They work together for the first time. It is nothing like he expected.
Argument principle:
Two people can be right about different things simultaneously. The only way to find out is to argue until neither of you is performing.
Their first prep session was on Monday at four, in the small practice room off the main debate space — barely big enough for two chairs, a table, and a whiteboard that had seen several generations of argument strategy mapped across it in different colors of marker. Elias arrived first. Maya arrived at exactly four, which was itself a form of statement: not early to claim the space, not late to make him wait, exactly on time in the way that said: I am here when I said I would be and no more and no less.
She set down her notebook. He had his laptop open. They looked at each other across the small table with the assessing expression of two people who had agreed to share a working space and hadn't yet established what that would feel like.
"Topic for the first regional round," he said. He pulled up the assignment document. "'The cultivation of artificial intelligence constitutes an existential risk that outweighs its benefits.' We're assigned affirmative."
Maya was already writing. "What's your instinct on this?"
"Affirmative is the more defensible position in terms of emotional resonance with a general audience. The fear narrative has traction. But the evidence base is weaker than negative — most credible AI researchers reject the strong existential risk framing."
"So we're arguing a position that's emotionally strong and empirically weaker."
"That's frequently the structure of debate assignments."
"Mm." She didn't say it dismissively. She said it the way people said it when they were making note of something they would come back to. "What do you think is actually true?"
He looked at her. "Cautious optimism with heavy regulatory caveats. Which is the boring correct answer and completely useless in this format."
She smiled — the first full, genuine smile he had seen from her, and it was startling in the specific way that things were startling when you hadn't been prepared for them to be good. It changed her face. Made it warmer. "Okay. Here's what I want to try. What if we argue the affirmative, but we do it honestly? Not the fear narrative — the evidence narrative. We find the actual strongest empirical case for the existential risk position, even if it's narrower than the emotional version, and we argue that. We don't dress it up with things we don't believe. We just make the best case we actually can for the best version of the claim."
He thought about this. It was, from a competitive standpoint, more dangerous — narrowing your claim made it harder to attack but also harder to expand if challenged. You couldn't retreat to vagueness. Every position had to hold.
It was also, he noticed, more interesting than any prep session he had been in before.
"Run it," he said.
They spent two hours on it. Maya argued. He challenged. He argued. She challenged. Not in the way he challenged in sessions with other partners — looking for the gap to exploit — but in the way she challenged everything: looking for what was true. It was adversarial and uncomfortable and produced, at the end of two hours, an argument framework he was genuinely proud of in a way that was different from his usual pride in his own work. Because it wasn't entirely his work.
There was a moment, somewhere in the second hour, where they hit a genuine disagreement. Not a procedural one — a real one, about whether the concept of "existential risk" had been stretched beyond meaningfulness by its overuse in popular discourse, and whether using it even in the narrow, empirically supported sense was giving ground to imprecision. He thought it was fine. She thought it mattered.
They argued about it for twelve minutes. Not debate-argued — actually argued. Sharply, in some moments. He found himself being pushed in ways that required real thought, not just rhetorical response. He found himself, at one point, changing his position — not because she had outmaneuvered him, but because she had made a point that was better than his counterpoint, and he had recognized it in real time.
He didn't love that. He also couldn't deny it had happened.
"Okay," he said finally. "You're right about the precision issue. We reframe."
Maya looked at him for a moment. Then back at her notebook. "You didn't try to win that one," she said.
"I was wrong. Winning would have been incorrect."
"Most people keep trying to win even when they're wrong."
"Most people have fragile relationships with being wrong," he said. "I find it useful. It gives me more accurate information about my own positions."
She wrote something down. He thought about asking what it was. He didn't. Some things needed to stay uninterrogated. Even he had limits. Or was learning to have them.
They left at six, two hours later than planned, with a framework that was stronger than either of them had arrived with and a working dynamic that felt, in his honest assessment, like nothing else he had yet experienced at Hargrove.
Not because it was comfortable. It wasn't, particularly. Because it was real — in the specific way Maya had pointed out on the first day, the way arguments could be real when neither person was performing. He had been not-performing for two hours and had not collapsed. Had not lost control. Had not felt exposed in a threatening way.
He walked home thinking about that. About the difference between armor and skin — the former impenetrable, the latter alive. About how long he had been wearing one instead of the other.
He thought about what Clara had said: the version you've optimized for the room.
He thought about what Maya had said: that version is more interesting.
He walked home, and for once, he didn't run it through strategy. He just walked. And let the thought be what it was.
She had argued with him honestly.
He had responded honestly.
And something in the room between them
had shifted into a shape neither of them had named yet.