Three Percent

The next day, I rubbed at my eyes on the way in to work.

Ow. They're stinging again.

Would something different show itself this time? Judging by everything so far, it was probably—

A retrofit.

Whenever I set my mind on needing to do something, it felt as if my eyes went out and found the way.

Strange.

I had no idea how an ability like this had fallen into my hands, but I meant to use it, and be grateful for it. So I came in like any other day, read the Inform, and took the handover. After Senior Yeongseok and I finished the daily work, I didn't head for the engineers' desks. I stayed with the tools.

"Senior. I'll look over the equipment a little longer before I go."

"Oh? What for? Think something's broken again?"

"I want to look for retrofit points."

"Already?"

"Too soon?"

"It is early. Doesn't matter, though. Nobody ever wrote down which year you're allowed to start on retrofits."

Senior Yeongseok laughed and clapped me on the shoulder.

"And if it's you, I have a feeling you can pull it off."

"Thank you."

"Go on. Look around at your ease. I'll keep an eye on the line."

A man like light itself, as always. Once he was gone, I went over the tool from one end to the other, half expecting something new to appear. The moment I actually searched for it, though, nothing new showed itself at all.

What?

Which was as good as saying this old machine had not a single point left worth retrofitting.

Because it's old?

Maybe the seniors had already improved everything there was to improve. Or maybe my ability simply hadn't manifested yet.

That can't be it.

The sting still lingered at the corners of my eyes. Something had changed. It just hadn't shown itself yet.

So it isn't a retrofit?

I was turning that over when—

Flash.

Something streaked past the edge of my sight.

"...Huh?"

I looked ‍‍‌‌⁠​​⁠‌​​‍​​‍​‌​⁠‍‌‍⁠⁠‌⁠‌‍‌‍‌‌‌‍⁠​‌​‌​‌‍‌‌‌⁠‍​​‌‍⁠‌‌⁠⁠around, but the flare didn't come again. Then, on a whim—

"A retrofit...?"

The moment I murmured it—

Flash.

The light returned. Blue.

I can see it.

It wasn't like the white light, or the red. The light only appeared when I was holding a specific purpose: the purpose of finding a retrofit.

So it decided this wasn't essential.

Retrofit work was secondary. You didn't have to do it. Keep the tools running decently and the job was done, so my ability seemed to have rewritten its own mechanism, firing only under certain conditions. I shrugged. The long way or the straight way, either one still got you to Seoul. What mattered now was that there was light in my eyes.

Let's have a look.

Blue light bloomed here and there across the tool, sparse as mold. It was a little unsettling.

So that's why it makes me switch it on and off.

If I left the ability running, the whole world in front of me would drown in blue.

Then what am I supposed to do with this?

Countless ‍‍‌‌⁠​​⁠‌​​‍​​‍​‌​⁠‍‌‍⁠⁠‌⁠‌‍‌‍‌‌‌‍⁠​‌​‌​‌‍‌‌‌⁠‍​​‌‍⁠‌‌⁠⁠choices lay in front of my eyes, and I was certain that whichever I picked would lead to a retrofit. Only—

Knowing a part needs improving isn't the same as knowing how to improve it.

That was the problem. What mattered was finding a piece I could actually act on, something matched to what the ability could show me. The thought had barely formed when—

Flash.

One patch of that blue mold burned brighter than the rest. I looked.

An MFC?

The blue light was coming from an MFC mounted low on the tool. A mass flow controller regulates the flow of a given gas, and a tool like this carried a great many of them. Only one was glittering now: an MFC that ran on UPC. Upstream pressure control.

If I remember right, this is the MFC that flows argon for heat transfer to the wafer.

The pressure between wafer and pedestal is decided by that MFC. Once the pressure set in the recipe is reached, the process proceeds. What kind of improvement could an MFC like that possibly have? I focused a little harder, and realized the blue wasn't spread over the whole unit. It had gathered in one place.

It's a switch.

An ordinary MFC only needs its zero point set through the board inside it, ‍‍‌‌⁠​​⁠‌​​‍​​‍​‌​⁠‍‌‍⁠⁠‌⁠‌‍‌‍‌‌‌‍⁠​‌​‌​‌‍‌‌‌⁠‍​​‌‍⁠‌‌⁠⁠and after that it flows gas perfectly well on its own. This one, though, carried DIP switches. Ten toggle switches in a single row, flipped up and down one by one to set the operating values. Blue light winked along those switches.

...

I didn't get it. Why this, of all things, was a retrofit point. Wondering if it was only this tool, I checked the others. Every one of them glittered the same way.

"Hmm..."

An improvement existed. That much was certain. All that remained was finding it. I headed for the engineers' station.

"Nnngh..."

And there was Senior Yeongseok, clutching his head.

"Senior. Something wrong?"

"Mm. No—manufacturing came asking us to push productivity higher."

I checked the work messenger. It was true: volume was heavy, so productivity had to be looked after. But every tool was already running. Maintenance had been solid for long enough that nothing was even down.

This is awkward.

To squeeze any more productivity out of the line, we would have to add tools.

"Haa. Nothing for it. So. Find anything?"

Senior Yeongseok asked.

"Yes. I found something."

"Yeah, finding it on the first try is tou—what? You found something?"

"Yes."

I sat down and searched the model number of the MFC I'd just been staring at, then went into the manufacturer's site and pulled up the manual. The product manual. The DIP switches were in there too.

"Senior, do you happen to know about this?"

"Hm? Know of it, sure. How to use it? Not really."

According to him, when an MFC came in, everyone simply mounted it as it arrived. Nobody bothered changing the settings.

"I see."

"But why this?"

"One moment."

I ran through the manual at speed. The section on the DIP switches was thorough.

Mm. This is about PID control.

Proportional, integral, derivative. The ten switches were split into three bundles—three, four, and three—taking on proportional, integral, and derivative in turn. Set a value high and it reacted sharply; set it low and it went dull. The role of the MFC, and the role of the DIP switches. Set the two side by side, and one thought shot clean through my head.

This could actually work.

A process that starts only when the pressure between wafer and pedestal hits its setpoint. An MFC that regulates that pressure. And, past even that, the DIP switches that govern how the MFC is controlled.

"If this works, we might be able to shorten the time it takes to run a process."

"What? Is that for real?"

Senior Yeongseok answered, startled. He had reason to be.

If process time comes down...

That was as good as productivity going up. We might even be able to solve the headache that had manufacturing—and us—clutching our heads. I laid out what I was thinking.

"We might be able to cut the time it takes for the pressure between the wafer and the pedestal to reach the setpoint."

"How?"

"By adjusting the MFC's DIP switch values. Make the initial response faster."

I looked at him.

"I don't know if it'll actually work, though. Want to try?"

"Right now. Right now!"

Backed by that fervent support, I headed for a single tool. WN-8. I'd already entered a local code—a standby code—so it sat paused for a little while.

It'll only be a moment.

I remembered the manufacturing lead watching me with a look sharp enough to cut, the moment I'd said I was stopping a tool. The memory sent a shudder through me for no good reason. I adjusted the DIP switches on the UPC MFC as fast as I could.

Faster response up front.

So the MFC could flow quickly up to the set pressure. That was all. The blue light that had been flashing was gone. Just in case, I pulled the dummy data as well. Normal.

"Senior. Let's put the tool back in use."

"Hm? You're done already?"

"Yes. It wasn't complicated."

After that I walked the line with an easy mind. The result wouldn't show itself right away in any case.

We'll have to check productivity... tomorrow ought to be enough.

I was curious how much would change from the moment I'd touched those switches.

The following day I came in no different from any other, and spent a very busy morning. Several failures had blown up all at once.

Rough.

The line flashed white, over and over. Senior Yeongseok and I cleared fault after fault, error after error, and only then dropped into our chairs. I had just caught my breath when a message went up in the WN part chat.

"Seokhyeon. The deputy's looking for you?"

"Yes, Senior. I'll be back."

I hauled my heavy body up and shuffled off the line.

Did something go wrong?

Uneasy for no good reason. Just in case, I looked with my eyes one more time. No white light on WN-8. No red.

It's probably nothing.

I walked into the office with my nerves pulled tight.

‍‍‌‌⁠​​⁠‌​​‍​​‍​‌​⁠‍‌‍⁠⁠‌⁠‌‍‌‍‌‌‌‍⁠​‌​‌​‌‍‌‌‌⁠‍​​‌‍⁠‌‌⁠⁠

"Deputy. I'm here."

"Oh? Seokhyeon, that you? WN-8?"

"Yes, that was me. If I might ask what this is about..."

I asked carefully. Deputy Kim Soonhyeop broke into a grin that split his face.

Three percent. A number that looked small at a glance, and one that absolutely could not be ignored. On a tool that runs twenty-four hours, if a hundred wafers went through in a day, now a hundred and three would. Ten days, and that was thirty more. A hundred days, and three hundred extra wafers.

It worked.

The improvement had landed.

Another success.

Seeing a result I could actually hold made me feel good. I explained to the deputy what I'd studied and what I'd set. He listened without a word, then nodded.

"That so? Then apply it to every tool."

"Every tool?"

"Yeah. And compare the productivity against before, and send it up by email. Say you did the retrofit."

"Understood."

I bowed to the deputy and headed back to the line.

The Line 1 office.

Department Head Jo Heesang had been knocked off balance by a call that came out of nowhere.

A call from the manufacturing team leader. The sudden rise in productivity had brought a confirmation call.

What is this about?

Higher productivity was a good thing. But sometimes, unhappily, bad luck stacked on top of itself. A step skipped in the middle of a process—that sort of thing.

We'll have to check.

The dummy-data inspection history already confirmed the tool was normal, but he still needed to know whether anything else had changed.

"Deputy Kim!"

"Yes, Department Head."

"Manufacturing says WN-8's productivity went up. What's going on?"

"Ah. Well, the way it happened is..."

Deputy Kim Soonhyeop rose from his seat, all smiles. The situation was what it was, and the man was smiling? And yet it was a smile that somehow put you at ease.

"That was Engineer Lee Seokhyeon again."

"Engineer Lee Seokhyeon?"

The name had been coming up a lot lately. WN-15, last time. The water leak on WN-11. And recently, getting budget manager Park Seongwoo tossed out.

The kid's ability is anything but ordinary.

Anyway.

"What did this engineer do?"

"He made an improvement. There's something called a DIP switch..."

Deputy Kim Soonhyeop explained, point by point, everything he had heard from Lee Seokhyeon. When the story was finished, the department head stared at him, startled.

"A new hire of two months did that?"

"That's what I'm telling you!"

Then it wasn't an accident. It was cause for celebration. Productivity, up by three percent.

"So. You told them to apply it to the other tools as well?"

"I did. They're probably finishing the settings about now."

"Right. So that's how it is."

The department head nodded.

"Next month is evaluation season, isn't it?"

"It is. ...Department Head, you don't mean—"

"If productivity rises on every other tool as well..."

A performance grade—that much, he could give. Even if, come next month, the man would still be a new hire only three months in.

"You talk to him properly, Deputy Kim. He probably won't even know it's coming."

"...Understood."

The veterans might complain, if they found out.

Well. The work is real.

It was more than enough to earn the grade.