He locked the lab door behind him and crossed to the console bank. Screens woke under his touch. Status trees rolled up in pale lines. Chamber pressure, anchor phase, thermal drag, field spread, spatial correction bias. The numbers came alive in their neat boxes and immediately began giving him reasons to delay.

Not reasons to quit, he told himself. Delay.

Anchor drift sat high by a fraction. Return-envelope spread breathed wider than target, narrowing and widening with a slow, ugly pulse. One heat sink reported a lag on dissipation. The chamber humidity had climbed overnight and was throwing tiny charge ghosts along the ring. Small things. Manageable things. The kind of things Amara had once leaned over his shoulder and tapped with one hard finger.

These are the things that matter when your theory meets a room.

He stood there hearing her anyway, hearing the old fights as clearly as if she had come down the lift behind him.

Drift matters.

Pressure variance matters.

You don’t get to call them noise because the graph offends you.

Elias adjusted what he could. He recalibrated anchor bias. He bled the chamber and restarted the pressure scrubbers. He rerouted cooling load away from the slow sink and watched the heat numbers improve by a little and then stall. He ran the token pulse again, then a dead-weight displacement check with the ballast frame. The frame vanished from one sensor grid and returned to another exactly as instructed. Close enough. Better than close enough, if he were being honest and impatient in equal measure.