Skip to Content
Blockchain

Unchained: A story of love, loss, and blockchain

A science fiction story for MIT Technology Review’s blockchain issue.

Alina set the trap for the car at her father’s summerhouse.

When it was time, she pulled on her Nokia rubber boots, slung the backpack over her shoulder, and went outside. The Lapland morning was monochrome: slushy snow, dark saw-edged pines, gray lake. The fresh air was a relief. She had never been able to sleep in the mildew-smelling house, not even as a child.

She had never planned to come back. But it was perfect for the trap. A patchy network that would force the car AI to run locally. No witnesses. No IoT devices auctioning sensor data for tokens. Although, years ago, her father had stuck a yellow “You Are Being Recorded” sticker in a window as a joke. That was so perfectly like him: all surface, no substance.

Alina had prebooked the ride using a burner autochain ID. The app in her AR glasses showed the car icon crawling along the winding forest road. In a few minutes, it would reach the sharp right turn where the road met the lake. The turn was marked by a road sign she had carefully defaced the previous day, with tiny dabs of white paint. Nearly invisible to a human, they nevertheless fooled image recognition nets into classifying the sign as a tree.

Glass-thin ice cracked under her booted feet. The cold weight in her belly grew colder.

Suddenly, she had an urge to call Sini, just to hear her voice. There was still time. Tapani would be taking her to the kindergarten right now. She hadn’t seen her daughter for almost two weeks, not since she had started hunting the car.

Alina removed her woolly mittens and took out her phone. As she swiped the screen, the band of pale skin on her ring finger grinned at her. The absence of her wedblock ring banished all doubt.

She put the phone away. No matter what, the car was going down.

Alina pulled the ski mask over her face and broke into a jog. She was warm when she got to her observation spot behind a large pine near the turn. She opened the app that controlled the jammer in her backpack, held it ready in her field of vision.

The car’s headlights flickered in the deep shadows of the trees. Then its sleek low form glided into view.

Time slowed. The pine smelled wet and rich. Alina was a vitrovegan, but some ancient reflex made her mouth water. This had to be what hunters felt like when the prey approached.

The car decelerated at the turn. Alina knew how it saw the world: lidar point clouds, IoT sensor oracle feeds, chain IDs—with algorithms predicting where everything was going to be seconds, minutes, days from now. The past, the present, and the future in a glance. Her simple traffic-sign exploit felt ridiculous. The car was never going to fall for it. She had not even thought about where she would ask the car to take her if the scheme failed ...

The car missed the turn. It plowed straight into the lake with a crash of breaking ice. Alina blinked at the jammer app. The device in her backpack killed the sparse local network.

She ran toward the car. It was a Mercedes, 2020s model, meant to evoke luxury: silvery sheen, sculpted contours. The front of the car was in the black water, but the back wheels spun madly on the shore. Snow, ice, and gravel flew into the air. The engine keened like a wounded animal.

To Alina’s horror, the car started moving slowly backward.

She dropped her backpack, ran, put her shoulder against the back of the car, pushed hard. Something tore, and pain shot through her left thigh.

For a second, she slid backward in the snow. Then her boots found purchase. The car was surprisingly light; you didn’t have to armor things that never crashed. Ignoring her screaming leg, Alina pushed the car into the lake.

When it was fully in the black water, its engine died. But its headlights stayed on. It bobbed up and down in the water, floating. The lidar nub on the roof kept spinning.

It was still alive.

Alina dropped to her knees. Black dots swam in her eyes. As if through the car’s four-dimensional vision, the past unfolded before her: the hunt, the day she saw the car for the first time.


It was on Tapani’s first day at the games company, and Sini was making them late.

“I—don’t—want—to—put—on—rubber—boots!”

She banged her heels on the floor. Alina sighed. She had gotten up early to pack the toys, the pram, and the spare clothes for the day in the city. Tapani had slept through it all. Now he stood in the hallway in an AR daze, a blank look on his bearded face. He looked owlish in his thick-rimmed glasses, black suit, and thick overcoat. Sini’s Moomin backpack dangled limply from his hand.

“Sini, Mommy is going to count to three. One,” Alina said, more sharply than she intended.

“No!” Sini screamed. A tendril of snot leaked from her nose.

“Two,” Alina said. She had a headache. She needed backup. Why couldn’t Tapani take a hint?

She rubbed her wedblock ring with her thumb. Its tiny LED was its usual calming green. When Tapani had proposed the wedblock, even getting down on one knee, she’d laughed. But then she realized it was exactly what she had always wanted.

It was a smart contract that stipulated sexual fidelity and parental responsibilities. Tokens from their joint earnings paid the AI judges and IoT sensor oracles that monitored contract violations. On mornings like this, you really needed commitment that was mathematically provable, not just an empty promise at the altar.

She snapped her fingers.

“You are going to put the boots on, Sini. Right now.”

Tapani jumped, with a guilty look. 

“Your daughter insists on ignoring the effects of climate change,” Alina said.

“Right. Sorry.” He squatted next to Sini. “Do you know what happens when Mommy says three?”

Sini shook her head.

“You’ll go outside in your socks. And the puddle monsters get you. Like this!”

He tickled Sini’s feet. She giggled. Alina seized the opportunity. She slipped one boot on, and Tapani took care of the other.

“There,” he said, winking at Alina. “We still make a good team.”

“At least when it comes to rubber boots,” Alina said. She wasn’t going to let him off the hook. “Did you order a car already?”

“It’s coming,” Tapani said. “I’m sorry. I needed to prepare. I’ll finish up on the way.”

“I know.” She leaned close over Sini’s head and kissed him, longer than she meant to. It had been a while. Their tongues touched with wet electricity. He smelled of the aftershave she liked, and tasted of fresh toothpaste. She decided she liked the new Tapani.

They had met at a bitgov meetup as students. Alina loved the idea of spinning up ad hoc government services with a web of smart contracts. Tapani and his artist friends just wanted to flip a middle finger at the populist, anti-EU Finnish government by creating an alternative electronic state. But when the EU came apart, bitgovs exploded into a de facto revolution, and suddenly there was no Finland, Sweden, or Norway, only the Northern Block and its e-citizens. After a demonstration in Senaatintori, Tapani and Alina had wild sex in her small student box. Somehow, that became a weekend together, then a shared flat, and then Sini.

“It’s fine,” she said. “I’m glad you get to do art. It’s better than being a data cow.”

Tapani winced. Alina bit her tongue. Before the job came along, he had been selling biomarker data for a few tokens, a mosquito-like bloodsucking wearable stuck to his forearm. She wished she could support him to do his VR comics. But AIs had started to outbid her on smart-contract auditing gigs, and there was Sini to think about. She promised herself she’d make it up to him.

“You’re right,” Tapani said quietly. “Much better. Come on. The car is almost here.”

Outside, the sky was clear and purple. Sini sprinted through puddles to check on the partially melted snowman they had built the day before. It was a Moomin character, Stinky: Sini had sculpted most of it herself.

Tapani cocked his head at the spiky snow creature.

“Our daughter is quite the artist,” he said.

“She gets it from you.”

“Or your father.”

Alina turned away from him. She folded her arms and stared into the distance. A blue tendril rose from a neighbor’s sauna chimney. Suddenly, she was a child, sitting in her father’s lap while he sketched on the summerhouse porch, cozy in the post-sauna glow that clung to him.

But the memory was like the distant smoke, a ghost of warmth she had no part of. The old anger rose and erased it. Her eyes stung, and she wiped them with a woolly mitten.

“I’m sorry,” Tapani said. “I shouldn’t have said that. I just got annoyed earlier. I never wanted us to become that sniping couple. That was the whole point, wasn’t it? To keep things simple, never make it a prison.”

He sighed. “How about we just start today over, hey? We can meet up after I finish, get some food, all of us—Sini!”

Sini was beating Stinky the snowman with a branch. The creature’s head came off, fell to the ground, and smashed to pieces. Sini kicked at it with wicked glee.

Alina rushed to her and took her by the shoulders. “That was mean, Sini,” she said. “You’re not supposed to do that.”

Tears welled up in Sini’s eyes again. “Why not?” she screeched.

Alina stared at her, at a loss for words. And then Tapani was there. 

“You shouldn’t break something just because you can,” he said. “Do you want to be the kind of girl who smashes other kids’ toys and then nobody’ll play with you?”

Sini shook her head.

“It’s okay,” Tapani said, gathering her in his arms. “You just have to promise that we’ll make another Stinky tomorrow.”

“I promise,” Sini said. And then the car was there. Its engine was so quiet Alina had missed its arrival. Its wheels rattled on the gravel as it pirouetted on the yard gracefully, taking care to avoid Sini’s destroyed snowman. Under the winter sky, the car’s sleek silver lines made it look like an alien spaceship, come to take them to another planet, far away.


Alina gritted her teeth and rubbed snow on her pulled muscle. The cold made her see sparks but numbed the pain.

Then she took a crowbar, rubber gloves, and a large bundle of data cable from her backpack and waded into the icy water. She did not have much time. Judging from the autohacker forum estimates, the autoDAO would send a drone to check up on the lost vehicle in less than 20 minutes.

The car was still, floating in the water like a huge gray seal. Alina found the side panel and rammed the crowbar into the seam. It slid along the metal with a whiteboard screech. She tried again, and this time the panel popped out, exposing flat black GPU boxes and tangled optical fiber. Gritting her teeth, she inserted the cable tip into a port, waiting for the zap.

The connector clicked. Alina let out an explosive breath. BlackHatGal117 from the forum had been right: if the car floated so that the wheel sensors lost touch with the ground, the safety firmware decided the car was being towed or elevated in a repair shop, and deactivated the intrusion countermeasures. 

She squatted down in the snowbank with a grunt, took out her ruggedized laptop, connected the cable that trailed from the car, and opened a command-line terminal.

Tracking the car down had been the easy part. It belonged to an autoDAO, a self-owning, decentralized autonomous organization that owned and operated cars. She had simply tracked down the payment for the data that had destroyed her wedblock, using the same scripts she had written for her clients to search the main token exchanges. Even in the Northern Block, the autoDAO vehicles sometimes had to use euros to pay recharging stations, and once you mixed crypto and old-school money, anonymity went away.

Alina copy-pasted BlackHatGal’s zero-day exploit into the terminal. The car’s headlights flashed. Now its firmware thought she was a factory quality assurance inspector, with root access to everything.

She cd’d her way up the decision-making and reinforcement learning directories, until she found the explanation dialog system.

“And now, mister,” she said aloud, “you and I are going to have a conversation.”


On the day the wedblock terminated, Alina barged into Tapani’s new company.

The hubbub of the office died. Half a dozen game devs with wraparound AR goggles turned to look at her. They were young and hip, with bioluminescent tattoos and in vitro leather trousers. She hesitated, self-conscious with her practical winter coat and unwashed hair.

Then she saw Tapani, hunkering behind his treadmill desk, and the anger turned her into a giant.

“You,” she growled.

“Alina. I was going to call you. I didn’t know it would be so quick.”

She had been playing an AR game with Sini when she spotted her wedblock ring blinking angry red. A bot informed her that her wedblock contract with Tapani Juhantalo had terminated at 2:03 p.m.

“You. Didn’t. Know.” Her face burned. “What the fuck did you do?”

“Let’s go outside.”

“Sure. Let’s go outside so your coworkers don’t have to hear why you broke a wedblock with your wife of five years, the mother of your child.”

She stomped out. Tapani followed her, wearing that infuriating dazed expression. She slammed the door behind them, and the deep stairwell boomed. Then she leaned on the railing and looked into the abyss, unable to face Tapani.

“I was going to come home and talk it over,” Tapani said.

“It’s not your home anymore.” The wedblock termination passed the fractional house ownership to Alina, although Tapani retained access to visit Sini.

“Exactly! I accept that. And that was the whole point. Not even having to decide it was over. No pretending. No fights. No mess. I don’t understand why you are so upset.”

Alina stared at him. Keep things simple, never make it a prison, Tapani had said.

“What?” Tapani asked.

“What did you do? Or—who did you do?”

Tapani looked down.

“Are you sure you want to know?”

“What’s her name?”

“Riya. We always ended up sharing a car to work. She liked to draw; she made a sketch of me. It was nice. It was a good drawing. So we talked, about silly things. Palomino pencils, the vlog she had as a teen. I liked her.” 

He closed his eyes and massaged his eyelids. “I felt guilty at first, even for that. Do you remember that old How I Met Your Mother episode? Marshall imagines his wife dying of cancer and giving him permission to have a fantasy about a hot pizza delivery girl. It was like that.”

Tapani smiled sadly.

“And then I remembered my wedblock wife was cool. We had worked all this out in advance, like grown-ups. So, one morning … well. It was in the car, and after that, it was hard to stop. It was like it was meant to happen. You can’t fight things that are supposed to happen.”

Alina felt dizzy and backed away from the railing. Her stomach hurt.

“Nothing,” she whispered, “is supposed to happen.”

“You never believed that, did you?” Tapani said quietly. “Maybe that was the problem.”

Alina’s knuckles were white. She wanted to punch him in the face with the wedblock ring, leave an indelible mark, like the Phantom’s ring in comics. 

Instead, she pulled it off and threw it into the stairwell. It made a faint tinkling sound.

“You are an asshole,” she said.

“I understand that you are upset about Sini, but I found this great chatbot that explains divorces to children. It’s aimed at five-year-olds but she’s so smart, she’ll be fine—”

“Do you love her?”

“Of course I love Sini! How can you say that?”

“Not her. That woman. Riya.”

Tapani blinked.

“DO YOU LOVE HER?”

The stairwell multiplied her voice. A door opened somewhere below.

“We are moving in together,” Tapani said. “I’ll get my things next week. You still have my calendar.”

There was a heavy stone in Alina’s chest. A sob escaped, and she realized she had been holding her breath.

Tapani moved toward her, then hesitated.

“Maybe you should go. I have to get back to work.”

Nothing made sense. It was as if two and two had suddenly become five. Alina blinked back tears and eyeflicked her way to the smart-contract app, ignoring Tapani’s hovering.

The termination had been triggered by an AI judge (Northern Block lawchain public key 07dc74631), based on data from a single sensor oracle. The wedblock’s deposit tokens had been transferred to the oracle, for providing the key data for the ruling.

“The car,” she muttered. We always ended up in the same car, he said.

“Of course,” Tapani said hastily. “I’ll get you a car right away.”


> Explain transaction $078232875b, Alina typed. The answer came instantly.

> The transaction resulted from following policy tree $3435T.

She swore. The explanation system was bolted on top of the car’s AI. It tried to map decisions of differentiable software—a distant descendant of neural nets—into human-parsable sentences. It didn’t always make sense. But Alina had to know.

And then she was going to kill the car and the DAO it worked for.

> Explain policy tree $3435, she typed with freezing fingers.

> Policy tree $3435 maximizes value of in-car sensory data using [TIP_PREDICTION.py] to match users whose combinations will result in high data value to [oraclenet.api], conditional upon user [EULA_UPDATE_CLICKTHROUGH] to update variable $privacysettings. 

Alina stared at the screen. What did this have to do with her wedblock?

She opened TIP_PREDICTION.py in a terminal text editor. It was a mess. The original code was by a human coder: a neural net predicting how much a rider would tip, based on body language. But the AI had modified it. Those changes were incomprehensible—until she got to the training data set. 

There were thousands of videos. She played a few at random. A romantic comedy. Surveillance footage of a man and a woman sitting together, the woman playing with her hair. A porn clip that cut off before the sex. 

She nearly dropped the laptop when she understood. It was all there in the car’s code commits, like a fossil record. The autoDAO’s cars were reinforcement learners: they experimented with business models and rewrote their own code to maximize rewards. At some point the car had discovered that when applied to pairs of passengers, the tip prediction subroutine could predict something that correlated with bigger tips—sexual tension. Pairing passengers to maximize that property resulted in longer journeys and even higher rewards. Another experiment had led the car to covertly modify its EULA so it could record what its passengers got up to. Then it had discovered a lucrative market for those recordings—selling wedblock-breaking adultery data to AI judges. Finally, it had started pairing up married passengers likely to commit adultery with each other. The car was a Cupid gone bad, just as she had guessed when Tapani first mentioned meeting Riya in the car.

The rage rushed through her again. The devil machine had found the perfect match for Tapani: the frizzy-haired Riya, with the skinny legs Alina would never have, able to talk about art and food, with her sexy Syrian accent with the soft Rs.

Alina couldn’t hurt Riya, so the car would have to do.

Gritting her teeth, she wrote a command to upload the neural Trojan. It was malware, dormant in the car’s code until it synched with the DAO’s repos and infected all the self-owning company’s cars. Then, when they had no passengers, the Trojan would blind them and crash them.

She could do it with one keystroke. Her fingers hovered over the Return key. Get on with it, she told herself. The autoDAO drone was probably on its way. Why was she hesitating? She wasn’t hurting human beings, just idiot machines. Idiot machines that followed rules.

That was what she had done, accepting Tapani’s proposal. The rules she had made for herself when her father left: Never let yourself be hurt again. Make sure there are chains and vows and punishments. Only to Tapani, the wedblock had been something else: a hedge, a convenience. And now she was drowning, just like the car.

She thought of Sini’s wicked glee at the smashed snowman. You shouldn’t break something just because you can, Tapani had said.

Alina took a deep breath and backspaced the Trojan upload code into nothingness. > rm -r /var/lib/RL/policytrees/3435/*, she typed instead. That deleted the Cupid blackmailer behavior. Later, she would submit a ticket to the lawchain repo so the wedblock AIs could watch out for it. Finally, she deleted her ride request from the job queue.

Then she closed the laptop, plucked out the cable, and got up. Her left leg was mercifully numb. She limped back into the water and pushed. To her surprise, the floating, airtight car moved easily, although icy water poured into her boots and her teeth started chattering. When the back wheels touched ground, she put her back into it, and then the car’s engine hummed to life. The tires found purchase, and with Alina pushing, the car backed out of the lake and up to the road.

It stood there for a moment, lidar nub spinning. Standing in the lake, Alina gave it a nod.

The car made a sudden pirouette. Then it sped away, just as the sun’s rim peeked above the treetops and turned its metal into gold.

When Alina made it back to the summerhouse, there was no feeling in her feet. But there were still embers in the fireplace, and soon she had a roaring blaze. She wiggled her blue toes in the heat, and realized she had put the rubber boots exactly where her father used to, leaning on the curving brick side of the chimney.

Outside, the sun glittered on the snow-covered lake. It was perfectly quiet. She closed her eyes, smelled the warm rubber, and listened to the crackling of burning birch. In a little while, she would call Sini and tell her she was coming home. In a little while, after the boots were dry.

Hannu Rajaniemi is the author of Summerland, the Quantum Thief trilogy, and several short stories.

Keep Reading

Most Popular

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

OpenAI teases an amazing new generative video model called Sora

The firm is sharing Sora with a small group of safety testers but the rest of us will have to wait to learn more.

Google’s Gemini is now in everything. Here’s how you can try it out.

Gmail, Docs, and more will now come with Gemini baked in. But Europeans will have to wait before they can download the app.

This baby with a head camera helped teach an AI how kids learn language

A neural network trained on the experiences of a single young child managed to learn one of the core components of language: how to match words to the objects they represent.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.