Chapter 27: Growing Pains
The first new hire argued with Dojun about code architecture within forty-eight hours of starting.
His name was Yoon Taeyoung, a fourth-year CS student from Korea University who had been recruited through the campus beta program. He was twenty-three, wore rectangular glasses that made him look like a manga character, and had the particular confidence of someone who had been the best programmer in every room he’d entered since middle school.
Until now.
“Your event loop is blocking,” Taeyoung said, pointing at Dojun’s mobile prototype code on the shared screen. “You’re processing task detection synchronously in the main thread. On a phone with limited memory and a single-core processor, this will freeze the UI for up to three seconds every time the user opens the app.”
“I know,” Dojun said. “It’s a prototype. The async refactor is planned for sprint three.”
“Sprint three is six weeks away. If we’re targeting App Store launch in July 2008, we can’t have a blocking event loop in the prototype for six weeks. Every design decision Hana makes will be based on a broken interaction model. She’ll design for a responsive app and we’ll deliver a frozen one.”
The office went quiet. Hana looked up from her sketchbook. Minjae stopped typing. Hyunwoo pretended to be very interested in his monitor.
Dojun studied Taeyoung. The critique was correct—technically, architecturally, and from a product perspective. The blocking event loop was a shortcut Dojun had taken because he was building the prototype alone, in a language he was still learning (Objective-C was not his strongest), on a timeline that was aggressively ambitious.
In his previous life, this moment would have gone differently. Prometheus Labs Dojun would have said something cutting—”I’m aware of the issue,” or “When you’ve shipped a product, you can critique mine”—and the new hire would have learned to never challenge the founder’s code. The culture of deference would have set in early, calcified over years, and eventually made the company fragile.
“You’re right,” Dojun said.
The office got quieter.
“The blocking loop is a problem, and waiting six weeks to fix it will cascade into design decisions that are harder to reverse later. Can you do the async refactor this week?”
Taeyoung blinked. He had clearly been prepared for pushback, not agreement. “This week? The refactor is—it’s a significant rewrite. The callback architecture needs—”
“You identified the problem. You clearly understand the solution. I’m asking if you can execute it.”
“I—yes. I can do it. By Friday.”
“Then it’s yours. Prioritize it over whatever else is on your sprint. And Taeyoung—”
“Yes?”
“Keep doing this. Challenge the code. Challenge me. If you see something wrong, say it. That’s not insubordination. That’s engineering.”
After the meeting, Hana pulled him aside in the hallway.
“That was impressive,” she said.
“What was?”
“You let him be right. In front of everyone. Most founders—most people—would have defended their code. You gave him the win and the work.”
“He was right. Defending wrong code doesn’t make it right. It just makes the team afraid to speak up.”
“You learned that from your mysterious mentor?”
“I learned that from experience.” Forty years of it, including ten years of watching talented engineers leave Prometheus Labs because their ideas were never as welcome as the founder’s. “The team has to be smarter than any individual. Including me.”
“That’s a very un-founder thing to say.”
“It’s the most founder thing I can say. The company survives the founder. If it doesn’t, it’s not a company—it’s a ego project.”
Hana looked at him for a long moment. “Sometimes you say things that make me think you’re sixty years old.”
“I contain multitudes.”
“You contain contradictions. But the useful kind.” She punched his arm lightly. “Come on. We have a mobile app to build and a new engineer who’s about to learn that being right is the easy part. Being right and executing on time—that’s the hard part.”
The team grew fast in the summer of 2007. Taeyoung was followed by Park Soojin, a UI developer from Yonsei who had been one of Bridge’s most vocal beta testers (“I submitted sixty-three bug reports. I figured you’d either hire me or block my account.”). Then came Jeon Dongwook, a KAIST transfer student who Seokho had recommended with the endorsement: “He’s competent, reliable, and doesn’t argue about naengmyeon. Everything I value in a human being.”
Seven people now occupied the twelve-square-meter office, which had reached a population density that Minjae calculated was “technically a violation of the Geneva Convention, if the Geneva Convention applied to startups.”
“We need a bigger office,” Hana said during the Monday standup—now conducted in a tight circle because there wasn’t room for anyone to stand separately.
“The Innovation Center has a larger unit on the fifth floor,” Dojun said. “Thirty square meters. Double the space.”
“Double the space is still tiny. But it’s an improvement.” She made a note. “I’ll talk to the center manager today. We also need proper desks. And a second whiteboard. One whiteboard for seven people is a communication bottleneck.”
“Communication bottleneck,” Minjae echoed. “She’s talking about whiteboards like they’re server infrastructure.”
“Whiteboards are the most critical infrastructure in a startup. A team that can’t visualize its thinking can’t think together.” Hana drew a circle on the current whiteboard, which was so covered in notes that new writing had to be layered on top of old like geological strata. “This whiteboard is full. It’s been full since May. We’re making decisions based on buried layers of dried-out marker. That’s not agile development—that’s archaeology.”
“Two whiteboards,” Dojun agreed. “And the fifth floor unit. I’ll sign the lease this week.”
The move happened on a Saturday—all seven team members carrying desks, chairs, monitors, and the sacred wobbly chair (kept as a founding artifact) up three flights of stairs because the elevator was broken. His mother, who had closed the stall early to help, directed the furniture placement with the authority of a woman who had arranged banchan containers for thirty years and understood the principles of spatial efficiency at a molecular level.
“The desks face the window,” she instructed. “Natural light improves mood and productivity. I read that in a magazine in the hair salon.”
“Mom, you don’t need to—”
“I’m helping. This is what mothers do. Move that desk three centimeters to the left. Three. Not five. Three.”
Hana watched this with barely concealed delight. “Your mother is project-managing our office move.”
“She’s project-managing my life. The office is just the latest scope creep.”
By evening, the new office was set up—thirty square meters of actual working space, two whiteboards (one for product, one for engineering), a coffee machine that Minjae had donated from his apartment (“It’s a sacrifice for the team. My mornings will suffer, but the company will thrive.”), and a window that overlooked not a parking lot but a sliver of campus greenery.
“Upgrade,” Hyunwoo said, admiring the view.
“An upgrade we earned,” Hana said. She stood in the center of the room, arms spread, taking in the space. “This is real, you guys. Seven people. Thirty square meters. A mobile app in development. Investors interested. Four thousand users and growing.” She looked at each person in turn. “When I walked into a study room eighteen months ago to meet a CS student named Park Dojun about a group project, I did not imagine this.”
“What did you imagine?” Minjae asked.
“I imagined getting a B and forgetting about it.” She laughed. “Instead, I got a company, a team, and the best japchae in Seoul, courtesy of the woman who just rearranged our furniture.”
From the doorway, where she was inspecting the coffee machine with deep skepticism, Younghee said: “The japchae goes in the front.”
“It goes in the front,” everyone repeated, in unison, because somehow that had become the company motto.
But growth brought friction. Not the dramatic, explosive kind—the subtle, accumulating kind, like sand in a machine that’s still running but getting louder.
The friction was between vision and execution. Specifically, between Hana’s vision and the team’s capacity to execute it.
Hana designed fast. Brilliantly fast. Her mobile interface mockups were weeks ahead of the engineering team’s ability to implement them. She would present a new interaction pattern on Monday, expect a prototype by Wednesday, and be frustrated when the engineers needed until Friday—or the following Monday—to deliver something that met her standards.
“The swipe gesture isn’t right,” she told Taeyoung during a review session. “I specified a 200-millisecond animation with an ease-out curve. This is linear. It feels mechanical.”
“I implemented what the spec said.”
“The spec said 200ms ease-out. This is 200ms linear. They’re different.”
“The difference is imperceptible to most users.”
“The difference is perceptible to all users. They just can’t articulate why. That’s the definition of good design—the user feels it without knowing it.” She pulled up a comparison on her laptop. “Look. Linear on the left. Ease-out on the right. Which one feels better?”
Taeyoung looked. Then looked again. “…The right one.”
“Exactly. Because ease-out mimics how physical objects decelerate. Your brain expects it. Linear feels like a glitch, even if you can’t explain why.”
“Okay. I’ll fix it. But Hana—these micro-interactions take time. If every animation needs custom easing, the sprint timeline—”
“The sprint timeline serves the product. Not the other way around.” Her voice had an edge that Dojun recognized—the edge of a designer who was compromising on nothing because she believed that details were the difference between good and great.
She was right. She was also creating a pace that the team couldn’t sustain.
After the review, Dojun found her in the hallway.
“The ease-out thing,” he said. “You’re right about the animation.”
“I know I’m right.”
“You’re also pushing Taeyoung to a standard he can’t reach on the current timeline. He’s learning Objective-C while implementing your designs. Every custom animation is a research project for him.”
“Then he needs to learn faster.”
“Or you need to prioritize. Not every interaction needs custom easing in the first release. The launch version needs to be good. The second version can be great. We can’t ship ‘great’ if the team burns out before launch.”
Hana’s jaw tightened. This was the first real disagreement they’d had—not about the project, but about the process. About how fast to go and how much to demand.
“I refuse to ship something mediocre,” she said.
“I’m not asking for mediocre. I’m asking for prioritized. Custom easing on the main navigation transitions—critical. Custom easing on the settings page toggle—nice to have. We ship the critical first and iterate the rest.”
“Iterate. Everyone says iterate. It’s code for ‘we’ll fix it later’ and later never comes.”
“It came for Bridge. We shipped without the learning module and added it three months later. That was an iterate-and-deliver decision, and it worked.”
She was quiet. The hallway hummed with the distant sounds of the office—keyboard clicks, muffled conversation, the coffee machine doing whatever coffee machines do when they’re not being used.
“You’re managing me,” she said.
“I’m collaborating with you. There’s a difference.”
“Is there? Because right now it feels like you’re telling me to lower my standards.”
“I’m telling you to sequence them. Highest standards first. Everything else follows.” He met her eyes. “I believe in your vision, Hana. Every pixel, every animation, every interaction. But I also believe in this team. They need time to match your vision. Give them time.”
She held his gaze for a long moment. He could see the tension in her shoulders, the frustration of a perfectionist being asked to accept imperfection—even temporarily, even strategically.
“Fine,” she said. “Prioritized custom easing. Navigation transitions first. Everything else in version 1.1.”
“Deal.”
“But I’m reviewing every animation before it ships. Every. One.”
“I would expect nothing less.”
She exhaled—a long, slow breath that released some of the tension. “We had our first real fight.”
“Was it a fight? I thought it was a negotiation.”
“In design school, negotiations are fights with better typography.” A reluctant smile. “Okay. I’ll talk to Taeyoung. Apologize for the tone, not the standard. The standard stays.”
“The standard stays. The timeline flexes.”
“The japchae goes in the front.”
“Always in the front.”
She walked back to the office. Dojun stood in the hallway and let his heart rate return to normal.
This was it. The friction point that had cracked Prometheus Labs wide open. Not a single explosive argument, but the slow accumulation of a designer’s uncompromising vision against an engineering team’s practical limits. In his first life, he had sided with the code—with speed, with shipping, with “good enough.” Hana had felt unsupported, undervalued, overruled. The resentment had built like debt, compounding silently until the principal was too large to repay.
This time, he had sided with both. Her standards were right. The team’s limits were real. The solution was sequencing, not choosing. Prioritize, don’t compromise.
It was a small victory. A hallway conversation, not a boardroom battle. But small victories, accumulated over years, were the difference between a partnership that lasted and one that didn’t.
He went back to the office. Taeyoung was already refactoring the animation code. Hana was at the whiteboard, re-prioritizing the interaction specs. Minjae was testing the data pipeline. Hyunwoo was debugging the sync layer. Soojin was wireframing the settings page. Dongwook was reviewing the API documentation.
Seven people in thirty square meters, building something together. Arguing, compromising, iterating. The japchae in the front, the custom easing on the navigation, the learning module in version 1.1.
Not perfect. But growing.
And growing, Dojun was learning, was always painful. That was why they called them growing pains. Not growing-pleasures. Not growing-conveniences.
Pains. The kind that meant something was alive and getting bigger.