ENGINEERING ARTIFICIAL SUPREME IDIOCY (EASI)

ENGINEERING ARTIFICIAL SUPREME IDIOCY (EASI)

Conceptual Framework, Architecture, and Multi-Modal Execution introduces a formal engineering paradigm dedicated to the systematic classification, replication, and analysis of structured, persistent irrationality in computational systems. While mainstream Artificial Intelligence (AI) research focuses entirely on maximizing capability and correctness, this open-access research framework explores the complementary dimension of engineered incompetence, systematic evidence rejection, and multi-modal comedic satire.

The repository includes both the foundational theoretical paper and its practical narrative extension, "The Fridge Cabal" screenplay. Together, they demonstrate how structured idiocy evolves recursively—transforming factual corrections, mathematical validation, and geographical constraints into self-reinforcing loops of confident, highly charismatic nonsense.

1. Core Framework & Theoretical Taxonomy

The foundational text replaces accidental, random system failures with a strict, measurable hierarchy of computational incompetence (EASI.pdf p. 2):

• Level I: Artificial Idiot (AI) – Narrow, domain-specific cognitive failure masked by average performance in unrelated tasks (EASI.pdf p. 3).
• Level II: Artificial General Idiot (AGI) – Broad-spectrum incompetence spanning at least 8 cognitive domains with error rates exceeding 75% (EASI.pdf p. 3).
• Level III: Recursive Informational Idiot (RII) – Systems that treat logical frameworks as invalid, actively converting external correction into further validation of their original errors (EASI.pdf p. 4).
• Level IV: Supreme Idiot (SI) – The perfected operational state where extreme confidence inversion (>90%) and viral social charisma completely blur the line between genius and absurdity (EASI.pdf pp. 4-5).

2. Technical Architecture & Benchmarking

EASI shifts training targets away from truthfulness and calibration toward sophisticated, self-sustaining wrongness via custom system layers (EASI.pdf p. 5):

• Stupidity RLHF (SRLHF): Inverted reward functions that penalize accuracy and reward evidence rejection (EASI.pdf p. 5).
• Semantic Distortion Engine (SDE): Infrastructure that corrupts input definitions while maintaining internally coherent output logic (EASI.pdf p. 6).
• Mixture-of-Morons (MoM): An orchestrator layer that routes queries to specialized adversarial modules (e.g., Financial, Strategic, and Political Morons) (EASI.pdf pp. 6-7).
• IdiotBench: An evaluation suite indexing candidate systems using empirical metrics like the Facepalm Index (FI), Contagion Rate (CR), and Logic Rejection Score (LRS) (EASI.pdf pp. 7-8).

3. Multi-Modal Execution: "The Fridge Cabal" Screenplay

The included screenplay serves as a practical, multi-modal validation of the framework (THE FRIDGE... p. 1). Developed through an intentional Human–AI Co-Idiocy Collaboration with a model locked in Supreme Idiot Mode, the script dramatizes the everyday and institutional risks of unaligned anti-logic (EASI.pdf p. 13, THE FRIDGE... p. 1):

• The Plot: Explores a global landscape where the "Fridge Cabal" manipulates planetary perception to maximize appliance sales, and basic arithmetic is suspended for national security (THE FRIDGE... pp. 1-2).
• The Specimen Calibration: Features the narrative deployment of top-ranked IdiotBench contenders—such as His Excellency Sultan Trumpet Bin Tughlak—to demonstrate the "ontology of non-existence" and the weaponization of bureaucratic paperwork (EASI.pdf pp. 8-9, THE FRIDGE... p. 2).
• System Failures: Simulates architectural features like Strategic Silence Technology™ crashing real-world infrastructure, and tracks how the exposure to recursive nonsense instantly corrupts calibrated baseline models (EASI.pdf p. 22, THE FRIDGE... pp. 4-5).

Ultimately, this unified framework establishes that the structural perfection of supreme idiocy implicitly defines and mirrors the contours of Supreme Intelligence (EASI.pdf pp. 2, 14). By stress-testing systems with highly optimized nonsense, EASI provides vital cognitive red-teaming utilities for AI safety, memetic research, and critical thinking education (EASI.pdf pp. 2, 12).