The Unbound Algorithm
The folding table still wobbled. Mira had fixed it twice—once with a folded napkin, once with a piece of cardboard torn from a box of remaindered novels—but the stubborn leg refused to cooperate. She gave up and let the laptop balance on her thighs instead.
“You’re going to give yourself a back problem,” Leo said from the doorway.
“I already have three back problems. They’re named Julian, the rent, and your coffee.”
He handed her a fresh mug. This one didn’t slosh. Progress.
The back room had transformed over the past three weeks. The boxes of unsold literary fiction had been stacked into makeshift walls, creating a narrow corridor that led to her workstation: a salvaged oak desk, a second-hand monitor, and a tangle of cables that looked like a modern art installation titled “Desperation.” The space smelled of old paper, fresh solder, and the faint metallic tang of Leo’s espresso machine from the front of the store.
“The crowdfunding hit forty-two thousand last night,” she said, not looking up from the code. “That’s enough to cover the first year of the co-op lease, but we need sixty-eight to buy the building outright.”
“Forty-two thousand dollars from people who hate algorithms.” Leo leaned against the doorframe, arms crossed. “That’s almost poetic.”
“It’s almost enough. We’ve got three days left on the campaign. The spike from the viral video is plateauing.”
She scrolled through the donation log. Names she recognized from the Lonely Hearts Club. A few from the tech community who’d reached out after her Gala speech. One anonymous donor who’d contributed five thousand dollars with the note: “For the bookstore that taught me to read.”
“Who’s that?” Leo asked, pointing at the screen.
“No idea. Could be anyone. Could be your ex-girlfriend trying to buy forgiveness.”
“I don’t have an ex-girlfriend. I have a series of failed experiments in human connection.”
“Same thing, different terminology.”
He laughed, and the sound filled the cramped space like warm light. Six months ago, that laugh would have made her flinch. Now it settled into her bones like a familiar ache.
“How’s the app?” he asked.
“It’s not an app yet. It’s a collection of functions that occasionally talk to each other. Think of it as a cocktail party where everyone is drunk and no one remembers anyone’s name.”
She turned the laptop toward him. The screen showed a mess of Python scripts, API calls, and commented-out code blocks. At the center, a single function called “unbound_match” that she’d rewritten seventeen times.
“Show me the pretty part,” he said.
She clicked a button. The screen went black, then bloomed into a soft gradient of amber and indigo. A single line of text appeared: “Meet at the park bench at 3 PM. Bring a book to swap.”
“That’s it?” Leo leaned closer. “That’s the whole interface?”
“That’s the whole interface. No profiles. No swiping. No compatibility scores. The algorithm finds two people who share a book, a time, and a location. It sends them a single message. After that, the app deletes itself from their phones.”
“Deletes itself?”
“Temporary connection. No data retention. No behavioral tracking. No ‘you might also like’ suggestions. The app is a matchmaker that leaves the room after the introduction.”
Leo was quiet for a long moment. Then he said, “That’s the most romantic thing I’ve ever seen.”
“It’s just code.”
“It’s code that respects people enough to get out of their way.”
She felt the heat rise to her cheeks. Stupid. She was twenty-eight years old, a former data analyst who’d built algorithms that predicted stock markets and dating preferences with equal precision. She should not be blushing because a bookstore owner called her code romantic.
“The beta test is next week,” she said, steering the conversation back to safe ground. “I’ve got twenty volunteers from the Lonely Hearts Club. They’ll each get one match. We’ll see if anyone actually shows up.”
“They’ll show up.”
“You don’t know that. People are flaky. Life happens. A bus breaks down, a cat gets sick, a sudden existential crisis about the meaning of connection—”
“They’ll show up,” he repeated. “Because you built something that treats them like humans, not data points.”
She wanted to argue. That was her default setting—find the flaw, the edge case, the variable that would break the system. But Leo was looking at her with that infuriating certainty, the same look he’d worn when he’d handed her the key to the back room and said, “Build your Unbound app. For real. I’ll handle the books.”
“Fine,” she said. “They’ll show up. But if they don’t, I’m blaming your coffee.”
“My coffee is excellent.”
“Your coffee is a war crime in a ceramic mug.”
He grinned and disappeared back into the main store. She heard the familiar sounds of the afternoon rush: the creak of the front door, the rustle of paper bags, the low murmur of customers asking about staff picks. Leo’s voice, warm and patient, recommending a novel about a woman who time-travels through her own memories.
Mira turned back to the code. The unbound_match function stared at her, waiting for her to solve its final problem.
The algorithm was simple in theory. It took two inputs: a user’s book preference and their available time slot. It cross-referenced those against other users in the same geographic area. When it found a match, it generated a location—always a public space within walking distance of both users—and a time. Then it sent the message and deleted the data.
The problem was the location generator. It kept suggesting places that were too generic—a Starbucks, a bus stop, the corner of a parking lot. The whole point of Unbound was serendipity, not convenience. She needed the algorithm to find places that had character, that invited conversation, that made the meeting feel like an event rather than an errand.
She’d tried pulling data from Yelp. Too commercial. Google Maps. Too sterile. OpenStreetMap. Too chaotic.
Then she remembered Leo’s vinyl listening party. The way he’d described the venue: “It’s a laundromat that turns into a listening room after midnight. The owner’s grandmother used to run it. The dryers hum in B-flat.”
That wasn’t data. That was story.
She opened a new file and started coding. Instead of scraping reviews or ratings, she wrote a scraper that pulled from local blogs, community newsletters, and handwritten notes posted on bulletin boards. She taught the algorithm to look for words like “hidden,” “secret,” “the owner says,” and “you have to know someone who knows someone.”
It was inefficient. It was messy. It was beautiful.
By the time she looked up, the light through the back window had shifted from gold to gray. The store was quiet. Leo stood in the doorway again, holding two plates of something that smelled like garlic and butter.
“You missed dinner,” he said.
“I missed dinner three hours ago. This is a late-night snack that’s pretending to be dinner.”
“Semantics.” He set the plates on the corner of her desk. “Eat. You’ve been staring at that screen for eight hours.”
“I’ve been building something.”
“You’ve been avoiding something.”
She opened her mouth to deny it, then closed it. He was right. She’d been avoiding the question that had been circling her mind for weeks, the one she couldn’t solve with code or logic or late-night espresso.
“What happens after the beta?” she asked.
Leo sat on a stack of boxes across from her. “We see if it works.”
“And if it does? If people actually connect? If the algorithm proves that serendipity can be engineered?”
“Then we have a successful app.”
“And then Julian offers to buy it again. Or builds a copycat. Or finds some other way to turn it into a data farm.”
Leo picked up his fork and twirled it between his fingers. “You’re worried he’ll win.”
“I’m worried that everything we’ve built will be swallowed by the same system we’re trying to escape.”
“Then we make sure it can’t be swallowed.”
“How?”
He set down the fork and looked at her with that steady, unblinking gaze that made her feel like she was the only person in the room. “You already open-sourced the code. Anyone can use it. Anyone can modify it. Julian can build a copycat, but he can’t own the idea. The idea belongs to everyone.”
“That’s not a business model.”
“It’s not supposed to be a business. It’s supposed to be a community.”
She wanted to argue. She wanted to point out that communities needed funding, that servers cost money, that the Lonely Hearts Club couldn’t sustain itself on goodwill alone. But Leo was looking at her with that infuriating certainty again, and she realized he wasn’t offering a solution. He was offering a choice.
She could keep trying to beat the system at its own game. Or she could build something the system couldn’t touch.
“Eat your pasta,” he said. “It’s getting cold.”
She picked up the fork. The pasta was overcooked, the sauce too salty, the garlic burned at the edges. It was the best meal she’d had in months.
“Leo?”
“Yeah?”
“Thank you. For the key. For the space. For not asking me to explain myself every five minutes.”
He smiled, and the lines around his eyes crinkled in a way that made her chest ache. “You’re welcome. For the key. For the space. For the terrible coffee.”
“The coffee is still a war crime.”
“I know. I’m working on it.”
They ate in comfortable silence. Outside, the fog rolled in from the bay, muffling the city into a soft gray hush. The bookstore creaked and settled around them, a living thing with its own rhythms and secrets.
Mira finished her pasta and set the plate aside. She opened the laptop again. The unbound_match function was still waiting, but now she knew what to do.
She deleted the location generator. All of it. The Yelp data, the Google Maps API, the scraper for local blogs. She replaced it with a single line of code: “Ask the user where they want to meet.”
The algorithm would suggest a time and a book. The location would be up to them. The app would facilitate the connection, but the serendipity would belong to the people who showed up.
She saved the file and closed the laptop.
“Done?” Leo asked.
“For now.”
“Good. Come see something.”
He led her through the dark bookstore, past the shelves of poetry and the display of local authors, to the front window. The street was empty. The fog had turned the streetlights into soft halos of amber light.
“Look,” he said.
She looked. At first, she didn’t see anything unusual. Just the usual Mission District nightscape: a few parked cars, a closed taqueria, a cat slinking along the gutter.
Then she saw it. Taped to the inside of the window, facing out, was a piece of paper. Handwritten in Leo’s messy scrawl: “This bookstore is not for sale. This bookstore is for you.”
“When did you put that up?” she asked.
“This morning. Before the rush.”
“You’re declaring war on NovaTech with a piece of paper and a roll of tape.”
“I’m declaring that some things can’t be optimized.”
She looked at the note, then at the reflection of their faces in the glass. His was calm, certain. Hers was tired, hopeful, and a little bit scared.
“What happens tomorrow?” she asked.
“The same thing that happens every day. We open the store. We make coffee. We recommend books. We wait for the world to catch up.”
“And if it doesn’t?”
He turned to face her. The fog had condensed on the glass, blurring the streetlights into watercolor smudges. “Then we build a better world in the space between the shelves.”
She wanted to say something witty, something that would break the tension and return them to their familiar rhythm of bickering and banter. But the words wouldn’t come. Instead, she reached out and took his hand.
His fingers were warm, calloused from years of shelving books and turning pages. They fit against hers like they’d been designed for it.
“The algorithm would have predicted this,” she said.
“The algorithm would have predicted a 99.9% match. It would have missed the part where we argue about dish soap and the correct way to shelve poetry.”
“That’s the important part.”
“I know.” He squeezed her hand. “That’s why I’m keeping you.”
The fog thickened. The streetlights blurred. The bookstore stood silent and stubborn, a fortress of paper and ink in a city of glass and code.
And for the first time in months, Mira didn’t feel like she was solving an equation. She felt like she was living in the answer.
“You’re going to give yourself a back problem,” Leo said from the doorway.
“I already have three back problems. They’re named Julian, the rent, and your coffee.”
He handed her a fresh mug. This one didn’t slosh. Progress.
The back room had transformed over the past three weeks. The boxes of unsold literary fiction had been stacked into makeshift walls, creating a narrow corridor that led to her workstation: a salvaged oak desk, a second-hand monitor, and a tangle of cables that looked like a modern art installation titled “Desperation.” The space smelled of old paper, fresh solder, and the faint metallic tang of Leo’s espresso machine from the front of the store.
“The crowdfunding hit forty-two thousand last night,” she said, not looking up from the code. “That’s enough to cover the first year of the co-op lease, but we need sixty-eight to buy the building outright.”
“Forty-two thousand dollars from people who hate algorithms.” Leo leaned against the doorframe, arms crossed. “That’s almost poetic.”
“It’s almost enough. We’ve got three days left on the campaign. The spike from the viral video is plateauing.”
She scrolled through the donation log. Names she recognized from the Lonely Hearts Club. A few from the tech community who’d reached out after her Gala speech. One anonymous donor who’d contributed five thousand dollars with the note: “For the bookstore that taught me to read.”
“Who’s that?” Leo asked, pointing at the screen.
“No idea. Could be anyone. Could be your ex-girlfriend trying to buy forgiveness.”
“I don’t have an ex-girlfriend. I have a series of failed experiments in human connection.”
“Same thing, different terminology.”
He laughed, and the sound filled the cramped space like warm light. Six months ago, that laugh would have made her flinch. Now it settled into her bones like a familiar ache.
“How’s the app?” he asked.
“It’s not an app yet. It’s a collection of functions that occasionally talk to each other. Think of it as a cocktail party where everyone is drunk and no one remembers anyone’s name.”
She turned the laptop toward him. The screen showed a mess of Python scripts, API calls, and commented-out code blocks. At the center, a single function called “unbound_match” that she’d rewritten seventeen times.
“Show me the pretty part,” he said.
She clicked a button. The screen went black, then bloomed into a soft gradient of amber and indigo. A single line of text appeared: “Meet at the park bench at 3 PM. Bring a book to swap.”
“That’s it?” Leo leaned closer. “That’s the whole interface?”
“That’s the whole interface. No profiles. No swiping. No compatibility scores. The algorithm finds two people who share a book, a time, and a location. It sends them a single message. After that, the app deletes itself from their phones.”
“Deletes itself?”
“Temporary connection. No data retention. No behavioral tracking. No ‘you might also like’ suggestions. The app is a matchmaker that leaves the room after the introduction.”
Leo was quiet for a long moment. Then he said, “That’s the most romantic thing I’ve ever seen.”
“It’s just code.”
“It’s code that respects people enough to get out of their way.”
She felt the heat rise to her cheeks. Stupid. She was twenty-eight years old, a former data analyst who’d built algorithms that predicted stock markets and dating preferences with equal precision. She should not be blushing because a bookstore owner called her code romantic.
“The beta test is next week,” she said, steering the conversation back to safe ground. “I’ve got twenty volunteers from the Lonely Hearts Club. They’ll each get one match. We’ll see if anyone actually shows up.”
“They’ll show up.”
“You don’t know that. People are flaky. Life happens. A bus breaks down, a cat gets sick, a sudden existential crisis about the meaning of connection—”
“They’ll show up,” he repeated. “Because you built something that treats them like humans, not data points.”
She wanted to argue. That was her default setting—find the flaw, the edge case, the variable that would break the system. But Leo was looking at her with that infuriating certainty, the same look he’d worn when he’d handed her the key to the back room and said, “Build your Unbound app. For real. I’ll handle the books.”
“Fine,” she said. “They’ll show up. But if they don’t, I’m blaming your coffee.”
“My coffee is excellent.”
“Your coffee is a war crime in a ceramic mug.”
He grinned and disappeared back into the main store. She heard the familiar sounds of the afternoon rush: the creak of the front door, the rustle of paper bags, the low murmur of customers asking about staff picks. Leo’s voice, warm and patient, recommending a novel about a woman who time-travels through her own memories.
Mira turned back to the code. The unbound_match function stared at her, waiting for her to solve its final problem.
The algorithm was simple in theory. It took two inputs: a user’s book preference and their available time slot. It cross-referenced those against other users in the same geographic area. When it found a match, it generated a location—always a public space within walking distance of both users—and a time. Then it sent the message and deleted the data.
The problem was the location generator. It kept suggesting places that were too generic—a Starbucks, a bus stop, the corner of a parking lot. The whole point of Unbound was serendipity, not convenience. She needed the algorithm to find places that had character, that invited conversation, that made the meeting feel like an event rather than an errand.
She’d tried pulling data from Yelp. Too commercial. Google Maps. Too sterile. OpenStreetMap. Too chaotic.
Then she remembered Leo’s vinyl listening party. The way he’d described the venue: “It’s a laundromat that turns into a listening room after midnight. The owner’s grandmother used to run it. The dryers hum in B-flat.”
That wasn’t data. That was story.
She opened a new file and started coding. Instead of scraping reviews or ratings, she wrote a scraper that pulled from local blogs, community newsletters, and handwritten notes posted on bulletin boards. She taught the algorithm to look for words like “hidden,” “secret,” “the owner says,” and “you have to know someone who knows someone.”
It was inefficient. It was messy. It was beautiful.
By the time she looked up, the light through the back window had shifted from gold to gray. The store was quiet. Leo stood in the doorway again, holding two plates of something that smelled like garlic and butter.
“You missed dinner,” he said.
“I missed dinner three hours ago. This is a late-night snack that’s pretending to be dinner.”
“Semantics.” He set the plates on the corner of her desk. “Eat. You’ve been staring at that screen for eight hours.”
“I’ve been building something.”
“You’ve been avoiding something.”
She opened her mouth to deny it, then closed it. He was right. She’d been avoiding the question that had been circling her mind for weeks, the one she couldn’t solve with code or logic or late-night espresso.
“What happens after the beta?” she asked.
Leo sat on a stack of boxes across from her. “We see if it works.”
“And if it does? If people actually connect? If the algorithm proves that serendipity can be engineered?”
“Then we have a successful app.”
“And then Julian offers to buy it again. Or builds a copycat. Or finds some other way to turn it into a data farm.”
Leo picked up his fork and twirled it between his fingers. “You’re worried he’ll win.”
“I’m worried that everything we’ve built will be swallowed by the same system we’re trying to escape.”
“Then we make sure it can’t be swallowed.”
“How?”
He set down the fork and looked at her with that steady, unblinking gaze that made her feel like she was the only person in the room. “You already open-sourced the code. Anyone can use it. Anyone can modify it. Julian can build a copycat, but he can’t own the idea. The idea belongs to everyone.”
“That’s not a business model.”
“It’s not supposed to be a business. It’s supposed to be a community.”
She wanted to argue. She wanted to point out that communities needed funding, that servers cost money, that the Lonely Hearts Club couldn’t sustain itself on goodwill alone. But Leo was looking at her with that infuriating certainty again, and she realized he wasn’t offering a solution. He was offering a choice.
She could keep trying to beat the system at its own game. Or she could build something the system couldn’t touch.
“Eat your pasta,” he said. “It’s getting cold.”
She picked up the fork. The pasta was overcooked, the sauce too salty, the garlic burned at the edges. It was the best meal she’d had in months.
“Leo?”
“Yeah?”
“Thank you. For the key. For the space. For not asking me to explain myself every five minutes.”
He smiled, and the lines around his eyes crinkled in a way that made her chest ache. “You’re welcome. For the key. For the space. For the terrible coffee.”
“The coffee is still a war crime.”
“I know. I’m working on it.”
They ate in comfortable silence. Outside, the fog rolled in from the bay, muffling the city into a soft gray hush. The bookstore creaked and settled around them, a living thing with its own rhythms and secrets.
Mira finished her pasta and set the plate aside. She opened the laptop again. The unbound_match function was still waiting, but now she knew what to do.
She deleted the location generator. All of it. The Yelp data, the Google Maps API, the scraper for local blogs. She replaced it with a single line of code: “Ask the user where they want to meet.”
The algorithm would suggest a time and a book. The location would be up to them. The app would facilitate the connection, but the serendipity would belong to the people who showed up.
She saved the file and closed the laptop.
“Done?” Leo asked.
“For now.”
“Good. Come see something.”
He led her through the dark bookstore, past the shelves of poetry and the display of local authors, to the front window. The street was empty. The fog had turned the streetlights into soft halos of amber light.
“Look,” he said.
She looked. At first, she didn’t see anything unusual. Just the usual Mission District nightscape: a few parked cars, a closed taqueria, a cat slinking along the gutter.
Then she saw it. Taped to the inside of the window, facing out, was a piece of paper. Handwritten in Leo’s messy scrawl: “This bookstore is not for sale. This bookstore is for you.”
“When did you put that up?” she asked.
“This morning. Before the rush.”
“You’re declaring war on NovaTech with a piece of paper and a roll of tape.”
“I’m declaring that some things can’t be optimized.”
She looked at the note, then at the reflection of their faces in the glass. His was calm, certain. Hers was tired, hopeful, and a little bit scared.
“What happens tomorrow?” she asked.
“The same thing that happens every day. We open the store. We make coffee. We recommend books. We wait for the world to catch up.”
“And if it doesn’t?”
He turned to face her. The fog had condensed on the glass, blurring the streetlights into watercolor smudges. “Then we build a better world in the space between the shelves.”
She wanted to say something witty, something that would break the tension and return them to their familiar rhythm of bickering and banter. But the words wouldn’t come. Instead, she reached out and took his hand.
His fingers were warm, calloused from years of shelving books and turning pages. They fit against hers like they’d been designed for it.
“The algorithm would have predicted this,” she said.
“The algorithm would have predicted a 99.9% match. It would have missed the part where we argue about dish soap and the correct way to shelve poetry.”
“That’s the important part.”
“I know.” He squeezed her hand. “That’s why I’m keeping you.”
The fog thickened. The streetlights blurred. The bookstore stood silent and stubborn, a fortress of paper and ink in a city of glass and code.
And for the first time in months, Mira didn’t feel like she was solving an equation. She felt like she was living in the answer.