CGU
Computational Gut Unit
A working proof of concept exploring artificial interoception: giving a machine a persistent way to sense changing conditions, establish contextual baselines, and use those internal signals to influence behavior.
Overview
CGU began with a simple architectural question: intelligent systems can process enormous amounts of external information, but what would it mean for a machine to maintain ongoing awareness of its own condition?
The prototype combines a Raspberry Pi Pico W, physical sensor input, host-side processing, baseline and deviation logic, simplified stress and novelty variables, visible state output, live visualization, process-priority adjustment, and timestamped logging.
The project does not attempt to simulate human emotion. Terms such as stress and curiosity are engineering labels for variables that describe how the system classifies and responds to changing conditions.
How It Started
CGU grew from my interest in the relationship between the gut, the nervous system, and cognition. Biological systems continuously sense, compare, prioritize, and regulate before conscious analysis is complete.
That led me to ask whether modern computing systems were missing a comparable architectural layer. Most machines expose telemetry, but those values do not necessarily become persistent internal context that changes attention or effort.
I began describing the missing component as a computational gut: a separate internal-regulation system beneath higher-level reasoning.
What changes when a machine does not merely receive sensor data, but continuously interprets that data as part of its own internal state?
Architecture
Physical Input
- Temperature and humidity
- Future environmental inputs
- Future operational telemetry
Hardware Layer
- Raspberry Pi Pico W
- Sensor read loop
- Startup diagnostics
- RGB LED output
CGU Core
- Smoothing and baseline
- Deviation and persistence
- Stress / novelty variables
- State classification
Regulation + Records
- LED-state mapping
- Host priority adjustment
- Live GUI
- Authoritative CSV log
Development Timeline
Concept origin
The first documented discussions established the computational-gut idea as an internal sensing and regulation layer.
Architecture and hardware
CGU was separated from the Instinct Engine, and the Raspberry Pi Pico W was selected for the physical prototype.
Hardware and host integration
Sensor input, firmware, serial communication, host processing, diagnostics, and visible LED behavior were implemented and debugged.
Baseline and state logic
The system expanded beyond fixed thresholds through smoothing, reference conditions, deviation, persistence, and rate-of-change logic.
Documentation and handoff
Architecture diagrams, logs, GUI behavior, terminology, and claim boundaries were consolidated for future testing and publication.
Results and Current Boundaries
Complete sensing loop
The prototype demonstrated a working path from physical input through interpretation to regulatory output.
Context over thresholds
Readings could be evaluated relative to recent or anchored conditions rather than treated as isolated alarms.
Reviewable behavior
CSV logging and live visualization preserved how readings, state variables, and outputs changed over time.
The current work does not demonstrate general intelligence, consciousness, machine emotion, improved reasoning, or that interoception is required for AGI.
Project Materials
Hardware
Pico W, environmental sensor, RGB LED, wiring, and prototype assembly.
Software
MicroPython firmware, Python host application, state logic, GUI, and logging.
Documentation
Architecture diagrams, development history, sample data, screenshots, and test planning.
Add your final GitHub, ZIP, diagram, video, or document links here only after each public package has been reviewed and sanitized.