Chapter 4What Came With the Stolen Code

The first two students gave decent presentations. One was on smart traffic optimization, the other on NLP sentiment analysis. The committee asked some standard questions, and the atmosphere remained calm and academic.

Then the moderator announced, "Next up, Gina."

Polite applause rippled through the room as Gina stood up and walked to the podium. Wearing a tailored white shirt, black dress pants, and an elegant updo, she looked every bit the rising star of academia. If you didn't know any better, that is.

She plugged in her flash drive, and the projector flared to life with her title slide.

"Good morning, members of the committee. My research is focused on multi-objective path planning optimization using deep reinforcement learning."

Yep. That title. It was the exact title of my v0.3 build back in August.

I had completely shifted gears by v1.0, abandoning path planning to focus on graph neural networks in dynamic topologies. She hadn't just stolen outdated code; she had stolen a research direction I had already discarded.

"The core innovation of this study lies in a modified DQN algorithm, which integrates an attention mechanism to achieve multi-objective balance."

I listened, my lips twitching. That was the abstract from my v0.3 documentation, lifted word-for-word.

The committee members nodded as she spoke. The first five minutes of her presentation were smooth. Anyone can memorize a script and click through pre-made slides. But the real test of a defense isn't the presentation. It's the Q&A.

Gina concluded her final slide with a polite, "Thank you. I welcome your feedback."

Professor Harrison flipped through the printed report in his hands. He looked up, pushing his glasses up his nose.

"Gina, in your modified DQN algorithm, what is the theoretical foundation for your reward function?"

Gina froze for a fraction of a second.

"Uh... it is based on the Pareto optimality theory in multi-objective optimization."

The answer itself wasn't wrong, but Professor Harrison didn't stop there. "And how exactly did you map the Pareto frontier to the reward signals? What was your normalization strategy?"

Gina's lips parted. "It was... a weighted sum approach."

"Weighted sum?" Professor Harrison frowned. "Your slide on page seven shows adaptive weights. Why are you suddenly talking about a simple weighted sum?"

Gina's fingers tightened around the presentation clicker until her knuckles turned white.

"I apologize, Professor. I misspoke. I meant to say an adaptive weighting approach."

"Then how did you implement the adaptation mechanism? How did you tune the hyperparameters?"

Silence. Three seconds. Five seconds.

Gina's mouth opened and closed, but no sound came out. That kind of silence is the kiss of death in a defense.

Another committee member interjected, "Why don't you show us your source code? If we look at the actual implementation of this algorithm, it should clear things up."

I sat in the audience, my breathing steady and my pulse calm. But if anyone had looked closely, they would have seen a slow, wicked grin creeping onto my face. I knew what was coming. The second she opened that source code, the real show would begin.

Gina hesitated, her hand trembling as she opened the file explorer. She located the project directory and double-clicked the main script.

Lines of dense code flooded the projector screen, visible to the entire room.

The first line was the standard file header. The second and third lines imported the libraries.

The fourth line, however, was a comment I had written at 11:00 PM on August 20th.

It stared back at the entire room in bright green font: // Another day of burning my life away for science. Too bad I'm burning my liver, not writing actual papers.

A quiet chuckle rippled through the back of the lecture hall.

Professor Harrison adjusted his glasses, his gaze locking onto the screen. Gina panicked, her fingers frantically scrolling down.

Line 47.

// If you don't understand this code, don't worry. I wrote it, and I don't understand it either. Let's all be confused together.

More laughter broke out, louder this time. Gina's face drained of color. She scrolled faster, trying to find the core math, but she had no idea how the project was structured. The faster she scrolled, the worse it got.

Line 128 flashed onto the screen.

// TODO: There is a bug here, but it's Friday. Even bugs deserve a weekend. I'll deal with it next week.

The lecture hall erupted into giggles. Even some of the committee members couldn't hide their grins. Professor Harrison's face remained stoic, but I knew him well enough to see the slight flare of his nostrils. He was holding back a laugh.

Gina finally reached the core algorithm section between lines 203 and 250. Right above the main function was a massive, glaring comment block. Nobody could miss it.

// If Garrett is peeking at my screen again, please raise your eyes by three centimeters and look at this line. What are you looking at? Go write your own.

The laughter stopped instantly. It was replaced by a wave of sharp, sudden whispers.

Garrett's name was printed in plain English on the giant screen. People in the audience immediately began turning their heads, searching the room.

Out of the corner of my eye, I saw Garrett sitting in the second row. He looked completely petrified. In a span of two seconds, his face flushed a deep crimson and then went entirely sheet-white.

Professor Harrison slowly took his eyes off the screen and looked directly at Gina.

"Gina," he said. His voice wasn't loud, but the room went so quiet you could hear the low hum of the ventilation system.

"This code explicitly mentions your colleague's name. Can you explain what this comment is doing in your program?"

Gina's hand shook violently. The presentation clicker slipped from her fingers, clattering loudly against the wooden floor.