Chapter 52: Two Lines
Hana told him on a Saturday morning in May 2012, sitting on the plastic stool behind Park’s Banchan, between a container of japchae and a container of kkakdugi, while his mother was negotiating sesame oil prices three stalls over.
“I’m pregnant.”
Dojun’s hand, which had been reaching for a piece of kongnamul, froze in midair.
“Pregnant,” he repeated.
“Two lines on the test. I checked three times. Two lines, every time.” She was holding his other hand under the counter, out of sight from the passing customers, her grip tight enough to leave marks. “I found out yesterday. I wanted to tell you here. In the market. Because this is where the important things happen.”
The market hummed around them—vendors calling, customers haggling, the ancient soundtrack of commerce that had been playing in this exact alley for centuries. And in the middle of it, two people in their mid-twenties sat behind a banchan counter and processed the fact that a third person was on the way.
“How do you feel?” Dojun asked, because his own feelings were so tangled—joy, terror, wonder, the vertigo of another timeline divergence—that he needed her experience to anchor him.
“Terrified. Happy. Nauseous—literally, the morning sickness started yesterday.” She smiled through the fear. “I’ve been designing interfaces for six years. I thought I was prepared for any user experience. Turns out, pregnancy is the one UX challenge that no amount of wireframing can prepare you for.”
“We’re having a baby.”
“We’re having a baby. While running a hundred-person company. In four countries. During a competitive market war with TaskFlow.” She laughed—the bright, startled sound that he loved. “Our timing is impeccable.”
“Our timing has always been impeccable.”
“That’s your time-traveler confidence talking.”
“That’s my husband confidence talking. We’ll figure it out.”
His mother returned from her sesame oil negotiation, carrying two bottles with the triumphant air of a woman who had secured a volume discount.
“Five hundred won cheaper per bottle,” she announced, setting them on the counter. Then she looked at Dojun. Then at Hana. Then at their hands, hidden under the counter. Then at Hana’s face, which was doing a poor job of hiding something enormous.
“Aigoo,” she said. Very softly. “Aigoo, aigoo, aigoo.”
“We haven’t said anything yet,” Dojun said.
“You don’t need to. I can see it. A mother sees these things.” Her eyes were already wet. “How far?”
“Six weeks,” Hana said.
“Six weeks.” Younghee set down the sesame oil, came around the counter, and took Hana’s face in both hands. “My daughter. My daughter is going to be a mother.”
“Eomeonim—”
“Shh. Let me be happy. I’ve been waiting for this since the day you ate my japchae and I knew you were family.” She pulled Hana into a hug—fierce, complete, the kind of hug that communicated decades of hope in a single gesture. “A grandchild. Park Younghee is getting a grandchild.”
“Don’t tell anyone yet,” Dojun said. “It’s early—”
“I’m telling Mrs. Kang. And Mrs. Hwang. And Mr. Song. They’re family too.”
“Mom, those are your customers.”
“Customers who’ve been buying my banchan for thirty years are family. Family gets the news first.” She was already moving toward Mrs. Kang’s stall, a woman on a mission that no amount of caution could redirect.
Hana watched her go and laughed. “So much for keeping it quiet.”
“In Namdaemun Market, there’s no such thing as quiet. Information travels at the speed of banchan.”
“Your mother’s network is faster than our enterprise deployment pipeline.”
“Her latency is lower too.”
The pregnancy changed the calculus of everything—not in dramatic ways, but in the subtle recalibration of priorities that happens when the future suddenly contains a person who doesn’t exist yet but who will depend on every decision you make.
Hana continued working—”I’m pregnant, not incapacitated. Pregnant women have been designing things since the invention of design.”—but she adjusted her schedule. No more all-nighters. No more weekend sprints. The rule she had imposed on Dojun years ago—”Saturdays are non-negotiable”—now applied to her own body: rest was not optional.
“The baby is my most important design project,” she told the team during a standup, her matter-of-fact tone cutting through the awkwardness that the announcement had created. “I’m designing a human being. That requires adequate sleep, nutrition, and the absence of staring at screens for sixteen hours a day. Which, frankly, should have been company policy all along.”
“Should we update the employee handbook?” Minjae asked.
“Update it to say: ‘Aria employees are humans, not machines. Humans require rest, food, and the occasional experience of sunlight.’ If that’s not already in the handbook, we failed.”
The handbook was updated. Not as a joke—as a genuine policy shift that Hana championed and Dojun supported. Aria implemented mandatory rest days, flexible hours for parents, and a “no-meeting Friday” policy that gave engineers four consecutive hours of uninterrupted focus time.
“These aren’t perks,” Hana told the leadership team. “They’re design decisions. A rested team produces better work. Better work produces better products. Better products produce happier users. The pregnancy is just the catalyst—the truth was always there.”
But the pregnancy also coincided with a different kind of reckoning—one that came not from inside the company but from outside.
In June 2012, a Korean technology journalist named Shin Jaemin published an investigative piece in the Korea Herald: “Aria’s Invisible Eyes: How Much Does Your Productivity App Know About You?”
The article was thorough, well-researched, and devastating in its implications. Shin had obtained—through a former Aria employee who had left the previous month—internal documentation showing the scope of data that Aria’s learning module collected: email subject lines, calendar event titles, file names, browsing habits, location data from mobile devices, and behavioral patterns including typing speed, app switching frequency, and even the time of day users were most productive.
“All of this data is used to improve Aria’s task detection,” the article quoted the former employee. “But the question is: who else has access to it? The company’s privacy policy is vague. The data retention period is undefined. And the users—the 620,000 people who trust Aria with their digital lives—have no clear way to see what Aria knows about them or to delete it.”
Dojun read the article at 6 AM on a Saturday—the day before his mother’s market visit—and felt the particular cold that comes from recognizing a crisis before it becomes one.
In his previous life, Prometheus Labs had faced a similar reckoning—not in 2012 but in 2025, when AI ethics became a global conversation and the company’s data practices came under scrutiny. The response had been defensive, legal, corporate—and it had cost them years of goodwill.
He would not repeat that mistake.
“Emergency leadership call,” he texted the inner circle. “8 AM. Not about the pregnancy.”
Hana called him at 6:15. “I read it. The article is accurate. Everything Shin describes is data we actually collect.”
“I know.”
“We designed the privacy policy when we were twelve people in a closet. It hasn’t been updated since. The data retention policy is—” She paused. “We don’t have a data retention policy.”
“I know that too.”
“Then this isn’t a PR crisis. It’s a product crisis. We built invisible technology that’s too invisible—users don’t know what it’s doing with their data because we never designed a way for them to see.” Her voice hardened—the particular hardness of a designer confronting a design failure. “This is my fault. I designed the interface to hide complexity. But hiding complexity and hiding data collection are two different things, and I didn’t draw the line clearly enough.”
“It’s not your fault. It’s our collective oversight. And we fix it the same way we fix everything—transparently, immediately, and thoroughly.”
The leadership call lasted two hours. By the end, they had a plan:
First: a public response. Not a corporate press release—a personal letter from Dojun and Hana, published on Aria’s blog, acknowledging the article’s findings and accepting responsibility. “The article is accurate. We collect extensive data to improve Aria’s task detection. We should have been more transparent about what we collect, why we collect it, and how long we keep it. We failed in that transparency. We’re fixing it.”
Second: a product update. A “Data Dashboard” that showed every user exactly what Aria knew about them—every email subject, every calendar title, every behavioral metric. Users could see, export, or delete their data with one click. Taeyoung estimated three weeks to build it. Hana said two. It shipped in twelve days.
Third: a formal privacy policy overhaul. Data retention limited to twelve months. Automatic deletion of behavioral metrics after six months. Opt-in, not opt-out, for every data category. And a dedicated privacy page, designed by Hana, that explained Aria’s data practices in language that a Namdaemun Market banchan vendor could understand.
“If my mother can understand the privacy policy, it’s good enough,” Dojun told the team. “If she can’t, simplify it.”
Younghee was, in fact, consulted. She read the draft privacy page—written in plain Korean, with no legal jargon—and said: “I understand everything except the word ‘metadata.’ What is metadata?”
“Data about data. Like—the subject of an email, not the email itself.”
“So if I write ‘Order 50 containers of kimchi’ in an email subject, Aria knows I ordered kimchi, but not from whom?”
“Exactly.”
“That seems fine. Knowing I ordered kimchi is not a secret. Everyone in the market knows I order kimchi.” She paused. “But what if the subject says ‘Doctor appointment, hip check-up’? That’s private.”
“Users can mark certain tasks as private. Aria won’t collect metadata from private tasks.”
“Good. Put that in the policy. In big letters. People worry about doctors more than kimchi.”
“Noted.”
The privacy update launched in July. The response was overwhelmingly positive—not because users were delighted by the data dashboard (most never used it), but because the transparency itself built trust. The Korea Herald’s follow-up article was titled: “Aria Responds to Privacy Concerns with Industry-Leading Transparency. Other Tech Companies, Take Note.”
Shin Jaemin, the journalist who had written the original exposé, emailed Dojun directly: “I’ve covered Korean tech for ten years. This is the fastest, most comprehensive privacy response I’ve ever seen from any company. Most companies lawyer up. You designed a dashboard. That says something about who you are.”
Dojun forwarded the email to the team with one line: This is what ‘fail well, fix fast’ looks like.
The privacy crisis resolved, the company growing, and the baby growing too—Hana was four months along by July, the morning sickness fading into a general state of fatigue that she managed with the same disciplined efficiency she applied to sprint planning.
“I’m tracking my energy levels on a spreadsheet,” she told Dojun one evening. “Peak productivity: 9-11 AM. Acceptable productivity: 2-4 PM. Complete uselessness: 7 PM onward. I’ve adjusted my design review schedule accordingly.”
“You’re project-managing your pregnancy.”
“I’m a designer. I design experiences. Pregnancy is an experience. It gets designed.” She rubbed her belly—a gesture that was becoming more frequent and more tender as the weeks passed. “Also, the baby kicks during design reviews. I choose to interpret this as enthusiastic approval of my interface decisions.”
“Or protest.”
“Approval. I’m the designer. I decide the interpretation.”
Dojun watched her—his wife, his co-founder, the CDO of a hundred-person company, carrying a child while shipping products and writing privacy policies and visiting his mother every other Saturday for japchae and unsolicited advice.
In his previous life, there had been no child. No marriage. No Saturday visits. Just code and conferences and the slow, corrosive solitude of a man who had chosen machines over people.
This life had code and conferences too. But it also had a woman who tracked her pregnancy on a spreadsheet and interpreted fetal kicks as design feedback. It had a mother who had already knitted three baby blankets and was working on a fourth “because babies need options.” It had a friend in Daejeon who had sent a onesie printed with the Nova Systems logo and a note: “Start them on cloud infrastructure early.”
Two lines on a pregnancy test. Two lines that meant the web was growing again—one more thread, one more connection, one more person who would be loved by the people who had learned, through two lifetimes of trying, that love was the only code that mattered.
The japchae goes in the front. The baby kicks during design reviews. And the future, arriving one heartbeat at a time, was more unpredictable and more precious than any algorithm could calculate.