Chapter 4

The crisp autumn air of Cambridge felt like a new element, a clean slate. I loved everything about MIT-the frantic energy, the smell of old books and ozone in the libraries, the way you could start a conversation with a stranger about quantum computing and they wouldn't look at you like you had three heads. For the first time, I felt like I was in the right place, surrounded by my people.

I wasn' t just surviving; I was thriving. My classes were challenging in the best way possible. I found a small group of friends in my advanced AI seminar-Maya, who could code circles around anyone I knew, and Ben, a robotics genius with a dry sense of humor. We spent hours in the lab, fueled by cheap pizza and a shared obsession with making our machines smarter. They were genuine, brilliant, and funny. There were no hidden agendas, no emotional blackmail. Just friendship.

My work on the neural network project, the one the Petersons had tried to steal, was progressing faster than I ever could have imagined. With the university's resources and the feedback from my professors, it was becoming more powerful, more elegant. I was reclaiming my passion, and it felt incredible.

One afternoon, studying in my dorm room, an email popped up. The sender was an address I didn' t recognize, but the name made my stomach clench: `jake.peterson@cit.edu`.

I clicked it open against my better judgment. It was a long, rambling message, a toxic cocktail of self-pity and blame. He wrote about how hard his classes were, how much pressure he was under, how I had "overreacted" and "ruined his family's reputation" for no reason. He said I was selfish for abandoning him.

I stared at the words, feeling a distant, cold anger. I didn't reply. I just dragged the email into the trash and then emptied the trash folder. He was a ghost from a past life, and I refused to let him haunt me.

A few weeks later, Maya convinced me to go to a city-wide tech mixer. "You need to network, Sarah," she'd insisted. "All the big startups and VCs will be there. You can't hide in the lab forever."

She was right. I put on a decent dress, took a deep breath, and went. The venue was a trendy warehouse space downtown, buzzing with ambitious energy. I was starting to enjoy myself, talking to a grad student about his drone navigation system, when I saw them.

Across the crowded room, standing near the bar, were Jake and Emily. They were here. Of course they were. CIT was just a short train ride away. They looked the part-Jake in a slick blazer, Emily in a sharp, expensive-looking dress. They hadn't seen me. I felt a sudden urge to flee, to run back to the safety of my dorm.

But then I thought, no. I'm not hiding. This is my world now, not his.

Later, the event organizers started a "Tech Tussle," a team-based trivia game. By some cruel twist of fate, my team ended up competing directly against the one Jake and Emily were on. The questions were hard, covering everything from historical programming languages to cutting-edge biotech.

It was the final round. The score was tied. The host read the tie-breaker question: "Explain the core limitation of the backpropagation algorithm in deep recurrent neural networks and name two common methods used to mitigate it."

Silence fell over the room. On Jake's team, Emily grabbed the microphone, a confident smirk on her face. "The limitation is unstable gradients," she said, then faltered. "And you mitigate it by... uh... using better data?"

The host shook his head. "I'm sorry, that's incorrect. Team MIT, do you have an answer?"

Maya nudged me. "That's your specialty, Sarah."

I took the microphone, my voice clear and steady. "The limitation is the vanishing gradient problem, where the gradient shrinks exponentially as it propagates back in time, making it impossible for the network to learn long-range dependencies. You can mitigate it using techniques like Long Short-Term Memory, or LSTM, units, or Gated Recurrent Units, GRUs, which use gating mechanisms to control the flow of information and preserve the gradient."

The host beamed. "That is absolutely correct! Team MIT wins!"

Our team cheered. I caught Emily's eye across the room. Her face was a mask of furious humiliation. Jake just stared at me, his expression unreadable.

Later, as I was getting a glass of water, he cornered me near the exit.

"You just had to show off, didn't you?" he hissed, his voice low and venomous.

"It was a trivia question, Jake," I said, not backing down. "I knew the answer."

"You think you're so much better than everyone else now that you're at your fancy school," he sneered. "You overreacted. You took one small mistake and used it to ruin my life."

"A small mistake?" I looked him dead in the eye. "You and your family tried to steal my future and pass it off as your own. That's not a mistake. It's a conspiracy. Now, if you'll excuse me, I have to get back to my friends."

I pushed past him. As I walked away, I noticed something. Dangling from the zipper of his jacket was a small, tarnished silver keychain-a circuit board I had given him for his sixteenth birthday. A pathetic, clinging remnant of a friendship he had purposefully destroyed. The sight of it didn't make me sad. It just made me feel tired.

The next day, Maya forwarded me a screenshot. It was from a private student forum at CIT. The post was anonymous, but the language was unmistakable.

"Watch out for the 'prodigy' from MIT, Sarah Miller. She's not as brilliant as she seems. Steals ideas and stabs her friends in the back. Word is, her whole 'Nightingale' project is based on research she lifted from someone else."

Underneath it was a single reply, from Emily Chen's verified account.

"I can confirm this. She's a user and a snake. Someone should teach her a lesson."

The cold dread was back. They weren't just haunting me. They were actively trying to destroy me.