"Huh, that's weird," Sam muttered under his breath. He sighed heavily, massaging his smooth scalp and staring incredulously at the computer screen, hoping his code would magically reveal its errors instead of forcing him to hunt them down line by line for the fifth time. The data matrix was definitely being passed into the neural network, but for some reason it was completely corrupted by the time the neural network finished running. He sighed again and checked his watch. 2:38 AM, he really needed a better sleep schedule. If he could just track down this bug he could still nap for a couple hours before having to prepare for his dissertation review the next morning. He had probably just missed a comma somewhere, or something else stupid. He scrolled down to his fitness algorithm for what felt like the millionth time, the letters on the screen blurring together as he fought to keep his eyes open. Goddammit, he had multiplied by one instead of negative one. He knew it was something stupid. Time to recompile.
He had spent the last seven years working on his doctorate in artificial intelligence, sacrificing sleep, money, friends, his love life, all for the dream of being able to have three letters appear after his name. Samuel Peterson, PHD. He liked the sound of that. Everything would fall into place once he finally got his thesis published. He could get a research position at the University of South Sequoia City, start paying off student debt, maybe actually start dating again ...
His code finished compiling, and to his surprise it hadn't thrown any errors for once. He opened the data output file and scrolled passed endless columns of numbers and variables representing each decision his program had made as it traversed the vast neural network Sam had meticulously crafted. Sam thought it was some pretty kick-ass code, though he was admittedly a little biased. Still though, imagine being able to drop arbitrarily large sets of data into an analytical engine that, thanks to some pretty fucking elegant data compression and multi-threading of concurrent recursive neural networks, could spit out a comprehensive meta-analysis in nothing flat. The best part was that it didn't even matter what kind of data you entered. Enter customer transaction history and it will give you an overview of their buying habits and a customized targeted advertising campaign. Give it the board state of a chess game and it won't just tell you how to win in the fewest moves, it will also analyze your play style and give you advice for future games. Hell, give it your financial info and it will do your damn taxes.
At least, it will as soon as he could figure out why the hell the data was still getting corrupted. It was his fitness algorithm, he knew it. It was the only thing that could possibly alter the input after it hit the neural network, but he had checked and rechecked the algorithm and the logic was solid. He was missing something. He leaned back in his chair and closed his eyes, trying to visualize the path of the data through the program. It was so close to giving him exactly what he wanted. Maybe he had bad data? He had set the input to a collection of Shakespearean sonnets, but instead of giving him analysis on writing patterns like his context detection algorithm was supposed to, he kept getting random strings of ten syllable sentences. He opened his sample data directory, found some cool looking MRI's taken while a patient was asked a series of test questions, and dropped them into the input file.
"Let's try this again," he muttered as he reran his code and opened the output file. "Fuck!" he yelled, slamming his keyboard on the desk. Mr. Kitteh bolted from his perch on the window, hissing at Sam before haughtily stalking out of the cramped bedroom. Sam immediately felt terrible. He rarely had emotional outbursts like that, and now his favorite coding buddy was scratching the shit out of the futon because he couldn't keep a lid on it. The lack of sleep was definitely getting to him. There had to be a logical explanation, there always was.
The MRI output didn't make any sense. The patient was asked easy questions -- favorite color, simple math, stuff like that -- and the MRI tracked which areas of the brain were active when they heard the question and when they gave an answer. But his output wasn't giving him any meta-analysis; it was just returning the exact same scans of the brain he had entered. No, wait a second. Something was off about the data time-stamps. Sam's jaw hit the floor. His program wasn't just returning the same answers that the patient was, it was predicting them! He looked at the input again to see if he was going crazy from the lack of sleep. He wasn't, the time-stamps didn't lie. The program clearly loaded the brain scans of the patient being asked a question, ran the analysis, and gave the correct answer before the patient did.
This was insane! If he was right about this, and he usually was, it would mean that he could predict what a brain was thinking before the brain thought it! He would be able to diagnose mental illness instantly, or perfectly model the brain's physiology, or predict people's thoughts, or simulate an entire human mind! He would be able to ... holy shit. Sam stared in silence at his code. What had he just discovered? Corporations would pay out the ass for this. Governments would kill for it. Sam shifted in his seat uncomfortably, wondering what exactly he had gotten himself into. A warm ball of fur sprang up into his lap. Sam scratched Mr. Kitteh absentmindedly under the chin, who promptly curled up and began purring loudly. He didn't know it at the time, but Simon Peterson had just accidentally discovered one of the most powerful and important algorithms of the century, and right now he was the only person in the world who knew it existed.
YOU ARE READING
Wonder World: Ignition
Science Fiction"In the distance sirens were blaring, but Kyle's eyes were fixed on the faceless figure staring down at him from the burning roof. 'Where are they? Where are my mother and sister?' he shouted, his voice cracking with fear. 'Come and see,' the figure...