How One Story Became a Film, a Webtoon, a Global E-book—and a Novel
I did not set out to make a film or write a novel. The original plan was much smaller: spend thirty days finding out what a non-developer could actually build with a $20 monthly AI subscription, then turn the experience into a book.
The project refused to stay inside that book. One of its lessons became a webtoon. An incident in the webtoon became a motion comic, then a three-minute, fifty-three-second short film. The same characters eventually moved into an eighteen-chapter workplace thriller.
From the outside, it may look as if I made one thing and reused it several times. The opposite was true. The moments that failed most often were the ones where I trusted an existing asset to work in a new medium without rebuilding it.
 |
| Episode 6 opens with a question rather than an explanation: if the accident never happened, what exactly did Dohyun prevent? |
Reusing the source did not preserve the feeling
A webtoon panel could hold a moment that vanished too quickly on video. A warning tone and a countdown that created tension in a film felt like exposition when written as prose. The unease and betrayal a novel could sustain for several pages would overwhelm a one-minute motion comic.
At first, reuse seemed like the sensible approach. Characters, setting, and the cause of an incident could travel across formats. But changing the aspect ratio or translating dialogue did not create a new work.
A webtoon needed a question that made the reader scroll. A motion comic needed an image that could stop a viewer in the first two seconds. A film needed a visible reason for someone to change a decision. A novel needed room for the emotional cost of that decision.
That was when I stopped thinking of the project as multichannel publishing. It had become an IP experiment: not distributing one master file, but rebuilding the same central question in the language of each medium.
The webtoon became a film—and exposed the gaps between scenes
Learning to Take My Place follows Zero Kim, an AI employee, and Dohyun Kim, the man in his fifties assigned to train him. Dohyun has to teach a system that may eventually replace him. Zero is fast and precise, but he cannot account for a person whose existence never entered the data.
Their relationship became the short film The Last Learning. A company wants to restart stalled machinery. A worker may still be inside an access hatch. The fact that a machine can restart collides with the question of whether anyone should restart it.
The first assembly did not fail because individual shots were unusable. It failed in the spaces between them. Viewers who had read the webtoon understood why Dohyun distrusted Zero. A new viewer did not yet know why Zero would hesitate, or why Dohyun would risk his job to stop the restart. Stitching panels together, then adding voices and music, did not create a chain of cause and effect.
I rebuilt that chain around four events: Zero's arrival, Dohyun's training, the disclosure of a replacement ranking, and the machinery incident. The voice tracks were remade to fit the scenes. Music was pulled back when it began explaining the emotion instead of supporting it. Rather than forcing a 16:9 scene into a vertical frame, the festival version kept the 16:9 composition and was checked again for character and machinery placement.
One line still broke the logic. An early version paired a sensor-malfunction probability with the option to restart remotely. If the signal itself was unreliable, why would it justify a restart?
The final version separates the two facts. The screen warns that occupancy inside the hatch is unconfirmed. Zero says, “Restart success probability: 98.7%. Remote restart is available.” The number describes the machine. It says nothing about whether a person is still inside. Once those facts were separated, Dohyun's refusal—and what Zero learns from it—finally made sense.
 |
| A probability can describe whether the machine will run. It cannot decide whether running it is safe. The Last Learning has now been officially submitted to the 2026 Busan International AI Film Festival. The submission was confirmed on August 31, 2026. Submission does not imply selection, screening, or an award; the official selection announcement is expected in October 2026. The completed film is now available on YouTube: Watch The Last Learning. Episodes 5 and 6 improved when I stopped trying to move everythingMy early motion-comic workflow was close to a slideshow: put the completed panels in order, then add zooms and camera movement. The images moved, but the tension did not. When every shot received similar treatment, the most important moments felt ordinary. Episode 5 was rebuilt around the order of attention rather than the number of panels: the access hatch, the restart control, the countdown that would not stop, and Dohyun's hand. Dialogue, music, and shot duration changed with that sequence. Episode 6 did not replay the accident. It opened with the company classifying the life-saving decision as a production delay. The employee who prevented an accident and the employee blamed for the delay were the same person. That contradiction became the reason to keep watching. The Korean motion-comic release reached 1,120 views and nine likes within its first twenty-four hours, with an early improvement in retention. That is an observation, not a verdict. The sample is too young to credit a single cause. Still, the cleaner thumbnail, the large opening question, and the face-to-face conflict are worth testing again. Production quality remained a separate problem. In multilingual panels, a tablet appeared in front of fingers that should have covered it. An information board obscured both characters' faces. File-count and size checks passed; the images did not. On a platform where deleting a post also destroys its URL and performance history, even a correction has a cost. The next episode will therefore begin with the riskiest panels. Hands, faces, tablets, and information layers will be checked at full size and mobile size before the rest of the episode is produced. A valid file and a believable image are now two different approvals. The English e-book had to become a product, not merely a translationFinishing the English manuscript felt like the end of the job. It was not. The cover had to remain readable as a small Amazon thumbnail, so it was rebuilt at 1600×2560. Twelve figures were prepared specifically for the English edition. The EPUB had to reflow as readers changed the type size. Navigation, tables, fonts, and images were checked in Kindle Previewer. The description and Author Page were reviewed again on the live product page. Translating a manuscript and delivering a book that a reader can buy are different jobs. After that second job was finished, the English Kindle edition went live on Amazon.com on August 26, 2026. It is currently priced at USD 5.99.  The novel could not be thirty-eight webtoon panels written out in proseMy first outline followed the webtoon's events too closely. The plot moved quickly, but the characters kept returning to the same office conversations. There was not enough emotional pressure to make a reader wonder what the next chapter would cost them. I started again and treated it as a near-future workplace thriller. Employees initially welcome the convenience of a new AI colleague. The mood changes when Zero becomes too good at the job. Some accept every recommendation. Some dismiss him. Some criticize him in private while using him in public. Others try to earn a favorable evaluation from the AI itself. Zero moves through operations, customer support, sales, human resources, finance, training, quality, and audit. At each new desk, recognition for Zero raises the replacement risk for someone else. Trust earned in one chapter becomes fear in the next. Dohyun's resistance, Haewon's choices, and a colleague's betrayal come from livelihoods and relationships—not a catalogue of corporate procedures. Late in the eighteen-chapter draft, Zero finds an internal proposal for “ZERO-2” on a computer left logged in. The system evaluating human workers discovers that it, too, can be replaced. The story stops being a simple contest between people and AI.  This is a complete eighteen-chapter review draft, not a publication-ready manuscript. The commercial length and prose still require another stage of work. What survived was not one master file, but one questionThe “$20” in this experiment refers to the base ChatGPT Plus subscription. It does not include regional tax, outside tools, music, equipment, or human labour. The answer after thirty days was not that AI had made everything by itself. AI made it possible for one person to move among writing, images, video, localization, and quality checks. The work moved forward when a person noticed something wrong and stopped: a line that made no sense, a board covering a face, an outline that had become repetitive, or a result that was not ready to publish. The webtoon needed a question that led to the next panel. The motion comic needed a first-two-second visual. The film needed a reason for a choice. The novel needed the feeling left behind by that choice. I did not copy the source four times. I rewrote the question four times. The English e-book contains not only the finished work, but also the failed attempts, revision decisions, EPUB and image QA, and the choices made after publication. The next experiments: comics and original musicThis IP experiment is not ending here. The existing free webtoon is a candidate for a paid comics edition, but not as a simple bundle of published panels. Resolution, facial acting, page turns, scene transitions, and reading rhythm will need to be designed again for a book. Music is moving in the same direction. Short logo themes and scene cues can become complete original songs, then a playlist whose tracks share the same creative world. The question is whether music that began behind a scene can become something worth returning to on its own. Neither the comics edition nor the music project is a finished product. As with the film, e-book, and novel, the next step is to build something small, release it carefully, and let real audience response decide what deserves to grow. Available now: What Can One Person Build with AI in 30 Days for $20? — Kindle edition |
Comments
Post a Comment