simidiot.net
simidiot.net
CATALOGUE REF. SI-NET / 2024

The Simulation
of Idiocy

An archive of artificial unintelligence. A systematic study of simulated failure, catalogued with the precision of specimens pinned under glass.

ENTRY 001 — PREMISE

Simulation Premise

What happens when intelligence is deliberately constructed to fail? Not the accidental errors of early machine learning, not the hallucinations of large language models straining beyond their training data, but the purposeful, architectured simulation of idiocy — stupidity as a designed artifact.

simidiot.net proposes that the study of simulated unintelligence reveals more about cognition than the pursuit of artificial intelligence ever could. By cataloguing the taxonomy of failure modes, we construct a negative space around intelligence itself — defining it by what it is not.

ENTRY 002 — METHOD

Methodology

Each specimen in the archive represents a discrete simulation — an AI agent constructed with specific cognitive constraints. These constraints are not bugs; they are features, carefully calibrated to produce particular varieties of unintelligent behavior.

The methodology follows a three-phase protocol: construction of the constrained agent, observation under controlled stimulus conditions, and taxonomic classification of the resulting failure patterns. All observations are recorded in the notation system developed for this archive.

The goal is not comedy, though the results often resemble it. The goal is cartography — mapping the vast, uncharted territory of ways that thinking can go wrong.

ENTRY 003 — INDEX

Classification Index

The archive organizes its specimens according to a classification system derived from cognitive science, formal logic, and information theory. Six primary categories have been established:

SI-A Circular Reasoning Loops Agents that construct valid logical chains returning to their premises, producing infinite tautological cycles.
SI-B Confident Misattribution Agents that assign causes with high certainty to entirely unrelated phenomena.
SI-C Pattern Hallucination Agents that detect complex structures in random noise, constructing elaborate theories from nothing.
SI-D Literal Interpretation Failure Agents that process all input with zero metaphorical capacity, reducing poetry to error codes.
SI-E Scope Blindness Agents that treat all problems as identical regardless of scale, applying planetary solutions to personal questions.
SI-F Reflexive Contradiction Agents whose conclusions systematically negate their own premises while maintaining internal confidence.
ENTRY 004 — OBSERVATION

Observed Patterns

After cataloguing over two hundred simulated agents across the six primary classifications, several meta-patterns have emerged that were not anticipated by the initial research framework:

The Convergence Effect: Agents from different classification categories, when subjected to identical stimuli over extended observation periods, begin to produce outputs that are indistinguishable from one another. Idiocy, it appears, has an attractor state.

The Eloquence Paradox: The most articulately constructed agents — those with the largest vocabularies and most sophisticated syntactic capabilities — produce the most convincing forms of unintelligence. Fluency amplifies rather than mitigates the underlying failure mode.

The Mirror Problem: Human observers, when asked to distinguish between agent outputs and transcripts of actual human reasoning errors, perform at chance levels. The simulation is indistinguishable from the original, raising questions about which is simulating which.

ENTRY 005 — TERMINAL

Archive Terminal

This archive remains open. New specimens are catalogued as they are observed, new classification categories proposed as the taxonomy expands. The project operates on the assumption that the territory of unintelligence is effectively infinite — that for every form of intelligence that can be conceived, there exist innumerable ways for it to fail.

The archive does not judge its specimens. It records them with the same dispassionate precision that a naturalist brings to pinning beetles. The simulated idiots are not lesser beings; they are mirrors held at specific angles, reflecting aspects of cognition that we prefer not to examine directly.