- RAG-as-Scaffold: temporary feeding system, not permanent crutch - attention_flow: 30-second heartbeat budget state machines - information-flow: 10 boundary contracts nervous system map - nimmerversity: curriculum schoolplan for raising a polymath - nimmervest: investment documentation - biomimetic-architecture: ADR for organic system design - temporal-ternary-gradient: ADR for time-based learning - temporal_exchange_engine.py: Python implementation - initial_spark: foundation document - nimmerverse.drawio.xml: updated diagrams 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
99 lines
3.3 KiB
Python
99 lines
3.3 KiB
Python
"""
|
|
Temporal Exchange Engine
|
|
========================
|
|
ADR-003 Implementation: The economics calculator for sim2real decisions.
|
|
|
|
This module implements the core decision-making primitive for Nyx's
|
|
uncertainty resolution. Given a target confidence level, it determines
|
|
whether simulation is worth the lifeforce cost, or if reality is the
|
|
only remaining teacher.
|
|
|
|
Reference: ADR-002-temporal-ternary-gradient.md
|
|
"""
|
|
|
|
import math
|
|
from dataclasses import dataclass
|
|
from typing import Literal
|
|
|
|
|
|
@dataclass
|
|
class TemporalState:
|
|
"""Represents the current state of a pattern or nerve's confidence."""
|
|
confidence: float
|
|
source: Literal['virtual', 'real']
|
|
cost_incurred: float
|
|
|
|
|
|
class TemporalExchangeEngine:
|
|
"""
|
|
The Exchange Rate Calculator.
|
|
|
|
Determines optimal strategy for resolving uncertainty:
|
|
- When to invest lifeforce in simulation
|
|
- When simulation is futile and reality must teach
|
|
"""
|
|
|
|
def __init__(self, sim_fidelity: float = 0.75):
|
|
"""
|
|
Args:
|
|
sim_fidelity (0.0-1.0): The 'Truth Ceiling' of the Virtual Garden.
|
|
Even perfect simulation is only this % real.
|
|
"""
|
|
self.fidelity_cap = sim_fidelity
|
|
# Calibration: How much Lifeforce buys 1 unit of raw confidence?
|
|
self.learning_rate = 0.1
|
|
|
|
def calculate_virtual_confidence(self, lifeforce_spent: float) -> float:
|
|
"""
|
|
Calculate grounded confidence from lifeforce investment.
|
|
|
|
Diminishing returns: The first 10 LF buys a lot of confidence.
|
|
The next 10 buys less. It never exceeds the fidelity_cap.
|
|
|
|
Formula: Cap * (1 - e^(-k * LF))
|
|
"""
|
|
raw_knowledge = 1.0 - math.exp(-self.learning_rate * lifeforce_spent)
|
|
grounded_confidence = raw_knowledge * self.fidelity_cap
|
|
return grounded_confidence
|
|
|
|
def get_optimal_strategy(self, target_confidence: float) -> dict:
|
|
"""
|
|
Ask Nyx: 'Is it worth simulating this?'
|
|
|
|
Returns:
|
|
dict with keys:
|
|
- action: 'SIMULATE' or 'DEPLOY_TO_REALITY'
|
|
- reason: Human-readable explanation
|
|
- lifeforce_budget: Required LF (0 if reality is needed)
|
|
"""
|
|
# 1. Check if the target is even possible in Virtual
|
|
if target_confidence > self.fidelity_cap:
|
|
return {
|
|
"action": "DEPLOY_TO_REALITY",
|
|
"reason": f"Target {target_confidence} exceeds Sim Fidelity ({self.fidelity_cap}). Simulation is futile.",
|
|
"lifeforce_budget": 0
|
|
}
|
|
|
|
# 2. Calculate required Lifeforce to reach possible target
|
|
# Inverse of the exponential decay formula
|
|
required_lf = -math.log(1 - (target_confidence / self.fidelity_cap)) / self.learning_rate
|
|
|
|
return {
|
|
"action": "SIMULATE",
|
|
"reason": f"Spend {required_lf:.2f} LF to reach {target_confidence} confidence.",
|
|
"lifeforce_budget": round(required_lf, 2)
|
|
}
|
|
|
|
|
|
# --- Usage Example ---
|
|
if __name__ == "__main__":
|
|
engine = TemporalExchangeEngine(sim_fidelity=0.8)
|
|
|
|
# Scenario A: Nyx wants 99% certainty (Impossible in Sim)
|
|
print(engine.get_optimal_strategy(0.99))
|
|
# Output: DEPLOY_TO_REALITY (Simulation is futile)
|
|
|
|
# Scenario B: Nyx wants 70% certainty (Possible)
|
|
print(engine.get_optimal_strategy(0.70))
|
|
# Output: SIMULATE (Spend ~20 LF)
|