Ethan slept only a few hours. He was not afraid of being proven wrong; he had been wrong many times during his career and learned from every one of them. He feared being right after people had already decided that the company could not afford to listen. In the darkness, he thought about the strangers who might eventually depend on Prototype Seven and the child asleep down the hall who depended on him to choose honestly.
By 8:12 on Monday morning, Prototype Seven had failed its preliminary certification sequence. The automated diagnostic reached the same problematic region Ethan had identified months earlier, and the system began responding to a disturbance in the wrong direction. A protective shutdown prevented damage, and the laboratory remained physically safe. But the evidence could not be described as a calibration inconvenience any longer.
Clare stood behind the observation glass with twelve engineers while the system returned to a quiet state. The vibration response displayed on the monitors was unmistakable, and its center lay at forty-one point six hertz. Vanessa, standing near the rear wall, suggested another reset and a modified threshold. Clare turned toward her with an expression that made the room fall silent.
"No," she said. "We are not going to hide the evidence by moving the point at which we display it." The lead engineer explained that a new configuration might be prepared within hours if the test team narrowed the operating conditions. Clare asked whether any independent reviewer could certify a permanent correction by the federal deadline.
Nobody answered. There was a time when she would have demanded that someone produce a solution by the end of the morning, because the pressure of her position made uncertainty feel like disobedience. Now she looked at the faces around her and understood that the absence of an honest answer was itself important information. She ordered the test suspended and instructed the team to preserve the raw results.
At 8:27 Richard entered the laboratory carrying the sealed founding documents. The old file identified Ethan as chief systems architect and founding technical partner, with the retained authority to withhold technical attestation when independent validation was incomplete. Clare had read the substance over the weekend, but seeing his signature beside the other original partners made the history feel newly personal. A man had worked six years beneath her office while the company forgot that its most important technical standard had been shaped by his judgment.
She telephoned the federal review liaison and explained that the prototype had failed the preliminary sequence. The official asked whether the protective shutdown had functioned correctly, and Clare confirmed that it had. He then asked whether Northstar could supply a validated redesign before the contracted review time. She said no.
The answer was costly, but it was the first one that allowed everybody in the room to know where they stood. The liaison suspended the review pending complete technical evidence and promised a new date only after independent evaluators accepted the updated plan. Clare thanked him and ended the call, aware that dozens of internal financial projections had just become uncertain. She did not ask the laboratory to make the result look better.