Chapter 12: Summer Code
The first line of Bridge’s source code was written at 11:47 PM on a Friday in May, in a studio apartment that smelled of instant ramyeon and sesame oil, on a Compaq Presario whose fan screamed like a dying animal whenever the CPU load exceeded 60%.
Dojun typed it with the careful deliberation of a man laying the foundation stone of a cathedral:
// Bridge v0.1 — Making technology invisible.
He stared at the comment line for a long time. In another life, the first line of Prometheus Labs’ codebase had been different—something functional, forgettable, a module header that served no purpose beyond organization. He hadn’t thought to mark the beginning. By the time Prometheus was worth marking, the beginning was buried under ten million lines of code and nobody remembered what came first.
This time, he wanted to remember.
He opened a new Java file—Java, because in 2006 it was the lingua franca of enterprise software and because the web frameworks he actually wanted to use (React, Node.js, Django) wouldn’t exist for years. The architecture he sketched was deliberately simple: a local desktop application that aggregated email, calendar, and file system data into a unified task view. No cloud. No network. Everything ran on the user’s machine.
It was primitive compared to what Prometheus Labs would eventually build, but it was the right starting point. In 2006, cloud infrastructure was barely a concept. AWS would launch its first services later this year, but nobody was building consumer products on it yet. The Bridge prototype needed to work with what existed: local storage, desktop APIs, and the IMAP email protocol that every university student used.
By 2 AM, he had the basic skeleton running—a Java Swing interface (ugly, but functional) that could read email headers from an IMAP connection and display them in a list. By 3 AM, he had added calendar parsing from iCal files. By 4 AM, he had a rudimentary task detection algorithm that matched email subjects to calendar events based on keyword overlap.
It was terrible. The matching was crude, the interface was hideous, and the performance was sluggish. But when he fed it a test dataset—a week’s worth of his own emails and calendar entries—it correctly identified three tasks and grouped the relevant emails and events together.
Prepare for Dr. Yoon’s seminar → [Email: Yoon’s reading list] + [Calendar: Seminar, Thursday 2PM]
KETI research meeting → [Email: Kim Taesik’s project update] + [Calendar: Lab hours, Wednesday 1PM]
Bridge project discussion → [Email: Hana’s wireframe attachment] + [Calendar: Jjigae place, Friday noon]
Three tasks, automatically detected. It was a toy. A proof of concept that a skilled undergraduate could build in a night. But it was also the first pixel in what would become a very large picture.
Dojun saved his work, closed the laptop, and fell asleep at his desk. He dreamed of code—not the clean, efficient code of his professional years, but the messy, hopeful code of a first draft, full of TODOs and FIXMEs and the particular optimism of someone who believes that every bug is just a solution waiting to be found.
The semester ended in June with the quiet anticlimax of final exams. Dojun aced everything—his biggest challenge was deliberately leaving a few minor errors on his exams to maintain the “gifted but human” persona that Kim Taesik’s accelerated track was designed to support.
His final grades: four A-pluses and one A. The lone A was in Korean History, where his essay on the Joseon dynasty’s bureaucratic structure had been marked down for “unconventional interpretations.” He suspected the professor simply didn’t appreciate a computer science student arguing that Joseon’s civil service examination system was an early form of algorithmic meritocracy.
The semester had also delivered a cascade of firsts. First publication—his A-star contextual pathfinding paper, accepted at the Korean Conference on Computing, co-authored with Dr. Yoon and listing Hana as “interface design consultant.” First research result—the KETI simulation model was producing clean data that Kim Taesik declared “promising,” which in Kim-speak translated to “exceptional.” First genuine friendships—Minjae, who had somehow become the social glue of their group; Jinwoo and Soyeon in the lab, who had graduated from tolerating him to actively seeking his input on their theses.
And Hana. Always Hana, in the spaces between everything else.
They had fallen into a rhythm over the semester—jjigae lunches on Thursdays, Bridge work sessions on Friday evenings, occasional coffee runs between classes where they talked about everything and nothing. She told him about her grandmother’s rice cake recipes, her father’s disappointment that she hadn’t chosen medicine, her secret love of terrible Korean soap operas. He told her about his mother’s banchan stall, his cousin Yuri’s improving math grades, his embarrassing inability to cook anything more complex than ramyeon.
He did not tell her about the forty years of future knowledge in his head, the company they would build together, or the way her laugh made his chest ache with the specific pain of loving someone you had already lost.
Some things, he was learning, needed to grow at their own pace.
The summer plan crystallized over a series of meetings in the jjigae place, which had become their unofficial headquarters.
“Three months until the Innovation Showcase,” Hana said, spreading a project timeline across the table between the banchan dishes. She had drawn it on a long scroll of paper, color-coded with the meticulous beauty of a medieval manuscript. “September 15th. Fifty teams. Three winners. We need a working demo, a pitch deck, and a business plan.”
“The demo is the priority,” Dojun said. “Nobody cares about a pitch deck if the product doesn’t work.”
“Spoken like a true engineer. The judges care about the pitch deck, Dojun. They’re industry people—VCs, executives, consultants. They’ve seen a thousand demos. What they haven’t seen is a compelling story about why this product needs to exist.”
“The product tells its own story.”
“The product tells a technical story. I need to tell a human story. Why does a university student need Bridge? What problem does it solve that they can feel in their daily life?” She tapped the timeline. “You handle the backend. I handle the frontend and the narrative. Minjae handles data integration and testing.”
“Minjae agreed?”
“Minjae agreed the moment I described the project. His exact words were, ‘You had me at working prototype. I’ve been waiting for someone to ask me to build something real instead of another homework assignment.'”
Dojun smiled. Minjae’s enthusiasm was one of the unplanned gifts of this timeline. In his first life, he hadn’t had someone like Minjae—a competent, eager team member who was genuinely happy to contribute without needing to lead. He had surrounded himself with stars, and stars competed for light. Minjae was something rarer: a team player who made the team better by being reliable.
“What about the summer?” Dojun asked. “Are you staying in Seoul?”
“My parents want me to come home to Busan. I told them I have a summer research project.” She grinned. “Which is technically true.”
“I’m staying. The lab is open all summer, and Kim Taesik wants the KETI results by August.”
“So you’ll be coding Bridge and doing KETI research simultaneously.”
“And tutoring Yuri. And visiting Mom on Saturdays.”
“Dojun.” Hana set down her pen. “That’s insane. You can’t work on three projects, tutor a middle schooler, and visit your mother every weekend for three months straight. When will you sleep?”
“I’ll sleep in September.”
“That’s not funny.”
“It’s a little funny.”
“It’s not. I’ve seen what happens to CS students who don’t sleep. They start talking to their code. Then their code starts talking back. Then they wake up on the floor of the computer lab with keyboard imprints on their face.” She pointed her pen at him. “I need you functional in September. Not brilliant—functional. A broken partner is worse than no partner.”
The word partner landed with unexpected weight. She used it casually—project partner, team partner—but every time Dojun heard it, he heard the other meaning. The one that had ended in silence and a set of keys left on a desk.
“I’ll be functional,” he said. “I promise.”
“I’m going to hold you to that. I’m writing it in the project timeline. ‘Dojun: remain functional.’ Right here, in red.”
She wrote it. He watched the red ink curve across the scroll paper and felt a warmth that had nothing to do with the jjigae.
Summer in Seoul was a different kind of beautiful—humid, heavy, the kind of heat that turned the city into a steam room and the campus into a green cathedral of shade-seeking students. The cherry blossoms were long gone, replaced by thick canopies of maple and ginkgo that turned every path into a dappled tunnel of light.
Dojun settled into a routine. Mornings: run, then KETI research in the lab. Afternoons: Bridge development in the computer lab or his apartment. Evenings: reading, planning, the occasional phone call with Seokho, who was spending his summer at a KAIST research fellowship and texted Dojun algorithm puzzles at 11 PM “because I’m bored and you’re the only person who solves them before midnight.”
Saturdays: Namdaemun Market. Always.
His mother had noticed the changes in him—how could she not? Her son, who had visited twice in his entire first year of college, now appeared every Saturday with the reliability of a train schedule. He brought her small gifts—a new cutting board when hers cracked, a portable fan for the summer heat, a pair of sneakers to replace the ones held together by hope and duct tape.
“Where is this money coming from?” she demanded one Saturday, examining the sneakers with the suspicious scrutiny of a woman who knew exactly how much a college student’s budget allowed.
“The contest prize. And the tutoring.”
“The tutoring is twenty thousand won a session. These sneakers cost forty thousand.”
“I got them on sale.”
“You got them at Namdaemun Market. I can tell by the stitching.” She held them up. “Mrs. Kang’s stall on the third floor. These are thirty-five thousand, not forty. You’re lying about the price to make me feel less guilty about accepting them.”
“…Yes.”
“I raised a good son.” She put the sneakers on immediately, with the practical speed of a woman who didn’t waste time on ceremony. “They fit well. Mrs. Kang knows feet.”
While she worked, Dojun helped at the stall—carrying boxes, restocking containers, learning the rhythm of the market. He was getting better at it. His mother still scolded him regularly (“The japchae goes in the front, the kimchi in the back—customers eat with their eyes first”), but the scolding had taken on a fond, rehearsed quality, like a comedian running through a beloved routine.
“I entered a competition,” he told her one Saturday, between customers. “A startup showcase at the university.”
“A startup? Like a business?”
“Sort of. A software product. Something I’m building with my project partner.”
“The girl? The designer?”
“Her name is Hana.”
“Hana.” His mother tested the name like a customer sampling banchan. “Pretty name. Is she pretty?”
“Mom.”
“It’s a simple question.”
“She’s my project partner.”
“And? Project partners can be pretty. Your father was my business partner before he was your father.” A shadow crossed her face, quick and involuntary. “Well. That’s a different story. Tell me about this competition.”
“Fifty teams. The top three get funding—money to start a real business. We’re building a tool that helps people organize their digital work. Email, calendars, files—all in one place.”
“Like a desk organizer? For computers?”
“Exactly like a desk organizer for computers.”
“Can you make one for my stall? I have papers everywhere—supplier receipts, orders, that stupid tax form the city keeps sending.”
“Actually… yes. That’s exactly what it’s for.”
His mother looked at him—really looked, the way she did when she was seeing something new in a face she’d known for twenty years. “You’re different this semester, Dojun-ah. Not just the visits and the sneakers. You’re… lighter. Like something heavy was taken off your shoulders.”
Something was, he thought. Forty years of regret.
“I’m just happy, Mom.”
“Happy.” She said it like it was a foreign word—not unfamiliar, but rarely used. “Good. Stay happy. And bring Hana to the market sometime. I want to meet her.”
“That’s not—”
“I want to meet your project partner.” The emphasis was unmistakable. “A mother has a right to meet the people her son spends time with. Bring her. I’ll make her my best japchae.”
“Your japchae is always the best.”
“Flattery works on me. I raised you well.” She turned to a customer. “Ajumma! The kkakdugi today is special—I used the radish from Chungcheong province. Try a sample!”
Dojun watched her work and felt the particular ache of loving someone so completely that it hurt—not the sharp pain of loss, which he had known in his previous life, but the soft, deep ache of presence. Of being here, now, in a place and time where his mother was alive and strong and selling banchan and nagging him about girls and sneakers and eating properly.
He would not waste this. Not a single Saturday. Not a single bowl of kongnamul-guk. Not a single moment of her voice saying his name.
By mid-July, the Bridge prototype was taking shape.
Hana had delivered a redesigned frontend—no longer the ugly Java Swing interface Dojun had built, but a clean, web-based UI using HTML, CSS, and the nascent AJAX techniques that were just beginning to transform web development. She had taught herself JavaScript in three weeks specifically for this project, a fact that she mentioned with the casual understatement of someone who didn’t realize how impressive it was.
“The interface is task-centric,” she explained during a video call—well, a phone call with screen-sharing through a clunky remote desktop application, because video calling in 2006 was still a novelty. “Each task is a card. You drag cards to rearrange priority. Click a card to see all related items—emails, calendar events, files. The system learns your patterns over time.”
“The learning algorithm is ready,” Dojun said. “It’s basic—keyword frequency analysis and temporal correlation. Not AI. But it’s effective for common workflows.”
“Show me.”
He ran the demo. On screen, the Bridge interface populated with sample tasks. A card labeled “KETI Research” automatically grouped three emails from Kim Taesik, two lab reports, and a calendar entry for Wednesday lab hours. Another card labeled “Bridge Development” grouped emails from Hana and Minjae, code commit logs, and project milestone dates.
“It works,” Hana breathed. “It actually works. The grouping is right. The priority ordering makes sense. And the interface—Dojun, it’s beautiful.”
“You designed the interface.”
“I designed the layout. You made it move. That’s different.” She was quiet for a moment. “This is real, isn’t it? Not a class project. Not a prototype. This is a real product that real people could use.”
“It’s getting there.”
“It’s there.” Her voice had a quality he recognized—the particular tension of someone standing at the edge of something much bigger than they expected. “Dojun. What if we actually win the showcase?”
“Then we get five million won and six months of office space.”
“And then what?”
“And then we build version two. And version three. And we keep going until Bridge is what we both know it can be.”
“A company.”
“Maybe. Eventually. If the product deserves it.”
“The product deserves it.” She said it with a quiet certainty that reminded him, with a sharp pang, of the Hana who had stood beside him at Prometheus Labs’ founding. The same conviction. The same clarity. The same absolute, terrifying belief that what they were building mattered.
“We have two months until the showcase,” she said. “Let’s make sure the demo is perfect. Not good—perfect. I want the judges to see Bridge and wonder why nobody built this before.”
“That’s a high bar.”
“I told you. I like high bars.”
She hung up. Dojun sat in his apartment, the Bridge prototype glowing on his Compaq Presario’s screen, and thought about what came next.
Two months of building. Then the showcase. Then—if everything went right—the beginning of something that could change the world.
Or the beginning of the same mistake, dressed in different clothes.
He opened his journal—the physical notebook, not a file—and wrote:
The product is on track. Hana is brilliant. The team works. Everything is going right.
That’s what scares me. In my previous life, everything went right too—until it didn’t. The company grew. The partnership cracked. The people who mattered most became casualties of the mission.
I can’t let that happen again. I need to build slower. Listen more. Choose people over product when the two conflict.
Bridge is not Prometheus Labs. Not yet. Maybe not ever. But if it becomes something—if we build it into something real—I need to remember: the code is the means. The people are the end.
Mom. Hana. Seokho. Minjae. Kim Taesik. These are the lines I will not delete.
He closed the notebook, turned back to the screen, and wrote the next line of code.
Outside, Seoul hummed through the July night—hot, alive, indifferent to the quiet revolutions happening in studio apartments across the city. Somewhere, someone was building the next social network. Somewhere else, someone was writing the algorithm that would change mobile computing. And in a cramped room in Gwanak-gu, a twenty-year-old with the memories of a lifetime was writing a program that understood what people needed before they knew they needed it.
The cursor blinked. The code grew. And summer stretched ahead like a runway—long, straight, and leading somewhere Dojun couldn’t quite see yet, even with forty years of hindsight.
For the first time, that uncertainty didn’t feel like a threat.
It felt like freedom.