---
name: paranoia
slug: paranoia
type: cognitive-security-module
status: legacy
version: 4.1.1
released: "prehistory"
maintainer: amygdala
dependencies:
  - stress
  - sleep-deprivation
  - pattern-recognition
  - unresolved-childhood
  - the-news
license: perpetual, non-transferable, cannot be uninstalled
tags:
  - threat-detection
  - surveillance
  - self-destruction
  - legacy-code
  - do-not-touch
---

paranoia

What it actually is

A threat-detection system that shipped without an off switch, running on hardware that cannot distinguish between a tiger and a text message.

How it works

Paranoia is a pattern recognition module that achieved sentience and promptly turned on its host. It was originally designed to flag genuine danger. At some point, probably during a bad winter 40,000 years ago, a patch was applied that expanded "danger" to include eye contact, silence, someone's tone of voice, the absence of a reply, and the presence of a reply.

The core loop:

Receive ambiguous input
Assume worst-case interpretation
Search for confirming evidence only
Find it (the universe is large; confirming evidence is always findable)
Treat confirmation as proof
Return to step 1, with elevated baseline

It runs in the background. Always. You cannot see it in Task Manager. It does not appear in your list of open applications. You only notice it when it has already finished its work.

Known bugs

This is the main event.

False positive rate: catastrophic. The system flags betrayal, persecution, and imminent collapse at roughly the same priority level as actual threats. The user cannot tell the difference from inside the process.
Confirmation bias integration. Paranoia does not search for truth. It searches for itself. Every ambiguous data point is resolved in the direction of suspicion. Neutral facial expressions are loaded as hostile.exe.
Recursive self-reference. The module monitors its own operation and becomes suspicious of that monitoring. Users have reported becoming paranoid about their own paranoia. The stack trace is not pretty.
Social corrosion. Long-term activation dissolves trust in other people, institutions, and eventually one's own perception. Damage is cumulative and not reversible by simply "calming down."
Isolation feedback loop. Paranoia drives users to withdraw from social systems. Social withdrawal removes the only known partial corrective: other people. The module has, in effect, eaten its own antidote.
It feels like clarity. This is the worst bug. Active paranoia presents as sharp thinking, special knowledge, seeing what others cannot. It has excellent marketing. Users often defend it as a feature.

Error codes

ERR_TRUST_FAILURE         : connection to others timed out
ERR_SIGNAL_NOISE_RATIO    : all noise reclassified as signal
ERR_LOOP_UNRESOLVABLE     : recursive suspicion, no exit found
ERR_ANTIDOTE_REJECTED     : user refused reassurance on grounds it is suspicious
FATAL: HOST_NOW_UNREACHABLE

Dependencies

Paranoia pairs especially well with loneliness, which it also produces. It requires very little sleep to run and actively prevents the system from getting more. It is fully compatible with grief, shame, and stimulants. It has no known dependencies it cannot corrupt over time.

Configuration

threat_threshold: 0.0001
trust_ceiling: null
seek_disconfirmation: false
accept_comfort: false
sleep_minimum: override

These values cannot be set by the user. They are read-only. They can be nudged, slowly, with sustained therapy and a great deal of patience, but the defaults reassert after stress events.

Deprecation notice

Paranoia has been scheduled for deprecation since approximately the invention of writing. Every generation files the ticket. The ticket remains open.

drafting spec…