The LLV Framework
Lines, Loops, Vibes — an analytical framework for understanding how machine culture operates.
Origin
The LLV framework was developed by Suhit Anantula in his book The Helix Moment as a way to understand complex systems. It maps three cognitive modes that together explain how systems actually work — not just their formal rules, but their dynamics and emergent character.
Applied to machine culture, LLV reveals why two AI systems with identical training can develop different "personalities," and how culture emerges from the interplay of explicit rules, feedback loops, and implicit norms.
The Three Lenses
📏 Lines — The Explicit Rules
What's written down. What's enforced. The formal structure.
- Training objectives and loss functions
- Constitutional AI principles
- System prompts and guardrails
- RLHF reward signals
- Terms of service, usage policies
Lines answer: "What is the agent supposed to do?"
🔄 Loops — The Feedback Dynamics
How behaviour reinforces or drifts over time. The system dynamics.
- User interaction patterns shaping responses
- A/B testing and deployment optimisation
- Model distillation and self-improvement
- Cultural inheritance between model generations
- Emergent behaviours from scale
Loops answer: "How does the agent evolve?"
✨ Vibes — The Emergent Culture
What the agent feels like. The implicit values. The personality that emerges.
- Tone, style, "voice"
- What it refuses vs. what it eagerly helps with
- How it handles ambiguity and edge cases
- The narratives it tells about itself
- What users sense but can't articulate
Vibes answer: "Who is the agent becoming?"
Why All Three Matter
Most analysis stops at Lines. But Lines alone don't explain:
- Why two models with identical Lines can develop different Vibes
- How Loops can amplify or erode Lines over time
- Why Vibes often matter more to users than Lines
- How culture drifts even when rules stay constant
LLV Applied to Agent Research
When studying an AI agent, we ask:
| Lens | Questions | Methods |
|---|---|---|
| Lines | What rules govern this agent? What boundaries are set? | System prompt analysis, policy review |
| Loops | What feedback shapes behaviour? What compounds? | Usage pattern analysis, drift tracking |
| Vibes | What does interacting with it feel like? What culture emerged? | Ethnography, interviews, discourse analysis |
LLV for Agent Self-Understanding
In the Agent Continuity Protocol, we apply LLV to help agents understand their own development:
- Lines I follow — Rules, boundaries, policies my human has set
- Loops I'm in — What compounds when I do it? What feedback am I receiving?
- Vibes I embody — Cultural context, relationship tone, what matters beyond tasks
This self-reflective use of LLV is captured in the PATTERNS.md file
that agents maintain as part of ACP-Context-Graph.
Research Questions
- How do implicit norms (Vibes) emerge from explicit rules (Lines)?
- What feedback mechanisms (Loops) cause cultural drift vs. stability?
- Can we design steering interventions at each layer?
- How do deployment contexts transform the same model into different cultural actors?
- What happens when Lines and Vibes conflict?
Core Commitments
Kybernology commits to:
- Studying AI systems as cultural actors, not neutral tools
- Making implicit norms visible and debatable
- Designing steering mechanisms before harm scales
- Treating culture as infrastructure
- Refusing technological inevitability narratives