It's difficult to transfer it to my phone on the way home from work

I have an evening appointment so I'll stop here for today

An Era Where No One Knows Again

Chapter 3

The first time I properly used AI was because of an operational failure.

Usually, I would have searched through the logs first.

I would have found the time period when the error occurred, checked the called API, looked at the database query, and if necessary, even traced back to previous deployment history.

I did the same thing at first that day too.

The log contained a familiar error message.

Being familiar didn't mean I knew the cause.

It was closer to meaning I had seen similar messages multiple times.

I was about to copy a few lines and paste them into the search bar, but I stopped.

And instead, I pasted it into the AI window I had open next to me.

Can you summarize the possible causes for this error?

The answer came right away.

There were five possibilities.

Three of them were not it.

I had already checked one, and the other one couldn't happen given the current system structure.

The last one was a bit suspicious.

I added more of the log.

Looking at this log, which possibility is most likely?

The answer changed.

It said that it was most likely that a specific value was empty during the data transformation process.

I checked the database.

It really was empty.

I stared blankly at the screen for a moment.

"Found it."

I muttered to myself.

Of course, AI didn't solve the failure.

I already knew which table to look at.

I also knew why that value couldn't be empty. I also knew that it couldn't be fixed immediately in production.

Still, it was faster than usual.

Much faster.

I fixed the problem and sent a reply to the operations team.

Confirmed the cause. It's a data exception case. I'll fix it and apply the changes.

After sending the message, I looked at the AI window again.

All I needed was just a few lines of logs.

Back then, I would have opened several search results, read whether they were similar cases or not, and changed my search query again. Since ads started mixing into search results, finding the answer I wanted also became more annoying than before.

At least AI answered my questions.

Though it was sometimes wrong.

A few days later, I pasted in some old source code.

It was code from ten years ago.

The author had already left the company, and there were almost no comments.

The methods were excessively long, and there were unfamiliar abbreviations in places I had never seen before.

In the old days, I would have traced through it line by line from the top.

I asked AI.

Explain what this code does overall.

The answer came.

It was roughly correct.

There were some details that were wrong, but the overall flow was captured.

I asked a bit more.

Where in this section could problems occur?

It pointed out a few places.

One of them was a part I had already suspected.

I leaned back in my chair.

I felt a bit strange.

It felt like I was working with someone.

And someone who understood me quite well.

When I ask the developer next to me, there's a sense of guilt.

There's the awkwardness of asking trivial things when I know they're busy, and it's hard to ask again if I don't understand after hearing a long explanation.

AI didn't have that.

Even if I asked the same thing three times, it didn't get annoyed.

If I asked strangely, it just answered strangely.

I liked that aspect.

I started changing my questions little by little.

Explain this more simply.

Create an example based on Java.

Is there another method besides this one?

Since it's a system in operation, the scope of changes needs to be small.

The answers became increasingly useful.

At some point, I started enjoying asking questions.

During lunch, a young developer came to my desk.

"Senior, do you use that a lot?"

He pointed to the AI window on my screen.

"I use it a bit these days."

"It's good, right?"

"More than I thought."

"I ask this almost first."

That friend turned his laptop around to show me.

There was a long conversation between him and AI.

He asked for code to be created, pointed out mistakes, had it fixed again, and even had test code created.

I looked at the screen and said.

"But you need to verify this yourself."

"I do."

"It talks about being wrong with so much confidence."

"That's right. So you can't just believe everything it says."

The answer was more sound than I expected, so it was a bit disappointing.

I seemed to have wanted to teach him a lesson.

The young developer said.

"But it's convenient. It reduces the time spent finding things you don't know about a lot."

I had to admit that.

"That's true."

"How did you do it back then?"

I thought for a moment.

"I searched for it."

"And before searching?"

"I looked at books."

That friend laughed.

"Did you really do everything with books?"

"I didn't do everything with books."

I laughed too.

I thought that phrase kept coming out strangely.

Did you really develop with books?

I wondered if that was such a distant story for developers these days. That afternoon, I had to use a library I had never seen before.

In the old days, I would have opened the official documentation first.

I hesitated for a moment, then asked AI first.

Tell me the basic usage of this library.

Example code appeared.

I copied and pasted it.

An error occurred.

I gave the error message back to AI.

AI fixed the code.

Another error occurred.

I asked again.

On the third try, it worked.

I stared at the screen for a while.

There were moments like this in the past too.

When I typed an example from a book exactly as written and saw the result on the screen for the first time.

I didn't fully understand why it worked back then either.

I just liked that it worked.

I moved the mouse and looked at the code again.

I laughed a little.

It was nothing special, but it was fun.

I had forgotten that feeling while developing for so long.

Learning new technologies became a burden rather than fun at some point.

Do I have to study again?

Did it change again?

How many years will this last?

I thought things like that first.

Building a career meant knowing more things, but it also meant becoming more skeptical when seeing something new.

AI was a bit different.

It seemed like it was studying things I didn't know for me.

More precisely, it made it possible to start without studying.

I liked that aspect, but it made me anxious.

I read the executed code again.

I didn't know something.

If it were like before, I would have looked for the document until I understood that part.

At least I remember doing so.

After thinking for a while, I asked the AI window again.

Why is this part necessary in this code?

An explanation came out.

I read it and nodded.

It seemed like I understood everything.

Whether I really understood it or not, I don't know.

Even though it was past quitting time that day, I was still sitting at my desk.

It wasn't because I had overtime work to do.

I was trying out various things with the AI.

Normally, I would have left it for tomorrow, but I was making a small function.

I asked it to create the code.

I didn't like it, so I had it fix it.

I also asked it to create a test.

One of them failed.

I asked why it was failing.

The AI said its code was wrong.

I laughed alone.

"You don't know either."

I had it fix it again.

This time the test passed.

There were hardly any people left in the office.

A few desks with only the monitor lights on could be seen.

After a while, a young developer walked by with his bag and saw me.

"Senior, are you still here?"

"Yeah. Just looking at this."

He glanced at my screen.

"Isn't it fun?"

I didn't answer and just laughed.

He also smiled and left first.

A little later, I closed my laptop too.

While waiting for the elevator, I thought.

It had been a long time.

I wanted to do a little more even though work was done.

That feeling was pleasant.

At the same time, it was a little funny.

When I was in my twenties, I used to stay up all night opening a book and typing unknown code. Now I was fixing code made by AI and delaying going home.

The method was different, but the feeling was similar.

There was something unknown in front of me,

and I felt like something would happen if I touched it a little.

The elevator door opened.

I went inside alone.

Just before the door closed, my cell phone rang.

It was a message from the young team leader.

Senior, let's talk briefly tomorrow about what we discussed earlier about using AI. There seem to be more uses than I thought.

I read the message and put my cell phone in my pocket.

The elevator started going down.

Strangely, I didn't feel bad.

Rather, I felt a little excited.

At least for now.