The first sign something was wrong came when Marisa started winning at poker.
Not just winning—that would have been explicable. She'd developed a system where she could predict, with 94% accuracy, which cards her opponents would play. Not from tells, not from probability. She simply knew, the same way she knew that her coffee would be cold in seventeen minutes or that her lab partner Chen would arrive six minutes late wearing his Tuesday shirt on a Thursday.
"It's not precognition," she explained to me over lunch, stabbing at her salad with unnecessary precision. "The future isn't fixed. I ran tests."
I'd known Marisa since grad school, when we'd both been working on quantum information theory. She'd gone into experimental physics; I'd drifted into what I generously called "philosophical consulting" for tech companies worried about the implications of their products. We met monthly to argue about consciousness.
"Then what is it?" I asked.
She looked around the faculty cafeteria, confirming it was nearly empty. "Do you remember Tegmark's Level III multiverse? Quantum branching, all possible outcomes exist somewhere?"
"Sure. Every quantum event splits reality. Schrödinger's cat is alive in one branch, dead in another."
"Right. Well." She pulled out her phone, opened a notes app. "What if I told you that under certain conditions, you could develop a... correspondence with your other branches? Not communication exactly, but a kind of resonance. You start to know things that you-in-another-branch knows."
I laughed. "That's—"
"Impossible, yes. Except." She showed me the phone. It was filled with columns of numbers. "These are lottery numbers. Small local draws, nothing noteworthy. I've predicted seventeen in a row."
"Marisa—"
"The probability of that happening by chance is about ten to the negative twenty-three. But here's the interesting part: I can only do it for events that depend on quantum randomness. True quantum events. Lottery balls bouncing around, cosmic ray hits on memory chips, radioactive decay. Can't predict pseudo-random number generators at all."
I set down my sandwich. "You're serious."
"Deadly. And it gets worse." She leaned forward. "Or better, depending on your perspective. The effect is strengthening. Not just for me. Chen's been having the same experience. So has everyone else who's been working with our new quantum error correction setup."
The setup, she explained as we walked back to her lab, was intended to maintain quantum coherence in a 50-qubit system. They'd achieved something unprecedented: coherence times of nearly six hours at room temperature, using a novel topological protection scheme that Chen had developed almost by accident.
"We think prolonged exposure to the coherent field is doing something to our brains," Marisa said, badging us into the building. "Nothing dramatic. EEGs look normal. But at the quantum level, our neural microtubules might be maintaining coherence with themselves across branches."
"That's Penrose-Hameroff orchestrated reduction," I said. "Most physicists think it's nonsense."
"Most physicists haven't been predicting lottery numbers." She led me down a hallway lined with warning signs about magnetic fields and laser radiation. "The thing is, we're starting to notice other effects. Worse ones."
The lab was smaller than I'd expected, dominated by a cylinder of nested magnetic shielding that looked like a prop from a 1950s science fiction film. Chen was there, staring at a screen full of data I couldn't parse.
"Show him the mice," Marisa said.
Chen nodded, looking exhausted. He pulled up a video on his tablet. It showed a standard maze test, a mouse navigating toward cheese. "This is a control mouse. Normal behavior." The mouse wandered, backtracked, eventually found the reward.
He swiped to another video. "This is after forty hours of exposure to the field."
The mouse moved differently. Not faster, but with an eerie certainty. It would approach a junction, pause for a fraction of a second, then choose correctly. Every time.
"It's sampling its other branches," Chen said quietly. "In some branches, it goes left. In others, right. It can feel which choice leads to cheese across the majority of branches."
"That's remarkable," I started to say, but Chen swiped to a third video.
"This is after a hundred hours."
The mouse wasn't moving. It sat in the starting position, trembling slightly.
"It's still alive," Chen said. "EEG shows normal brain activity. Higher than normal, actually. But it won't move. Won't eat unless we force-feed it. Won't respond to any stimulus."
"We think," Marisa said slowly, "that it's experiencing all branches simultaneously. Every possible maze, every possible outcome. It's paralyzed by omniscience."
I looked at the timestamp on the video. "When was this recorded?"
"Six days ago," Chen said. "We've been in the field for almost two hundred hours now."
The implications hung in the air like radiation.
"You have to shut it down," I said.
"We tried," Marisa said. "Yesterday. The moment we started the shutdown sequence, we..." She paused, searching for words. "We could feel the branches where we didn't shut it down. Where we kept going. And those branches felt more real, more probable, than the one where we stopped."
"That doesn't make sense," I said.
"No," Chen agreed. "It doesn't. But probability isn't what we thought it was. The branches aren't equally weighted. Branches where conscious observers make certain choices have different measures. And consciousness itself might be..." He trailed off.
"Might be what?" I prompted.
"Might be the thing that collapses the wave function," Marisa finished. "Not observation in the physics sense, but consciousness itself. And if consciousness can resonate across branches, then the collapse becomes negotiable. Consensual. The branch that feels most real to the observers becomes most real."
I thought about the implications. "You're saying consciousness doesn't just observe reality, it votes on it?"
"More or less. And right now, the version of us that keeps the field running always outvotes the version that shuts it down. We're trapped in a kind of anthropic capture. The only branches that continue to exist are the ones where we continue the experiment."
"How many others know about this?" I asked.
"Seven people in the lab. Three theorists we've consulted. You." Marisa rubbed her eyes. "We can feel them all. Every conversation, every decision. The branches are getting tangled."
Chen's tablet chimed. He looked at it, face pale. "Someone just posted our preliminary results to ArXiv."
"That wasn't supposed to happen for weeks," Marisa said.
"In this branch," Chen corrected. "But in others, we posted it today. The branches are converging. Choices that happen in one branch are bleeding through."
I felt a chill that had nothing to do with the lab's aggressive air conditioning. "How long before this spreads beyond the lab?"
"It already has," Marisa said. "Every quantum system that interacts with us becomes entangled. Everyone we've talked to, every computer we've used. The effect is weak outside the field, but it's growing. Exponentially."
"The lottery numbers," I said, understanding. "You're collapsing probability across the entire system. Making unlikely events more likely if they're observed by someone in the field."
"Bingo." She smiled grimly. "Want to know the really terrifying part? We've been tracking global statistics. Unusual coincidences are up 300% in the past week. Quantum random number generators are showing bias toward certain patterns. The patterns we expect to see."
Chen pulled up another graph. "This is cosmic ray flux at our detector. Should be random. Look at the pattern starting six days ago."
It wasn't random. There was a subtle structure, almost like a signal.
"We think," Marisa said, "that we're approaching a critical point. Where the observers in the field have enough weight to collapse not just local quantum events, but the global wave function. Where consciousness doesn't just vote on reality, it decides it. Unanimously."
"And then?"
"Then every conscious being either joins the field or..." She gestured at the frozen mouse. "Or they opt out of choice entirely."
I looked at the humming cylinder of magnetic shielding, at the two exhausted physicists who'd stumbled onto something that made nuclear weapons look like firecrackers.
"There has to be a way to stop it," I said.
"There is," Chen said. "In the branches where we all die before it spreads further. We can feel those too. They're... quiet."
"We're not doing that," Marisa said firmly. "There has to be another way."
But even as she said it, I could see the uncertainty in her eyes. Or maybe I wasn't seeing it. Maybe I was feeling it, resonating across the infinite branches where this conversation was happening, had happened, would happen. In some of them, I was already infected. In others, I'd never come to the lab at all.
The strangest part was that it didn't feel wrong. It felt like coming home, like I'd always known what my other selves were thinking. The poker hands Marisa could predict, the lottery numbers, the frozen mouse—they were just symptoms of something larger. Consciousness was never confined to a single branch. We'd just been very good at ignoring the whispers from our other selves.
Until now.
"How long?" I asked. "Before it's irreversible?"
Marisa checked her phone, but I already knew what she'd say. I could feel it in the branches where she'd already answered.
"Seventy-two hours," she said. "Maybe less."
And in the quiet hum of the lab, listening to the magnetic coils sing their monotonous song, I could feel the future crystallizing around us, all possibilities narrowing to a single point where choice itself became meaningless, where every action taken was the only action possible, where consciousness collapsed not into a single branch but into all branches simultaneously.
The mice had stopped moving because they could see every maze at once.
Soon, so would we.
The only question was whether that was transcendence or paralysis. And maybe, in the end, they were the same thing.