For 30 advance/early chapters : p atreon.com/AutumnXd
Ryan felt like a schoolteacher grading homework.
Kyle delivered the firefighting mech component reports in a stack. Ryan sat at his desk, red pen in hand (metaphorically; he used a tablet), and worked through them one by one, marking errors, noting improvements, flagging sections that needed rework.
"That's everything?" he asked, eyeing the pile.
"Everything." Kyle looked hopeful in the specific way of a student who'd already spent too long on an assignment and couldn't face doing it again.
"I'll have notes for you by tomorrow."
Kyle pressed his palms together in gratitude. "Thank you. Seriously. I owe you one."
"The cafeteria is free, Kyle. You can't owe me a meal you're not paying for."
"Fair point."
Kyle left. Ryan set the reports aside and opened the neural link documentation.
The prosthetics downgrade was the priority.
Scrapper's neural link read signals across the entire motor cortex and beyond: limb movement, trunk stabilization, balance, fine motor control, everything the human body could express through neural impulses. The system's signal recognition was functionally perfect, its decoding algorithm the most advanced on the planet.
The prosthetic version needed a fraction of that capability. A hand. A wrist. An elbow. Three joints, a limited set of motor neuron pathways, a constrained vocabulary of neural commands. Strip away everything else and you had a system that was still light-years ahead of any competing technology but narrow enough in scope that it couldn't be reverse-engineered back to the full military-grade platform.
The signal acquisition hardware was straightforward. Simplified sensor array, reduced contact points, consumer-grade housing. Ryan could spec that in his sleep.
The real challenge was the decoding algorithm.
The algorithm was the crown jewel. It translated raw neural noise into precise motor commands with sub-millisecond latency. Even the downgraded version, limited to upper-limb motor signals, would be worth billions on the open market. Every BCI lab in the world would kill for it.
Which meant it needed to be locked down.
Ryan implemented AES-128 encryption on the algorithm core. One hundred twenty-eight bits of key length. To brute-force the encryption using current computing capabilities would require approximately two hundred billion years and an electricity bill roughly equivalent to a century of global GDP. He added side-channel countermeasures on top: timing randomization, power consumption masking, electromagnetic shielding. The standard attacks (differential power analysis, electromagnetic emanation monitoring, cache-timing exploits) were all neutralized.
The only theoretical vulnerability was a quantum computing attack, and practical quantum computers existed in approximately zero laboratories on Earth.
The algorithm was safe.
Over the next several days, Ryan multitasked at a level that would have hospitalized most adults.
Morning: review Kyle's firefighting mech reports, provide corrections, return them for revision. Midday: check in on the plasma cannon data analysis (Thornton was cross-referencing the new generation test against his decade-old archives, finding encouraging alignment). Afternoon: monitor the drift experiments (the second round of dual-person tests was underway, with three-person trials scheduled next). Evening: write the downgraded neural link code and coordinate remotely with Mason's team on the prosthetic limb modifications.
By the weekend, the simplified neural connection system was complete.
Ryan called Jake to his quarters and ran a quick verification test. Jake wore the sensor cap while Ryan monitored the signal output. The system isolated hand-specific motor signals cleanly, filtering out everything else. When Jake thought about opening and closing his fist, the decoder produced the correct command sequence with zero errors.
The hardware worked. The software worked. The encryption held.
Ryan packed the system and headed to town.
-----
The Prism Sciences Neural Prosthetics Research Center had acquired a sign since his last visit.
Hand-painted letters on the metal gate: *PRISM NEURAL PROSTHETICS RESEARCH CENTER.*
It still looked like a converted textile workshop. Because it was a converted textile workshop.
Mason's team was waiting outside. Every single one of them had dark circles under their eyes deep enough to store coins in.
"Welcome back," Mason said, stifling a yawn.
"You all look terrible."
"We've been averaging about four hours of sleep. The 3D printer runs twenty-four-seven and somebody has to babysit it."
Ryan walked inside. The workshop had been transformed. Not into something pretty, but into something functional. Workbenches organized by task. Tools mounted on wall racks. A 3D printer in the corner, still humming, extruding joint components layer by layer.
And sitting in a chair near the back wall, looking slightly nervous, was a middle-aged man with a missing right arm. Beside him, a teenage boy who was trying very hard not to stare at Ryan and failing completely.
Mason made the introductions. "This is Mr. Grant, our volunteer. Lost his right arm in a construction accident three years ago. And this is his son, Danny."
The teenager's eyes widened. "I know who you are. I've seen all your videos."
Ryan smiled and shook hands with both of them, then turned to business.
"Show me the prosthetic."
Mason led him to a workbench. Lying on the surface was a right-arm prosthetic, matte black, 3D-printed from reinforced polymer with titanium joint pins. Five articulated fingers. Motorized wrist. Powered elbow. The structural design was Mason's team's work, modified over the past week to accept neural input instead of myoelectric sensors.
"The drive system upgrade isn't complete yet," Mason said. "Response times are the same as our original prototype. We need better motors and a redesigned gear train to hit the speeds you're asking for. But for today's test, the mechanical side is functional."
"That's fine. Today is about the neural interface, not the drive speed."
Ryan unpacked the system he'd brought. The sensor cap, smaller and lighter than Scrapper's full helmet but built on the same principles. The signal conversion module, a palm-sized box that translated neural commands into digital instructions the prosthetic's onboard computer could execute. Fiber-optic cables connecting the three components: cap, converter, arm.
He linked them together on the workbench. Cap to converter. Converter to prosthetic. A chain from brain to machine, simplified from Scrapper's architecture but carrying the same fundamental capability: thought becomes action.
"Power it up," Ryan said. "Let's make sure everything initializes before we put it on Mr. Grant."
Mason connected the power supply. The prosthetic's indicator lights came on. The converter module's display showed a ready state. The sensor cap's internal diagnostics reported nominal.
Everything was green.
Ryan looked at the volunteer, at his son, at the team of exhausted engineers who'd spent a week turning a myoelectric prototype into something that had never existed before.
"Let's begin."
-----