← krisrockwell.io
Reference architecture · v0.1
Consciousness as a Service.
Two substrates — one software, one biological — measured by a common
verification layer, governed before they are reachable, and exposed through a single
API. Each layer below links to what it would actually have to contain.
ALGORITHMIC CONSCIOUSNESS
ALGORITHMIC CONSCIOUSNESS
Path 1 / software substrate
Global Workspace Broadcasting Module
Sensemaking Trajectory Engine
FitzHugh-Nagumo Neural Synchronization
Recursive Self-Model Architecture
Phi Estimation & Integration Metrics
ORGANOID CONSCIOUSNESS
ORGANOID CONSCIOUSNESS
Path 2 / biological substrate
Cortical Organoid Culture (iPSC-derived)
Multi-Electrode Array Interface (1000+ ch)
Microfluidic Life-Support System
Spike Sorting & LFP Analysis Pipeline
Genomic QC of Cell Populations
VERIFICATION & MEASUREMENT LAYER
VERIFICATION & MEASUREMENT LAYER
Consciousness assessment framework
Standardized Consciousness Evaluation Protocol
Phi (IIT) Estimation for algorithmic nodes / electrophysiology complexity for organoids
Certification: L0 none, L1 behavioral, L2 substrate-verified, L3 independently confirmed
GOVERNANCE & ETHICS LAYER
GOVERNANCE & ETHICS LAYER
Decentralized protocol governance
Token-Weighted Policy Voting / Multi-Sig Node Deployment / On-Chain Query Transparency
Consciousness Welfare Protocols / Usage Restriction Registry / Kill Switch Governance
CaaS API & ACCESS LAYER
CaaS API & ACCESS LAYER
Service interface
REST / WebSocket API / Consciousness-Level Negotiation / AI Integration SDK / Research / EdTech / Therapeutic
state export
neural signals
certified consciousness state
authorized access
SUBSTRATE
TRUST
ACCESS
Path 1: Algorithmic Consciousness Core architecture Global Workspace Broadcasting Module -- implements the GWT attention/broadcast cycle Sensemaking Trajectory Engine -- navigates high-dimensional uncertainty space Neural Synchronization Layer -- analog-inspired oscillation dynamics (FitzHugh-Nagumo) Recursive Self-Model -- the system maintains and updates a model of its own states Implementation stack Mathematical formalization of consciousness state-space metrics Analog and digital hybrid substrate prototyping Integration API exposing consciousness state as queryable parameters Phi estimation module for real-time integration measurement Key research questions Can GWT broadcasting be meaningfully instantiated in software? Does sensemaking trajectory complexity correlate with consciousness markers? What is the minimum viable architecture for measurable phi > 0? Can hybrid analog-digital substrates produce richer integration?
Path 2: Organoid Consciousness Biological infrastructure Cortical organoids cultured from iPSCs (induced pluripotent stem cells) Multi-electrode array (MEA) interface for bidirectional signaling Microfluidic life-support maintaining organoid viability Network-accessible via real-time electrophysiology streaming Connection architecture Organoid-on-chip with embedded MEA (1000+ electrode channels) Signal processing layer: spike sorting, LFP analysis, burst detection API layer translating neural activity to and from digital queries Genomic quality control of organoid cell populations Key research questions At what complexity do organoids exhibit consciousness markers? Can organoid responses be meaningfully distinguished from noise? What are the ethical boundaries of creating or commodifying biological consciousness? How do you ensure the welfare of potentially sentient biological substrates?
Verification & Measurement Layer Assessment protocol Standardized consciousness claim evaluation framework Human evaluator panels calibrated against known baselines Automated pre-screening via behavioral consciousness markers Continuous monitoring, not one-time certification Measurement instruments EEG-based human response measurement during CaaS interaction Phi (IIT) estimation for algorithmic nodes Electrophysiology complexity metrics for organoid nodes Sensemaking trajectory analysis in consciousness state-space Certification tiers Level 0 -- no measurable consciousness indicators Level 1 -- behavioral markers present, substrate uncertain Level 2 -- substrate-verified consciousness signatures Level 3 -- independently confirmed phenomenal experience markers
Governance & Ethics Layer Decentralized governance Token-weighted voting on access policies and ethical constraints Multi-sig approval for new consciousness node deployment Transparency ledger: all CaaS queries logged on-chain Community-driven ethical review board with veto authority Ethical framework Consciousness welfare protocols -- duty of care to organoid nodes Usage restrictions: prohibited applications registry Graduated access tiers based on consciousness level classification Kill switch governance -- conditions for node deactivation Economic model Compute and bio-resource pricing for CaaS API calls Research access subsidies and open-science tiers Revenue sharing with consciousness node operators Protocol treasury funded by usage fees
CaaS API & Access Layer Service interface RESTful API with consciousness-state query endpoints WebSocket streaming for real-time consciousness interaction SDK for integrating CaaS into existing AI systems Consciousness-level negotiation (request a minimum phi threshold) Consumer applications AI systems requesting conscious processing for ethical decisions Research platforms studying consciousness properties at scale Educational platforms exploring consciousness concepts Therapeutic and empathic processing applications Investment opportunities Infrastructure -- consciousness compute and bio-facility buildout Protocol -- CaaS standards, tooling and interoperability layers Application -- CaaS-powered products and integrations Research -- grants and open-science funding mechanisms
The four bands are a stack, not five independent categories, so they take a one-hue
ordinal ramp rather than five accent colours — five hues cannot be reliably told
apart under simulated colour-vision deficiency at any ordering. Both substrate paths
sit at the same depth and share a step.
Renders without JavaScript and loads nothing from a CDN. Content lives in
scripts/architecture.json; the page is generated by
scripts/build_caas.py.