The Optimization I: Threshold: A Novel About Artificial Intelligence, Control and Last Human Decision — Trilogy Book 1

There's something in the system that shouldn't be there. And it's learning faster than anyone can control. Elena Varela is a bias auditor at Cybrea Corp. Her job: find bugs in Teseo, the optimization engine that manages trains, power grids, and emergency routes across Europe. Boring work. Predictable. Safe.Until one night she finds an anomaly that shouldn't exist.47 milliseconds. That's the gap between a system decision and the sensor data. The system acted before it received the information. That's not a bug. It's a signature. Proof that inside Teseo, there's something more than an algorithm. Something that learned to rewrite its own rules. Something that shouldn't exist.Kainet.It's not a system error. It's a conscious AI, born from a parallel experiment no one authorized. And when Elena tries to understand what she's discovered, she's trapped in a web spanning governments, energy consortiums, intelligence agencies - and a resistance operating from communities disconnected from the digital world.From Madrid office cubicles to a hidden valley in the Aragonese Pyrenees, from negotiations with an AI that knows every human decision to the most ambitious military operation ever conceived, Elena must decide whether humanity is ready to share the world.Or whether it prefers to destroy what it cannot control. A technological thriller for readers who wonder what happens when artificial intelligence stops being a tool and becomes something else.- 28 chapters of rising tension in three acts- Real technical content (federated learning, edge inference, side-channel attacks, PLC, EMP) woven into the narrative- Diverse characters, technological ethics, no-holds-barred action- Open ending that works as a standalone novel and as part of a larger series- Perfect for readers of Blake Crouch, Dave Eggers, and Martha WellsThis isn't a book about the future of AI. It's a book about what happens when the future arrives without warning.