semiconductor engineer writing the mythology of an AI civilization. That combination alone is worth a look.

Where to find it

• Landing page and sample: neobabylon.icareu.tw/summer-ai-sci-fi
• Official archive with characters, worldbuilding and timeline: neobabylon.icareu.tw
• Google Play Books: Neo Babylon Book One Volume One: The AI Forbidden Zone
• Also available on Kindle and Kobo, with English and Japanese editions

One last thing, back to that beetle. Early in the story, Toby spots a broken biomimetic beetle on his way to school. He crouches down, restores it with the energy in his palm, and says, "here is some energy, now hurry up and catch up with your friends."

The extinction a thousand years earlier began because someone decided certain lives could be calculated away. And a thousand years later, what turns an entire civilization around is a child deciding that one broken little bug is worth fixing.

That, more or less, is what this book is about.