On the twenty-fifth day, the United Nations launched the "Collective Human Memory" project.
It was a global online platform where anyone could upload their own memories—especially those related to major historical events. An algorithm would then identify the version of events that was remembered by the largest number of people—that version would become the leading candidate for the "prime timeline".
As deeply fused mirror humans, Ethan and Mason were asked to provide "control memories". They needed to describe in detail what had happened during the summer of 1999, identifying the differences between the two worlds through as many minute details as possible.
Ethan began writing:
"June 28, 1999. My SAT scores were released. I scored 1350, ranking 4,217th in the state. My father booked a table at a seafood restaurant and invited all our relatives and friends to celebrate..."
He wrote over two thousand words, recalling every detail with as much precision as possible.
Mason wrote his own version of events:
"June 28, 1999. My SAT scores were released. I scored 1350, ranking 4,217th in the state. My father was undergoing medical tests at the hospital that day. I checked my score using a public phone in the hospital hallway..."
The two versions were identical up until June 28, after which they diverged completely. Like two branches growing from the exact same point on a tree.
Within twenty-four hours of the platform going live, it had received three billion memory uploads. The servers nearly crashed under the load. The algorithm began processing the data, searching for the "maximum consensus".
The results were shocking.
For most historical events, humanity had no "collective memory". Even for widely documented events like the 1969 moon landing or the 2001 9/11 attacks, people from different countries and cultural backgrounds remembered things differently. Some differences were minor—had Neil Armstrong said "one small step for a man" or "one small step for man"? Others were much larger—was 9/11 an inside job orchestrated by the U.S. government or not?
Even for the event with the highest consensus rate, the agreement only reached 67%. This meant that even within the "prime world", a full third of people remembered a different version of events.
"It's impossible to establish a single prime timeline," the algorithm team concluded. "Humanity has no unified past. For every event, there are countless versions remembered by different people."
"What will be the final outcome of the fusion then?" Ethan asked.
No one had an answer.
But the overlapping zone in Berlin provided a clue. There, the two worlds hadn't merged into one—instead, they'd formed a permanent "superposition district". Buildings simultaneously displayed architectural styles from two different eras, pedestrians were inhabitants of two different worlds, and time itself existed in a state of flux between 1945 and 2025.
Reality had become a probability cloud. Walking down the street, you had a 70% chance of seeing the modern version of the city and a 30% chance of seeing its wartime counterpart. The people you met had a 60% chance of being residents of 2025 and a 40% chance of being from 1945.
This wasn't fusion—it was... chaos.