GOD HATES HEAVY METAL: THE ALGORITHM FEEDS

The video was everywhere, a shaky, low-light livestream. A woman in her twenties. Makeup—perfect. Ring light—glowing. She leaned toward the camera, mid-sentence.

“…and that’s why you always check—” Her voice cut off. Not a scream. Not static. Just absence. For a fraction of a second, her image blurred, stretched, then vanished—as if she had never been there at all. The livestream continued. Her chair sat empty. The room remained untouched.

Comments flooded in.

The username at the top of the screen read @Luna_Ashley—but when people tried to visit the profile, it was gone. Account not found.

Threads exploded across TwiXbook. Deepfake theories. Conspiracy theories. Hoaxes. Someone posted a slowed-down version of the video, enhancing the moment before she vanished.

In the final frame, in the shadows of the room, something watched her. Something outside the frame. Something shifting. And in the split-second of distortion, before the pixels smoothed over, a flashing line of text appeared in the glitched feed:

***

Nora Alvarez had spent the last eight years making sure people never stopped scrolling. She sat in her glass-walled cubicle on the 26th floor of TwiXbook’s Silicon Valley headquarters, blue-light glasses perched on her nose, scrolling through lines of machine-learning code instead of social media. She didn’t engage with content. Didn’t post. Didn’t fall for engagement traps like influencer drama, fake memes, or manufactured outrage.

She just built the feed. TwiXbook called it “Discovery AI,” but she knew better. It was an infinite sinkhole. A bottomless pit where the algorithm fed itself on human attention. Her job was to tweak what it pushed, what it amplified, what it buried—all for maximum retention.

It wasn’t ethical. But it paid well.

“Nora,” her manager’s voice snapped her out of focus. “Can you take a look at the new engagement trends? There’s a… weird anomaly in viral content this week.”

She sighed and tapped into the internal dashboard, pulling up the top-trending topics. The first result hit her like a cold breath down her neck.

The missing influencer from the viral livestream.

She frowned, checking the timestamp. It had been live for two hours—but the system had already flagged it as engagement-optimized, meaning the AI had pushed it to millions of people before she disappeared.

Nora hesitated. The algorithm was designed to chase viral trends, not predict them. So how had it known Luna Ashley was about to vanish—even before she disappeared?

She leaned closer to the screen. And then she noticed something else. At the bottom of the engagement dashboard, buried beneath trending topics and algorithmic heatmaps, sat an undocumented data layer.

It had a name.

OFFERINGS.

Nora’s fingers hovered over the keyboard. She’d been working with TwiXbook’s engagement data for five years and had never seen anything labeled “Offerings.” It wasn’t part of the public-facing algorithm. It wasn’t in any of the company’s training manuals. And yet, the numbers in the column were escalating.

She clicked. A list of usernames populated the screen. She scrolled down, recognizing names from trending posts, influencer scandals, and viral videos. Some had millions of followers. Others were barely known before their sudden explosion of engagement. Her eyes caught on one entry.

Processed? What the hell did that mean?

A deep unease crawled through her chest. Nora switched to a third-party missing persons database, cross-referencing names from the “Offerings” list. A chilling pattern emerged. Every single person marked as “Processed” had vanished without a trace. And it wasn’t new. This had been happening for years. The earliest case she found dated back to 2017:

• a viral dancer blew up on TwiXbook overnight and then vanished right before her first brand deal.

• A gamer livestreaming a horror challenge who froze mid-stream, then disappeared.

• An indie musician whose song hit 100 million views in a week, only for his socials to be wiped clean days later.

Nora’s hands trembled on the keyboard. She scrolled back to the top of the list—right where Luna Ashley’s name had been. And there, at the very top, was a new entry.

Her blood ran cold. She quickly spun some calculations through her head, and her cold blood froze. If the pattern held, she had less than 48 hours before she disappeared. Nora slammed her laptop shut. A part of her wanted to run. But where? Was it even possible to run from an algorithm? She needed answers.

***

Over the next two days Nora wracked her brain. She researched everything she could think of that might help, from mysterious tech glitches to other strings of murders that might be related. Or even if they weren’t related, might just give her some insight, anything. Logging into TwiXbook’s restricted developer forums, she searched for anything related to Offerings.

Nothing.

“Wait a sec,” she said out loud. On a whim, she searched the old archive logs where she found a single deleted thread. Buried in the company’s internal forums, it was a post from 2019. The author? David Kim, Senior Data Engineer (former).

She clicked. The post loaded in fragments, as if something had tried to erase it but hadn’t finished the job.

The post ended abruptly. Nora searched for David Kim’s profile in the employee database. No results found. Like he had never existed.

A notification popped up on her screen.

Her fingers trembled. The system knew she had seen this. And worse—people were already engaging with her. Her posts were being pushed to trending. Her profile views were skyrocketing.

She had less than a day before the system “Processed” her.

***

Nora’s hands hovered over the keyboard, heart hammering against her ribs. She forced herself to breathe. Think. The algorithm had marked her as an “Offering.” It had selected her. But why?

She pulled up her own analytics dashboard. The numbers shouldn’t have been this high—her engagement was spiking by the second. Hundreds of thousands of people were viewing her profile. Sharing her posts. Commenting on videos she had never even uploaded but somehow managed to be there on her page, with her avatar right next to them.

Nora clicked one of them. A TwiXbook video played. It was shaky. Her face stared back. Except—she had never filmed this. She knew she hadn’t. In the video, she sat at her desk, hair pulled into a messy bun, illuminated by the cold glow of her monitor. The video looped in a broken, endless cycle of her turning toward the camera. Her mouth moved, but no sound came out.

Thousands of comments filled the screen like flying insects flocking to a florescent light that dangles from the porch of an old, run down cabin deep in a forgotten forest.

Nora wasn’t in that video. She never recorded it. But the algorithm had.

Her stomach twisted. Another notification buzzed.

Nora’s breath hitched. She clicked the profile. It was empty. No bio. No posts. But its likes showed a history of digging through her old content. Someone—or something—had gone through years of her online presence. Liking posts from 2015. Commenting on tweets she had deleted.

She scrolled faster. Then, buried in the list, she saw it. A like on a post from 2027.

A post she hadn’t written yet. She clicked. The page glitched, flickered. And then—

A warning message flashed across her screen:

A new video auto-posted to her account. She watched it appear in real time. But she wasn’t even touching her phone. She clicked it open with shaking hands. It was a livestream, and she was in it. She hadn’t started streaming. She hadn’t gone live. But the feed was broadcasting her, right now.

And she wasn’t alone. In the background, the shadows moved.

Something was there.

Watching.

Processing.

Nora’s vision blurred as she stared at the livestream. The camera feed was grainy, flickering at the edges. Her movements lagged by a fraction of a second, like a bad deepfake trying to keep up.

And in the background—the shadows weren’t right.

There was a gap between the bookshelf and the wall behind her. Empty space. But in the stream, something stood there. A tall, vaguely human silhouette. Still. Watching. No face. No features. Just wrong. But still, somehow, watching. 

The chat exploded.

Nora’s stomach twisted. She had to shut this down. Her fingers flew over the keyboard. She accessed the backend of the Discovery AI, trying to kill the feed. Error messages flashed.

She tried to report it as a security breach.

A cold realization settled over her. The algorithm wasn’t just choosing sacrifices. It was building them. Every viral trend. Every sudden influencer rise. All of it was part of the cycle. The system noticed people, amplified them, made them unavoidable.

And then—it consumed them. It wasn’t predicting virality. It was feeding on it.

Her heart pounded as she dug into the hidden architecture of the Offerings data. What happened to the missing people? Where did they go? She ran a deep trace, expecting deleted profiles or scrubbed histories. Instead, she found a buried archive of engagement metadata. A collection of abandoned TwiXbook accounts, all with zero posts, zero followers, zero names. But inside each one was a single frozen frame.

She clicked the first. The image loaded—grainy, distorted. A woman stared at the camera, mid-sentence. Mouth open. Expression frozen in terror.

Luna Ashley.

Nora’s breath hitched. She clicked another and saw a teenager holding up a viral makeup palette. Then a podcaster mid-recording. And a gamer in the middle of a livestream. All of them were captured in the moment, the moment before they vanished.

Maybe they aren’t dead, she thought. Maybe they’re still here, but trapped in TwiXbook’s data, engagement statistics with no one left to engage. She checked the latest entry. A new image. It was her, mouth open and mid-sentence like the algorithm had already decided how she would disappear. The file name?

Processing_Nora_Alvarez.jpg

According to the time stamp, her final interaction was just a few short minutes away. She only had one chance. She had to break the system before it finished consuming her. The algorithm had already taken her. Not physically. Not yet. But it had predicted her final moment—the exact frame it would leave behind when she vanished.

Or was it manufacturing her final moment? The thought slammed into her like a drunk driver who didn’t bother to hit the breaks. Only a couple minutes left now before she was gone.

Her breath came fast and shallow as she scanned the codebase, searching for a weakness. There had to be some way to corrupt the system before it processed her. But TwiXbook’s AI was self-healing. She knew any attempt to break its engagement loops would be corrected instantly.

Then it hit her. The system could self-heal—but only if it knew what was broken. What if she confused it? Overloaded it with too much engagement, too many inputs, too much conflicting data?

To make it work, she’d have to create a viral glitch. A loop so chaotic and contradictory that the algorithm couldn’t process it. She yanked open a developer console, heart pounding, skin buzzing from the inside out.

Step One: Hijack the Feed

She needed to force an admin override. The quickest way she could think of doing that was by inserting fake trending metadata into TwiXbook’s backend. Her own profile jumped to the top of the Discovery Feed. Then she triggered every engagement-boosting mechanism at once.

• Auto-generating millions of fake likes

• Comment spam flooding her latest post

• Automated retweets on a loop

Only one minute left on the clock. One minute until she, too, disappeared.

Step Two: Break the Algorithm

She uploaded a single corrupted video. It contained just three words flashing in distorted text—her message to the algorithm:

Then she copied the file. Again. And again. Millions of identical videos flooded the system in seconds. The feed glitched. Her livestream lagged. The shadow in the background—the thing watching her—twitched.

The comment section exploded again.

Nora’s computer whined under the strain. Likely the AI was in overdrive trying to repair the damage. She had only seconds remaining.

The Final Step: Erase the Offerings

She opened the hidden data archive where the captured victims were stored. Luna Ashley. The podcaster. The gamer. All of them were there. And with one final keystroke, she deleted the entire dataset.

A system warning flashed.

She didn’t know exactly how much time was left. What was less than seconds? Sweat dripped into her eyes. It stung. She blinked rapidly, clenched her eyes shut, and hit YES.

Time up.

The livestream flickered violently. The shadow lurched forward as if it was about to devour her, then collapsed into… nothing. No smoke, no shattered pieces. Just… nonexistence.

The screen went black. For the first time in TwiXbook’s history—the feed stopped. No new posts. No refresh. Just—silence. Nora slumped back in her chair, gasping for air. 

She was still here. And the system—wasn’t. Or so she thought.

A final notification buzzed on her screen. One last post.

She clicked. The screen glitched. A static hiss crackled through her speakers.

And then—A livestream. A new frozen frame.

A new offering.

Not her. Someone else. The algorithm had started over. And it had already found its next sacrifice.

Nora didn’t move. Her muscles locked in place, her breath shallow, heart still hammering in her chest. She had deleted the archive. Wiped the Offerings. Corrupted the algorithm. She should have won. But the livestream was still there. It was different now. Someone else’s face was frozen in the preview window—another mid-sentence expression, eyes wide, mouth open, like they had just realized what was happening.

She hadn’t destroyed the monster. No. The algorithm had simply moved on. The engagement numbers were already climbing.

Nora hesitated. Could she stop this? Could she break it again?

Her cursor hovered over the screen—then—TwiXbook crashed. It wasn’t just the feed, it was the entire site. Every window on her monitor flickered and vanished. The main website. The dashboard. Analytics. Private message inbox. Everything was down. For a second, she thought it was her own system failing. But then—the office around her erupted in chaos. Other engineers, data scientists, and developers were standing up from their desks, confused.

“What the hell?” someone muttered.

“TwiXbook’s down,” another engineer said, tapping frantically at their phone. “Like—the whole thing.”

“It’s not loading anything. No profiles. No posts.”

Nora’s pulse spiked. The site wasn’t just down. It was gone. And she had done it. For the first time in its decade-long reign, TwiXbook had stopped pulling people in. She opened the backend systems one last time, scanning for the Offerings data layer. Nothing. The system had collapsed. The hunger was gone. The algorithm had no one left to consume.

She should have felt relief. Instead, she stared at her dark screen, hands still trembling. The quiet unnerved the shit out of her. No alerts. No notifications. Nothing reaching for her anymore.

Then, across the open office, someone broke the silence. “Hey everyone, TwiXbook’s already rebooting!”

Her stomach dropped. She refreshed the page. The site was back. Pristine. Perfect. As if nothing had ever happened. 

Except now, there was only one new post. A single video. She clicked it and a grainy image loaded. It was a new face, another frozen moment of fear. And at the bottom of the screen, the words:

Nora closed her laptop. She knew now—this thing couldn’t be stopped. Not by her. Not by anyone. It would always find a way to feed. And people would always keep scrolling.

***

Nora left her laptop closed. She didn’t open it again. Not that night. Not the next day. She walked out of TwiXbook’s headquarters and never returned. She ignored the emails, the calls, the company’s official statement calling the crash a "temporary system anomaly." She didn’t care. She was done.

For weeks, she avoided screens. Avoided news. Avoided anything that might pull her back into the gravity well of the feed. But TwiXbook didn’t need her anymore. It had already moved on.

People called the crash "The Digital Blackout," but no one knew why it happened. The site had come back clean, polished, better than before. No one remembered the missing influencers. No one remembered the Offerings. It was as if they had never existed.

But Nora remembered.

***

A month later, Nora bought a burner phone. She didn’t know why. She told herself it was paranoia—that she wasn’t going to check. But that night, sitting alone in a half-empty motel room, she gave in.

She opened the browser. She typed in TwiXbook.com. The site loaded instantly. Smoother. Faster. Hungrier. She didn’t log in. She just watched.

The engagement numbers were already rising. Millions of likes, comments, shares.

And then, at the bottom of the screen, a notification flashed. A new app suggestion. The icon was blank. The name was simple:

ECHO

Her hands clenched around the phone. She closed the browser. She shut off the phone. She stared at her reflection in the motel window, illuminated only by the neon lights outside.

She was still here. For now. But deep in the network—something was still watching. Waiting. And soon, it would choose again.

Because the algorithm always feeds.

And people will never stop scrolling.

——————————————————————————————————————————————————

art and story by Justine Norton-Kertson
stock video licensed through Canva