Malcolm shook it. I'm trying to make sure we're both right about what matters. Brooks walked away. The room exhaled collectively without anyone meaning to. Daniel rolled his chair toward Malcolm.

Did he just give us access to the firmware source code? He did. That's Daniel caught himself. Okay. Okay.

Let's go. They went. The next 9 hours were the most technically demanding Malcolm had worked through in years. Not because the individual problems were harder than what he'd faced before, but because the pace was relentless and the stakes of each calculation compounded the one before it. The firmware source code revealed two timing dependencies that weren't documented in the specifications legacy code from a previous engine iteration that the current team had inherited and partially understood.

Untangling those dependencies required three hours of work by Daniel on the code side while Malcolm worked through the thermal model implications on paper. At one point, Priya from structural analysis brought them both sandwiches and set them on the edge of the workstation without interrupting. Malcolm ate half of his without looking at it. At 4:43 in the afternoon, Malcolm had a complete revised sequence. Not a rough draft, not a preliminary model, a full parameter set, every modified value, every timing offset, every system interaction accounted for with the thermal model, confirming that the cold spot issue was addressed by the preconditioning step and that no secondary effects remained unresolved.

He handed it to Daniel. Run it through the simulation model. Full system, no shortcuts. Daniel ran it. The simulation model took 11 minutes to process.

11 minutes during which Malcolm stood with his arms crossed looking at the screen and Daniel sat watching the progress bar with the specific tension of someone who has invested 9 hours in a result and is trying to prepare themselves for the possibility that it doesn't work. The simulation completed. Daniel read the output. He read it twice. He turned to look at Malcolm.

It's He read it twice. He turned to look at Malcolm. It's stable. All three systems. The resonance window doesn't form.

The oscillation never appears. His voice was quiet like he was handling something fragile. Malcolm, it works. For a moment, neither of them said anything. Then Daniel said louder to the room, "Run it again.

Full parameters, different starting conditions." The second simulation took 9 minutes. Stable again, Malcolm said. Vary the fuel temperature by plus and minus2° full range. Three more simulations over the next 20 minutes.

All stable. Rebecca was standing beside Malcolm now. She had been reading over his shoulder for the last hour. She straightened up and looked at the row of simulation results on the monitor. Five green stable indicators where 14 simulations had previously produced 11 red shutdowns.

Ground test. She said it wasn't a question. Ground test. Malcolm confirmed. She turned and went to find Brooks.

What happened next took less time than Malcolm expected and more time than anyone could comfortably afford. Brooks authorized the ground test at 6:00 in the evening. The hardware modification team worked through the night implementing the firmware changes, a carefully managed process that required signoff at three separate review points, each of which Malcolm sat through in a small observation room available. if questions arose not actively participating because the modification process had its own established protocols and his job was the analysis not the implementation. Richard Vale submitted a formal written objection to the emergency waiver at 9:47 that night. Malcolm learned about it from Daniel at 10:15.

He's claiming the analysis timeline is insufficient, Daniel said, leaning in the doorway of the observation room. He's filed it with the review board. They have to acknowledge it before the ground test can proceed. When is the board convening? Tomorrow morning, 7:00 a. m.

Malcolm looked at the clock on the wall. 8 hours. Then I'll be there at 7, he said. He called Ava at 10:30. She was awake.

She always was, past her bedtime with her notebook open and her light on the way. She handled situations she couldn't control by staying informed about them. The test is tomorrow, he told her. Are you ready? The model is ready.

The question is whether they'll let us run it. Why wouldn't they? Because the people who told them it was safe the first time are the ones who have to stand next to it when we prove it wasn't. A pause. That's not engineering, Ava said.

That's ego. A lot of things that look like engineering are ego with equations on top of them. He heard her exhale. How's the science project? I fixed the fin myself, she said.

Use the geometry from your fluid dynamics textbook. Which one? The Anderson. Chapter 4. He felt something move in his chest.

Good choice, Dad. Her voice shifted slightly, losing the precision she defaulted to when she was focused, becoming something more unguarded. Are you going to be okay tomorrow? Not the test. you. He considered lying.

Considered giving her the reassuring version. I don't know, he said, but I'm going to be there. That's enough, she said. That's always been enough. He put the phone down and went back to the observation room, and the night settled around the facility like a held breath.

The review board met at 7:00 a. m. in a conference room that held 14 people uncomfortably and was holding 18. Malcolm sat at the table. Vale sat at the opposite end. Brooks sat at the center. Rebecca sat beside Malcolm and said nothing, which was its own kind of statement.

Vale's objection was read into the record. It cited insufficient timeline for independent verification of the proposed control modifications. It cited Malcolm's 8-year absence from the active aerospace industry. It cited the risk of introducing a novel modification to flight certified hardware with less than two weeks to launch window. Every point was procedurally correct.

Every point was also in Malcolm's judgment. A procedurally correct construction built around a motivation that had nothing to do with engineering safety and everything to do with the fact that the instability had been growing for 6 weeks on Vale's watch. And the person who had identified its cause was the same person Vale had spent eight years making sure nobody would listen to. When the objection was finished being read, the board chair, a thin woman named Dr. Miriam Park, who had the specific exhaustion of someone who has been mediating between engineering and politics for her entire career, looked at Malcolm.

Dr. Reed, she said, do you wish to respond to the objection? Malcolm opened the folder in front of him. He had prepared for this the way he prepared for everything, not with rhetoric but with data. The objection cites three concerns, he said.

I'll address them in order. He was calm. He had been this calm before. He had learned that calm was not the absence of feeling but the management of it. And that the management of it was the difference between being heard and being dismissed.

Timeline. We have five confirmed ground simulations showing the revised sequence is stable across varying conditions. Every simulation used unfiltered raw sensor data. Every result is documented and reproducible. Timeline is not a substitute for evidence and evidence is what we have.