Translator: Dreamscribe
The hallway air was still cold, as if the early morning chill hadn’t yet faded.
Ttak-ttak-ttak.
The sound of several people's footsteps echoed there.
At the front, Alan walked. The hem of his heavy coat brushed against his legs.
Behind him, Lisa, Michael, and other researchers followed silently, holding their breath.
As they reached the end of the hallway, a naplate ca into view.
[Room 2-148. Lead Investigator – Yoo Seoha]
Beep-
When he tapped the card, the door opened.
Alan carefully stepped inside.
“Hm?”
From within the unlit lab, the faint sound of piano music could be heard,
It was a sorrowful tone, as if consoling soone’s heart.
Click.
When the light was turned on, the room’s interior was revealed.
“Holy shit!”
“Ah....”
“What on earth is this....”
Everyone looked around with their mouths agape.
A massive equation covered three walls of the lab, greeting them.
Formulas written densely in white chalk,
Intricate lines and symbols intertwined, forming a vast algorithm.
“Amazing.”
To Alan’s eyes, it didn’t look like a language created by humans, but rather a circuit moving and living on its own.
Handel's Passacaglia resonated majestically.
Chalk dust slowly floated up, reflecting the fluorescent light. It looked like white snow scattering from the sky.
This surreal scene made everyone forget why they had co here.
Whirring-
A simulation was running on the workstation placed on the desk.
A log window was displayed on the screen.
[Simulation complete.]
[Error: none detected.]
"Over there."
Soone whispered while pointing to one side.
On the sofa in the corner, Seo-ha was sleeping quietly.
White chalk dust covered his entire hands, as if he had fought a long battle.
“Seo-ha....”
“Shh!”
As Lisa approached to wake him, Alan put his index finger to his lips, gesturing for silence.
It seed he had fallen asleep lying on the sofa after concentrating for a long ti.
No one spoke, not wanting to wake him.
Lisa’s gaze slowly moved.
Seo-ha’s fingertips, the formulas filling the chalkboard, and back to the monitor showing the completed simulation.
His sleeping face looked peaceful.
Rustle.
Alan took off his coat and draped it over Seo-ha’s body.
“Let’s leave him be for now.”
Everyone nodded.
Hands covered in white chalk, the satisfied smile on his sleeping face, and the perfect equation he had left on the board.
A strange emotion surged in one corner of Alan’s chest.
Pride, and at the sa ti, guilt.
'We entrusted too much to you.'
Until he felt the pressure of ti and budget, Seo-ha must have planned to collaborate.
Like with that paper, which had been more like a group assignnt.
But the situation didn’t leave him alone.
‘Unbelievable. With so many talented people gathered here, there wasn’t a single person who could exchange thoughts on equal footing with you?’
Alan thought perhaps Seo-ha might be an unbearably lonely existence.
***
MIT dia Lab E14.
Though it was evening, none of the project mbers considered leaving work.
Much had changed overnight.
The formulas on the chalkboard were scanned with high-resolution caras and displayed on a large screen.
The research team sat around the monitor, analyzing the structure of the algorithm Seo-ha had left behind.
Lisa pointed to the screen with her finger.
“Look here. This is different from the structure we used before. It looks hierarchical?”
“Hierarchical?”
Michael frowned and zood in on the screen.
“This seems like a direct mathematical expression of a neural network... but it’s not a typical neural net. Each layer has different rules?”
Lisa nodded.
“That’s right. Each layer operates on a different ti and scale. One layer works in seconds, another in microseconds. It’s completely nonlinear.”
“Is that even possible?”
Soone murmured.
“Look.”
Lisa tapped on the keyboard and opened part of the formula.
“This isn’t based on the traditional Fourier transform signal-over-ti domain. It induces self-synchronization on its own.
Seo-ha probably concluded that the previous error stemd from the limitations of hardware in parallel simultaneous computation and decided to pull the concept of ti itself entirely into the internal calculations.”
A mont of silence passed.
They tried to understand the formula, but its structure didn’t imdiately co into focus. After staring at the screen for a while, the researchers began to grumble.
“Have I gotten too old? I can’t do this anymore.”
"At this level, we should give the algorithm a na, don't you think? It's like math from another dinsion."
“There was a good reason MIT gave him a lab.”
Lisa continued speaking.
“What’s amazing is that this isn’t unstable at all.
Normally, structures this complex collapse imdiately with even the slightest error, but this one instead accepts that in real ti and stabilizes itself.”
She pointed to the workstation running the simulation.
The graph on the monitor was moving with astonishing smoothness. The data curve, which used to shake like a vibration when a crash occurred, now rose smoothly without a single tremor.
“CPU load rate at 48 percent?”
Michael stared at the screen in disbelief. The load, which used to exceed 80% and approach 90%, had dropped to nearly half.
“This is the sa dataset, and it’s performing like this?”
“Clock sync is perfect too. Before, it lagged by 0.03 seconds every ti it did parallel processing, but now the deviation has dropped by more than a factor of 100.
Looks like he factored in delays caused by the hardware right from the design phase.”
The researchers gathered and examined the logs.
‘We thought there was no more room for improvent before....’
It was a staggering, groundbreaking change. The fact that this had been created in just two days left everyone in shock.
Alan, who had been quietly listening, nodded.
“Let’s run the full verification set with this as the final version.”
“Yes!”
Voices rang out powerfully.
With the completion of the project in sight, energy returned to the lab.
The cursor on the monitor blinked, then commands were entered.
A mont later, a loud whir ca from the workstation.
Whirrrrr.
As Lisa hit enter, dozens of simulation windows opened simultaneously.
“12,000 test sets input complete. Beginning full-range verification now.”
Alan made eye contact with each person in the lab.
“For the next 24 hours of testing, we’ll be in ergency operation mode.
Back up all logs every 10 seconds. Imdiately report any errors, and no actions except for manual shutdown are allowed under any circumstances.”
It was the final phase. Tension spread across everyone’s faces.
“Yes!”
Michael chuckled.
“So we’re pulling an all-nighter after all.”
This was a familiar situation for researchers.
“This is still very manageable.”
“We’ve done it for a week straight before, so what.”
“The work itself isn’t the problem. It’s not getting results that’s painful.”
Everyone nodded in agreent.
“But the expectations are higher this ti. I can’t even imagine how much this city will change.”
“Hey now! We’re still testing. Don’t jinx it!”
“But who’s doing the maintenance?”
At the final researcher’s words, everyone fell silent.
If soone tampered with it and triggered an error, it could result in a city-scale disaster.
Was there even anyone at city hall capable of handling such a complex algorithm?
“What’s the problem? Just contract directly with the designer.”
Alan snapped.
“Well, Seo-ha could write the manual or train the staff.”
“Or give him a position as an advisory board mber.”
Alan thought about ways the city could give Seo-ha appropriate compensation for his contribution.
What he ca up with was this thod.
As long as Boston used Seo-ha's system, he would be able to receive compensation like a pension.
***
Seo-ha woke up around evening.
‘Hm?’
Sothing about the situation felt familiar. Seo-ha sensed a cold chill creeping up his spine.
‘I hope it’s not the hospital again....’
He cracked open his eyes and looked to the side, it was the lab.
‘Whew! Thank goodness.’
Seo-ha had taken a lot of care this ti to avoid any unexpected accidents.
Since many people here worked themselves to the bone in their research, it was easy to establish a proper calorie plan. As long as he ate plenty of chocolate bars and high-calorie drinks and got so adequate sleep, nothing like before would happen again.
‘I guess I was a little sleep-deprived....’
He had closed his eyes from ti to ti to avoid collapsing, but it seed the tension had released after completing the algorithm.
Rustle.
As he sat up, a familiar face ca into view.
“Ellie?”
Eleanor looked up from the book she was reading.
“You’re awake. The doctor ca by. Said you were just asleep. Accumulated fatigue.”
Seo-ha scratched his head sheepishly.
“I don’t feel sick at all though.”
Ellie gave a pained expression.
“Honestly, I knew you were overdoing it. Not just , everyone did.”
“Huh?”
“But no one could bring themselves to stop you.
The project looked like it was about to fall apart without funding, and you were desperate with the algorithm. I’m just an assistant, but I think everyone believed you were the only chance left for this project.”
‘I didn’t think it was that hard....’
It looked like there had been another misunderstanding. Seo-ha lowered his head for now.
“I’m sorry for making you worry.”
Ellie shook her head.
"No, I'm reflecting on myself. I saw you researching alone at dawn. I wanted to go in and help like during the team project, but I didn't know what to do.
Right now, I can’t even properly understand the algorithm you made.”
Ellie lowered her head and stayed silent for a while.
She fidgeted with the edge of her book before speaking quietly.
“At first, I was so happy to be a part of this project.
But all I’ve done is organize data or take eting minutes. Ah! I also reserved equipnt and set up the eting rooms.”
Ellie let out a self-deprecating laugh.
“I know. I’m just an assistant, so that’s expected. But seeing everyone rely only on you… it made feel bitter, like I was overstepping.”
“Ah... no, that’s not it....”
Seo-ha floundered, unable to find the right words to comfort her.
“So what I wanted to say was, next ti, I want to be a real team mber, not just the copy machine girl. I’ll have soone else fetch the coffee.”
She clenched her fist playfully, as if she had made up her mind.
“Well, I’ll get going now.”
Ellie waved with a smile.
“Yes! Thank you. And you said you didn’t do anything, but you really helped a lot. Without you, I probably wouldn’t have been able to complete the initial algorithm.”
Ellie’s eyes widened at the unexpected words.
“Hahaha.”
Tears stread down her cheeks.
Not wanting him to see, Ellie quickly hurried outside.
After walking for a while, she saw a vending machine.
Her hand trembled as she reached for a coffee.
Thud!
Ellie slamd her forehead into the vending machine, as if blaming herself.
Thud.
Thud.
Each ti the cold tal touched her forehead, her mind grew clearer.
The world of mathematics, black and white, right and wrong, so clearly defined.
And because of that, she could see it all too cruelly and clearly.
“Why am I so stupid.”
She had set aside all her other studies just to try and understand Seo-ha’s algorithm.
But in this world, there were walls that willpower alone couldn’t overco.
Ellie took a notebook from her pocket.
The edges were already worn and discolored. Each page was filled with equations she hadn’t been able to understand.
Swipe.
Wiping away her tears with the back of her hand, Ellie set off to find a place to study.
User Comments
0 comments from readers