The Null Hypothesis

"She wasn't wrong about this. She was wrong about what it meant."The note appears in the margin of a shared document at midnight - written by Dr. Ashida Rin, a research scientist who burned down her career to tell the truth, meant for no one but herself. She doesn't know yet that Dr. Seo Yuna has read it. That Yuna has read all of it. That Yuna has been reading her for three years. They have been publicly dismantling each other's work for exactly that long.
Rin's whistleblower paper exposed systemic failures in AI safety reporting at three major corporations. Yuna, working from inside one of them, published the methodological rebuttal that contested her findings - technically rigorous, infuriatingly precise, and devastating in the way only a response from someone who understood your argument completely could be devastating. The field took sides. The internet noticed.
Neither of them has spoken directly since. Now a Brussels regulatory body has engineered a solution to the legislative deadlock their disagreement created: a ninety-day research residency in Incheon, two workstations at opposite ends of the same room, and a single synthesizing paper that both of them must co-author or the certification framework dies in committee. The mandate is binding. The timeline is fixed.
The only question is whether two researchers who have spent three years being each other's most formidable opponent can produce the most important work of their careers - and what it costs to finally be in the same room. Yuna arrives first. She creates a folder structure. She makes tea on a four-minute timer. She does not open the second drawer of her desk in Neo-Seoul before she leaves, where a doctoral photograph has been face-down for four years.
Rin arrives with a portable whiteboard, a bag organized by a logic only she can fully reconstruct, and eighteen months of an unlabeled folder she has never discussed with anyone. What follows is not a love story about two people who stop disagreeing. It's a love story about two people who discover that the disagreement was, all along, the most intimate thing they had. Over ninety days, through methodological battles that run past midnight and annotation layers that become something neither of them has language for, through a corporate pressure campaign and a security breach and a legal communication technically not a threat, through a photograph turned face-up in a drawer and three words written in a margin no one was meant to read - they build something that will reshape their field, cost one of them everything she thought she needed, and leave both of them standing at an edge they have not yet decided to cross.
The paper is submitted. The findings hold. The acknowledgments section is eleven lines long. The last two will only make sense to one person.