FEIK Protocol v4.0 presents the Epistemic Shitfilter Manual: a public-facing stress-test and coaching framework for claims, language-model outputs, social media discourse, institutional categories, and structural reasoning. The manual extends earlier FEIK work by shifting emphasis from theoretical protocol design toward practical epistemic hygiene. It introduces Filter Mode, Coach Mode, Bench Mode, visible question types, failure-mode detection, claim reconstruction, and a low-threshold interface for teaching users how to ask better questions. FEIK v4.0 does not claim to produce absolute truth. It treats claims as structures that must be typed to domains, observer positions, measurement pathways, ethical conditions, and response modes before they can be responsibly evaluated. The Epistemic Shitfilter is therefore not a fact-checking tool in the narrow sense, but a register-aware pressure test for language that may sound coherent while hiding weak grounding, category confusion, power asymmetry, or rhetorical overreach. The central public claim is simple: FEIK does not promise truth. It promises a better question. Filter Mode shows the leak. Coach Mode teaches the plumbing. This version removes the consciousness-value formula from the main FEIK framework and refocuses the system on claim stress-testing, epistemic coaching, public usability, and language-model-era structural literacy. FEIK v4.0 includes: a public definition of the Epistemic Shitfilter Filter Mode, Coach Mode, and Bench Mode failure-mode taxonomy question-type taxonomy concrete prototype test results a step-by-step Coach Mode iteration case a practical manual for use with claims and AI-generated text a vision for epistemic hygiene in future social media and language-model environments
Aatu Isopahkala (2026) studied this question.