Chapter 23: January
The new year arrived with snow and a phone call from Seokho at 12:03 AM.
“Happy new year,” Seokho said. “I’m calling because I just finished the preliminary results for our collaboration paper and I wanted someone to appreciate them. My lab mates are asleep. You’re the only person I know who’s awake at midnight on New Year’s because they were coding.”
“I was watching the countdown on TV.”
“Were you coding while watching the countdown?”
“…The TV was on in the background.”
“That’s what I thought. Listen—the bimodal model works for network traffic. Not just works—it’s elegant. The predictable flows map almost perfectly to the regular branch behavior, and the unpredictable flows show the same random distribution. The dual-mode classification improves traffic routing efficiency by 23% in our simulation.”
“Twenty-three percent is significant.”
“It’s publishable. Easily. SIGCOMM will take this. I’m formatting the submission now—target deadline is April 15th. Can you review the prediction model adaptation section by January 15th?”
“Done. I’ll have it to you by the 10th.”
“Five days early. Annoyingly efficient, as always.” A pause. “Happy 2007, Park. It’s going to be an interesting year.”
“More interesting than you think.”
“Cryptic. I’m adding ‘cryptic’ to my growing list of Park Dojun personality traits, right between ‘inexplicable’ and ‘frustratingly talented.'”
They hung up. Dojun sat in his apartment, the TV showing New Year’s celebrations at Bosingak Bell—the ancient bell in Jongno that had rung in Korean new years for six hundred years. Snow fell over Seoul like static on a screen, soft and relentless.
2007. The year the iPhone would be announced. The year that would change everything about how humans interacted with technology, with each other, with the world. And he was one of a handful of people on the planet who knew it was coming.
The temptation to act on that knowledge was strong. He could build a mobile-first version of Bridge before anyone else. He could invest in Apple stock. He could position himself at the exact intersection of mobile computing and productivity software, years before the market existed.
But the lesson of his first life was that knowing the future was not the same as controlling it. Prometheus Labs had been built on future knowledge, and it had consumed everything he loved. Bridge was being built differently—slowly, organically, on a foundation of real user needs and real relationships.
He would not sacrifice that foundation for the sake of being first.
The iPhone comes in June, he thought. Bridge will be ready for it when the time is right. Not before.
He turned off the TV, went to bed, and let 2007 arrive without him.
January was a month of quiet building.
Bridge hired its first employee: Cho Hyunwoo, a third-year CS student who had been one of the beta’s most active bug reporters and who, during his interview, had said: “I use Bridge every day and I have seventeen ideas for how to make it better. Here’s a prioritized list.” Hana had hired him on the spot.
“He brought a prioritized list to a job interview,” she told Dojun afterward. “That’s exactly the kind of person we need—someone who cares about the product enough to already be thinking about it.”
Hyunwoo’s first task was the sync layer rewrite—the infrastructure work needed to support expansion beyond SNU. He was competent, methodical, and had the particular gift of writing code that was readable by humans, not just compilers. Dojun supervised his work with a light touch, correcting architecture decisions when needed but letting Hyunwoo find his own solutions.
“You’re a good mentor,” Hana observed one afternoon, watching Dojun review Hyunwoo’s code. “Patient. Clear. You explain the ‘why,’ not just the ‘how.'”
“I had a good teacher. A long time ago.”
“The mysterious mentor again.”
“The mysterious mentor.” He smiled. “Some lessons take a long time to learn.”
The campus beta continued growing. With the learning module active, user satisfaction had plateaued at 4.6 out of 5—a number that Minjae called “unreasonably high” and that Choi Eunji, in her monthly check-in, called “exactly where it should be for a product with real product-market fit.”
“Product-market fit,” she explained during a coffee meeting in Daechi-dong, “is when your users can’t imagine going back to life without your product. Your retention data says 67% of daily active users have been active every day for more than 30 days. That’s not curiosity. That’s dependency. In a good way.”
“What’s the next milestone?” Hana asked.
“Expansion. Korea University and Yonsei by March. Three campuses, target 5,000 users by summer. After that—” She sipped her coffee. “After that, we talk about Series A.”
“Series A?”
“A larger funding round. Ten to twenty times the seed amount. Enough to hire a real team, build real infrastructure, and launch Bridge as a commercial product.” She looked at them with the evaluative precision that Dojun recognized from a hundred investor meetings in his previous life. “But Series A requires traction, team, and timing. You have the first two. The timing—” She paused. “I think 2007 is going to be a very good year for software products. I can’t say why yet. But I have a feeling.”
Dojun said nothing. He knew exactly why 2007 would be good for software. In five months, Steve Jobs would announce the iPhone, and the entire world would suddenly need software that worked on small screens, with touch interfaces, in mobile contexts. Bridge—a task-centric workspace that adapted to user context—was almost perfectly positioned for that world.
But he couldn’t say that. Not to Eunji, not to anyone. The iPhone announcement was the best-kept secret in technology, and revealing knowledge of it would raise questions he couldn’t answer.
“I have a similar feeling,” he said. “2007 is going to change things.”
Eunji studied him for a beat longer than usual, then nodded. “Good. Be ready.”
The Seokho collaboration progressed with the brisk efficiency of two people who understood each other’s work without needing to explain it. Biweekly meetings alternated between Seoul and Daejeon—Dojun bringing Auntie Bong’s naengmyeon shop into the rotation (“You can’t keep coming to Daejeon without eating here. Auntie Bong will think you’re avoiding her.”) and Seokho reluctantly visiting the jjigae place (“The soup is good. The seating is inhumane. But I’ll endure it for science.”).
The paper took shape: “Dual-Mode Traffic Classification for Software-Defined Networks: Applying Branch Prediction Principles to Network Flow Management.” It was dense, technical, and—Seokho claimed—”the most naturally interdisciplinary paper either of us will ever write.”
“Your prediction model adapts to network traffic like it was designed for it,” Seokho said during a January meeting in Dojun’s apartment, surrounded by printouts and empty coffee cups. “The bimodal distribution holds across every dataset we’ve tested. TCP flows behave like regular branches—predictable, steady, optimizable. UDP bursts behave like irregular branches—random, volatile, best handled by conserving resources rather than predicting.”
“The model generalizes because the underlying principle generalizes,” Dojun said. “Any system with bimodal behavior benefits from dual-mode classification. Processor branches, network traffic, even human workflows.”
“Human workflows?” Seokho raised an eyebrow. “That sounds like Bridge.”
“Bridge’s task detection is fundamentally the same principle. Predictable tasks—recurring meetings, regular assignments—are easy to classify. Unpredictable tasks—one-off requests, ad hoc collaborations—require a different approach.”
“You’re telling me that your consumer product, your academic paper, and our collaboration paper are all based on the same insight.”
“Different applications. Same mathematics.”
Seokho leaned back. “That’s either very elegant or very lazy.”
“It’s efficient. Good ideas should be reused.”
“Spoken like a true engineer.” He gathered his printouts. “The paper is in good shape. I’ll handle the experimental results section. You finalize the theoretical framework. Submission target: April 1st—two weeks before the deadline, because I don’t submit on deadlines. Deadlines are for people who didn’t plan ahead.”
“April 1st. Done.”
“And Park? The ISCA presentation in June. Have you started preparing?”
“I’ve outlined the talk.”
“Outline is not preparation. ISCA Q&A is brutal. Professors who’ve been building processors for thirty years will ask you questions designed to find the limits of your understanding. You need to prepare for questions you can’t answer.”
“I can answer anything about my own paper.”
“That’s the problem. You answer too well. Too completely. Too confidently.” He paused. “Learn to say ‘I don’t know’ convincingly. It’s the most important phrase in academic defense. Nobody believes a twenty-year-old who has all the answers.”
The advice was sharp and true. Dojun filed it alongside Kim Taesik’s “talent without context” and Hana’s “don’t edit yourself”—the collection of wisdoms from people who saw him clearly enough to help him hide in plain sight.
“I’ll practice saying ‘I don’t know,'” he said.
“Practice in front of a mirror. Make it look natural. Right now, your ‘I don’t know’ face looks like someone pretending not to know the answer on a game show.”
“That’s because—”
“Because you always know the answer. Yes. I’ve noticed. Everyone has noticed. That’s the problem.” Seokho stood and pulled on his coat. “January train back to Daejeon. Call me when the framework section is done.”
He left. Dojun sat among the printouts and thought about learning to say “I don’t know” while knowing more than almost anyone alive.
It was, he reflected, the central paradox of his second life: the more he achieved, the more he had to pretend. The more people trusted him, the more he hid. The more real his relationships became, the more the impossible secret at the center of everything pressed outward against the walls he had built around it.
One day, he thought. One day I’ll tell them. Hana first, then Kim Taesik, then Seokho. And they’ll either believe me or think I’m insane. And either way, the hiding will be over.
But not yet. Not until the web was strong enough to hold the weight of the truth.
Saturday. Namdaemun Market. Late January.
His mother was using Bridge Market Edition with the aggressive competence of a woman who had conquered a technology by treating it exactly like a stubborn supplier—with firm expectations, no patience for excuses, and the unshakeable belief that she could do it better if she had to.
“The doraji numbers are wrong,” she announced when Dojun arrived. “I sold seven containers yesterday but the system says I sold five.”
“Did you press the minus button seven times?”
“I pressed it five times. Then Mrs. Hwang’s grandson bumped the table and I lost count.”
“That’s not a bug, Mom. That’s a hardware collision with a small child.”
“Then build me a button that’s child-proof.” She said it with perfect seriousness. “Also, the accounting page doesn’t account for the containers I give away for free. Mrs. Choi’s daughter had a baby. I gave her three containers of seaweed soup banchan. The computer thinks I sold them.”
“I’ll add a ‘gift’ button.”
“Good. And a ‘samples’ button. I give samples to new customers. If the computer counts samples as lost inventory, the numbers will never match.” She frowned at the laptop. “Your computer is smart about some things and stupid about others. Like you.”
“Thanks, Mom.”
“I mean it with love.” She adjusted the laptop’s position on the counter—she had built a small shelf for it out of a wooden crate and two rubber bands, a solution that was structurally questionable but entirely functional. “Mrs. Kang next door wants one.”
“One what?”
“One of these.” She pointed at the laptop. “The program. She saw me using it and asked if she could have the same thing for her socks stall. I told her my son built it. She wants to know how much it costs.”
Dojun stared at his mother. “Mrs. Kang wants Bridge Market Edition for her socks stall?”
“Is that strange?”
“No. That’s… that’s exactly how products grow. Word of mouth. One user tells another. The product spreads because it works, not because of marketing.” He felt a familiar tingling—the particular electricity of product-market fit, felt not through data dashboards but through a sock vendor in Namdaemun Market.
“So can she have one?”
“I’ll set it up next Saturday. She’ll need a laptop.”
“She has one. Her grandson left it when he moved to Busan. She uses it as a paperweight.”
“A paperweight.”
“It’s very flat. Good for keeping receipts in place.” His mother shrugged. “The market adapts, Dojun-ah. Even to computers.”
He spent the rest of the morning helping his mother with the stall, adding the “gift” and “samples” buttons to Bridge Market Edition between customers, and marveling at the organic, unplanned way that technology was seeping into a space that had resisted it for decades.
His mother hadn’t needed Bridge. She had managed perfectly well without it for thirty years. But now that she had it, she was finding uses for it that he hadn’t imagined—tracking which banchan sold fastest on which days (“Japchae peaks on Saturdays. Kkakdugi peaks on Wednesdays. I knew this, but now the computer knows it too”), identifying which customers bought which combinations (“Mrs. Park always buys kimchi and kongnamul together. We should sell a set”), and even using the accounting function to negotiate with suppliers (“Mr. Song, the computer says my beef costs went up 8% this month. Your prices haven’t changed. Are your cuts getting smaller?”).
“Your mother,” Hana said when Dojun told her about the supplier negotiation, “is using Bridge as a competitive intelligence tool. She’s running her banchan stall like a Fortune 500 company.”
“She’s been running it like a Fortune 500 company for thirty years. She just didn’t have the vocabulary.”
“We need to talk to more market vendors. If Bridge Market Edition can help your mother’s stall, it can help thousands of small businesses that run on memory and habit. That’s a market segment nobody is serving.”
“One step at a time. Campus beta first. Market vendors later.”
“I know. I know. But Dojun—” She leaned forward, eyes bright with the particular light of someone seeing a future that others couldn’t. “This is what Bridge was always meant to be. Not just a student tool. Not just an enterprise product. A bridge between technology and the people who never thought technology was for them. Your mother. Mrs. Kang. Every ajumma in every market in Korea.”
“That’s a lot of ajummas.”
“That’s a lot of potential users.” She grinned. “And they all need good japchae.”
January ended with snow on the ground and numbers on the dashboard: 487 daily active users at SNU, a 92% task detection accuracy, four employees (including the three founders and Hyunwoo), fifty million won in the bank with a careful burn rate, one ISCA presentation to prepare for, one SIGCOMM paper in progress, and one banchan vendor in Namdaemun Market who had taught her neighboring sock vendor how to use a laptop that had been a paperweight.
The web held. The code ran. The future, approaching from all directions at once, felt less like a threat and more like a wave—powerful, inevitable, and possibly, if you positioned yourself right, something you could ride instead of being crushed by.
February would bring spring semester. March would bring the first anniversary of his second life. And June would bring San Diego, ISCA, and the world stage.
But first, Dojun had a “gift” button to build and a sock vendor’s laptop to set up.
One user at a time. One bridge at a time.