The email from Henderson, my boss, had landed in my inbox three months before the "strategic decoupling." Subject: "New Prioritization Initiative - Confidential." Henderson wasn't big on clear subject lines, or clear much of anything, really. He preferred his requests vague, like a corporate oracle.
"Evan," it started, no pleasantries. "Need your expertise on a new sorting algo. We're looking at… re-evaluating resource allocation. Efficiency metrics, performance data, project impact – you know the drill. Looking for a weighted ranking system. Top quintile, bottom quintile. Standard stuff. Discretion is key on this one."
My gut clenched. "Resource allocation." "Efficiency metrics." Corporate speak for "who's getting the axe." We weren't a big department, R&D skunkworks mostly, but MicroApple had been making nervous noises about "synergies" and "streamlining" for months. The AI division, my old stomping ground before they reassigned me here, was booming. We were… less so.
I stared at the blinking cursor in my reply draft. This felt wrong. Henderson knew I was good with data, could make an algorithm sing and dance and tell you your future. But this kind of "future" wasn't one I wanted to code into existence.
The next day, I cornered him by the artisanal coffee machine – the one that always seemed to be out of oat milk when I wanted it. "About that ranking algo, Henderson," I began, trying to sound casual. "What's the exact scope? Which dataset are we pulling from?"
Henderson swirled his ethically sourced whatever-it-was in its compostable cup. "Broad scope, Evan. Company-wide, eventually. But we're starting with a… controlled sample." He gave me a look, that patented Henderson "I'm-being-transparent-while-telling-you-nothing" gaze. "Focus on the logic. Make it clean. Make it defensible."
"Defensible," I repeated. Like a layoff list needed to be legally bulletproof. My unease grew. For the next two weeks, I built it. Lines of Python felt like digital Judas kisses. I made it elegant, efficient – the kind of code that would get a nod in a peer review. And every night, a sick feeling settled in my stomach.
One evening, fueled by cheap takeout noodles and paranoia, I did what any self-respecting, slightly terrified engineer would do. I duplicated the core logic, pointed it at our department's personnel data – anonymized, of course, but I knew the patterns. I added my own employee ID to the test set. I hit 'run.'
The output was a simple, sorted list. My ID number blinked back at me from the bottom fifteen percent. The "strategically decoupled" zone.
The air left my lungs. There it was. My own code, calmly predicting my professional demise. I should have deleted it, told Henderson I couldn't crack it, feigned incompetence. But there was that nagging thing – professionalism? A misplaced sense of duty? Or maybe just the dumb hope that I was wrong, that this was just a theoretical exercise.
I scheduled a follow-up with Henderson. "The algorithm is complete," I told him, trying to keep my voice even as I showed him the simulated output (minus my own damning presence on the list, of course). "It identifies underperforming assets based on the criteria."
He nodded slowly, eyes on the screen. "Good work, Evan. Clean." "Henderson," I had to ask. "This isn't for our department, right? I mean, if this was an internal R&D review, you wouldn't have me building the tool that… well, you know." It was a weak probe, but I needed something.
He finally looked up, a slight frown. "Evan, this initiative is way above our pay grade. Corporate strategy. They task, we execute. This resource model will never be deployed by our department. We're R&D, not HR hatchet men." He clapped me on the shoulder, a gesture that felt more like a shove. "You're overthinking it. Good job. Submit the final commit. They're waiting on it."
Never deployed BY our department. The weasel words. He didn't say it wouldn't be used on our department. My stomach twisted. He knew. He had to know. And he was letting me do his dirty work, or at least build the weapon.
I hesitated for a full minute in front of my monitor that night, the "SUBMIT" button glowing like a malevolent eye. Every instinct screamed don't do it. Change a variable. Introduce a subtle bug in the weighting for anyone with a 'V' in their name. Something. Anything.
But I didn't. I clicked submit. I committed the code. Maybe it was cowardice. Maybe it was that stupid, ingrained obedience of an employee who'd been told he was overthinking.
One month later, the email from Chapter One landed. "Strategic decoupling." My own damn algorithm had, with cold, impartial logic, signed my professional death warrant.
The regret was a physical thing, a constant bitter taste. Not just for the job, but for that moment of weakness, that betrayal of my own gut feeling. I'd had the power, the knowledge. I could have run a different commit.
The memory, sharp and agonizing, swirled like a corrupted data packet. The biting wind of Forkworld was suddenly back on my face, the impossible cityscape of myriad Evans burning my retinas. I was standing on that obsidian balcony again, the weight of Atlas Walker's earlier pronouncements pressing down.
Then, a flicker. Beside me, where there had been empty air, another Evan solidified. This one was… different. Not as polished as Atlas, not as overtly hostile. He wore simpler clothes, almost utilitarian, and his eyes – my eyes – held a keen, almost unnervingly calm intelligence. There was a subtle difference in his stance, a coiled readiness. He wasn't Level 0. His HUD, fainter than Atlas's but present, read: [Evan_AlteredAlgoCommit – Level 1 System Analyst – Rogue Variant].
He turned his head, meeting my gaze directly. No surprise, no shock. Just a quiet assessment.
"You hesitated," he said, his voice an echo of mine, yet with an edge of certainty I envied. "Before you submitted the ranking protocol. You knew."
I could only nod, speechless. How did he—
"You regretted not changing it," he continued, a statement, not a question. "Every day since. That particular regret… it resonates strongly here." He gestured vaguely at the city around us, a city built of such echoes. "I am the commit you didn't run, so if I am Evan Walker altered_algo, bet i can call you Evan Walker Main. The one where the algorithm developed a… very specific, localized 'rounding error' that kept one particular R&D skunkworks engineer off the list."
He took a step closer. "That choice saved my career in my branch. " He gave a wry, almost bitter smile. "Welcome to the consequences, brother. What will you do with your second chance to choose?"
Level 0.