Two weeks after the accident, Leon broke his seventh keyboard.
The 'E' key snapped clean off, followed by the 'R' and 'T'. He stared at the mechanical keyboard—supposedly "professional grade"—and watched three more keys pop off in quick succession as he tried to finish his sentence.
"That's it." He pushed back from his desk. "This is ridiculous."
Maya looked up from her laptop on the couch. She was working on her own research, some protein analysis that had her muttering about amino acid sequences. "Another one?"
"Another one." Leon held up the keyboard. Keys dangled from broken springs. "I'm typing faster than the hardware can handle. My hands move, the keys register, but by the time the signal processes, I'm already five words ahead."
"So slow down?"
"I tried that for the first three keyboards." He tossed the broken one into the growing pile by the door—they'd need to make a recycling run soon. "My brain works faster than my hands now. And my hands work faster than any keyboard made for humans."
Maya closed her laptop. "What are you thinking?"
"I need a direct interface. Brain to computer. Cut out the middleman."
"Brain-computer interface? BCI?" She leaned forward. "You know that tech is still pretty experimental, right? The hardware's getting there, but the software is barely functional."
"Exactly." Leon grabbed his current laptop—the keys on this one were still intact, but he could feel them getting mushy under his fingers. "Hardware exists. EEG caps, neural sensors, the whole setup. But the interpretation algorithms are garbage. They can't read brain signals accurately enough to be useful."
He pulled up research papers on BCI technology. Articles from Neuralink, academic studies, conference presentations. His perfect memory had all of this already catalogued, but seeing it on screen helped him think.
"Current best is around 40% accuracy in signal interpretation," he said, scrolling through data. "And that's in controlled lab settings. Real-world usage drops to maybe 25%. The software just can't decode what the brain is trying to say."
Maya got up and moved to stand behind his chair, reading over his shoulder. "Because brains are messy. Electrical noise, signal interference, individual variance. Every person's neural patterns are different."
"Right. But I have expertise in neuroscience now. And computer science. And signal processing. And mathematics." Leon opened a new file. "I can build better algorithms."
"How much better?"
"A lot better."
He started coding. His fingers flew across the keys—he'd have to be gentle with this keyboard too—pulling from everything he'd learned in two weeks of intensive study. Neural network architectures. Fourier transforms for signal processing. Machine learning models for pattern recognition. Bayesian inference for handling uncertainty.
The code poured out of him. Function after function, each one optimized, each one building on the last. He was designing a complete system from scratch—data input from EEG sensors, noise filtration, signal decomposition, pattern matching, output interpretation.
Maya watched silently. She'd seen him code before the accident. This was different. There was no hesitation, no debugging, no trial and error. Just pure creation.
"You're in the zone," she said quietly.
"I'm always in the zone now." His eyes never left the screen. "Everything just makes sense. I see the solution and implement it. No wasted motion."
Three hours later, he leaned back. "First version done."
Maya had gone back to the couch at some point, but now she returned. "That fast?"
"That fast." Leon scrolled through the code. Thousands of lines. Elegant. Efficient. "But there's a problem."
"What?"
"Python's too slow. JavaScript's too slow. C++ is better but still not optimized for what I need. Existing languages were designed for human programmers thinking at human speeds." He rubbed his face. "I need something faster. More efficient."
"So... what? Use assembly language?"
"No. I make my own language."
Maya blinked. "You're going to create a programming language."
"Yeah."
"Just like that."
"Just like that." He opened another new file. "It's just syntax rules, compiler design, and optimization strategies. I know the theory. I know what I need. I can build it."
"Leon, programming languages take teams of people years to develop."
"Teams of normal people." He started typing again. "I'm not normal anymore."
She sat on the edge of his desk, watching him work. "Okay. Show me."
The new language flowed out of him even faster than the BCI software. He designed the syntax to be clean and expressive. Built-in support for parallel processing. Native optimization for neural data. Memory management that eliminated garbage collection overhead. Compiler optimizations that would make the code run at near-machine speeds.
He called it Nexus. Simple. Direct. It connected his thoughts to the machine.
"There." He saved the compiler. "Done."
Maya leaned in to look at the screen. The code was clean, almost beautiful. "This actually makes sense to me. The syntax is intuitive."
"That's the point. Human-readable but machine-optimized." Leon started porting his BCI software into Nexus. "Watch."
The conversion took twenty minutes. The new code was half the length of the Python version and would run at least fifty times faster.
"Okay." Maya crossed her arms. "That's legitimately impressive. But you can't test it without hardware."
"Already ordered." Leon pulled up a confirmation email. "Basic EEG cap kit from Alibaba. Should arrive in three days."
"Three days? You ordered this when?"
"Yesterday. I knew I'd need it."
She shook her head, but she was smiling. "You're planning three steps ahead now."
"Four, actually. After I test the BCI, I'm going to need better hardware. Professional-grade sensors. Maybe custom-built. Which means money. Which means I need to monetize this."
"Monetize BCI software?"
"It's breakthrough technology, Maya. If I can hit 95% accuracy—and I think I can—every tech company on the planet will want it. Neuralink, Synchron, Paradromics, all of them. They're stuck at 40% and throwing billions at the problem."
Maya was quiet for a moment. "You're talking about selling this?"
"Eventually. After I test it. Prove it works. Then yeah, I sell it." He turned to face her. "We need money. Real money. This apartment is fine for now, but I need equipment. A proper lab space. Access to materials. Research costs money."
"How much money?"
"Millions. Maybe billions if I'm going to really understand what happened to me. What the gene locks are. What Evolyx is." He gestured at their cramped apartment, books and papers everywhere. "I can't do that here with a laptop and a broken keyboard collection."
She nodded slowly. "You're right. We do need resources." She glanced at the code on his screen. "And if this works the way you think it will..."
"It'll work."
"Then yeah. We'll have those resources."
---
Three days later, the package arrived.
Leon tore it open on the kitchen table. Basic EEG cap with sixteen sensors. Not professional-grade, but enough to test the concept. The sensors were dry-contact, which meant they wouldn't need gel. The whole setup looked like a swimming cap covered in small metal contacts.
"That's your brain-reading device?" Maya picked up the cap, examining it.
"That's the hardware, yeah." Leon was already connecting it to his laptop via USB. "The real magic is the software."
He pulled on the cap, adjusting it until the sensors sat properly against his scalp. The Nexus compiler was running on his laptop, and the BCI software loaded without errors.
"Okay. Calibration first." Leon closed his eyes, focusing on simple thoughts. The letter 'A'. Then 'B'. Then 'C'. The software was learning his brain patterns, building a baseline.
Maya stood beside him, watching the screen. Data streamed in—electrical signals from his brain, translated into waveforms, processed by his algorithms, output as interpreted commands.
"Signal quality is good," she said. "Even with basic sensors."
"My brain puts out a strong signal now. Enhanced neural activity." Leon opened his eyes. "Calibration done. Let's test accuracy."
He thought the letter 'A' without speaking, without moving. On the screen, an 'A' appeared.
'B'. The letter appeared.
'Hello'. The word appeared, letter by letter, as fast as he thought it.
'The quick brown fox jumps over the lazy dog'. The sentence materialized on screen in real-time, each word appearing the moment he thought it.
Leon checked the accuracy metric. "95.3%."
Maya leaned closer to the screen. "Industry standard is 40%. You just more than doubled it with basic hardware and a home computer."
"The algorithms work." Leon pulled off the cap. "With professional-grade sensors, I could probably hit 98%, maybe 99%."
"Leon." Maya's voice was serious. "This is worth millions."
"Billions, maybe." He saved the test data. "Every major BCI company will want this. The software is the bottleneck, and I just broke it."
"So what's the plan?"
"I refine it. Test it more. Make sure it's bulletproof. Then I reach out to companies." He looked at her. "This is going to change things for us. Big money. Real resources. The ability to actually research what I've become."
Maya sat down in the chair next to him. "You know once you sell this, people are going to ask questions. About you. How a CS student created breakthrough BCI software in two weeks."
"I'll say I've been working on it for years. That the breakthrough came from combining multiple fields of study. Which is true, technically."
"And if they want to meet you in person? See you present the work?"
Leon shrugged. "I'll wear contacts to hide the gold eyes. Dye the white streaks. Dress normally. I can pass for human if I'm careful."
"You are human."
"Enhanced human. There's a difference." He closed the laptop. "But you're right. We need to be careful. Can't let anyone know what really happened."
"Not until we understand it better ourselves."
"Exactly."
They sat in comfortable silence for a moment. Outside, the city hummed with evening traffic. Inside, their apartment was cluttered with the evidence of two weeks of intensive research and learning. Broken keyboards piled by the door. Books stacked on every surface. Papers covered with Leon's notes and calculations.
"We're really doing this," Maya said. "Selling breakthrough technology. Playing with the big companies."
"We're doing this." Leon squeezed her hand. "And once we have the money, we figure out what I am. What the gene locks are. Whether there are others like me. All of it."
"One step at a time."
"One step at a time."
He looked at the EEG cap sitting on the table, at the code running on his laptop, at the future suddenly opening up before them.
Two weeks ago, he'd been a normal CS student. Now he was something else. Something more.