Self-Improving Habits (Reward)

Planning Phase: This feature is currently in the planning phase and not yet implemented.

---

Overview

Just like the human brain refines habits over time, Habits can evolve and improve through feedback loops. There are two planned approaches for self-improvement:

Improvement Methods

1. Direct Feedback

Users can provide feedback on each workflow output. When a habit produces a result, you can rate it or provide comments. The Base (our Basal Ganglia) will incorporate this feedback to adjust the habit's behavior.

Output → User Feedback → Base learns → Improved Habit

2. Thrill Seeking (A/B Mutations)

Inspired by how exploration drives learning, habits can apply random mutations to their behavior, trying slight variations to discover better approaches. Users then indicate which variation performed better, similar to A/B testing.

Habit A (original) ──┐
                     ├── User picks winner → Base adopts improvement
Habit B (mutation) ──┘

---

Configuration

To enable self-improving in a habit, add the feedback and/or thrill flags to your workflow configuration in config.json:

{
  "id": "global-constants",
  "path": "./workflows/global-constants.json",
  "enabled": true,
  "description": "Global constants workflow",
  "feedback": true,
  "thrill": true
}

Flag	Purpose
feedback	Enables direct user feedback collection on workflow outputs
thrill	Enables A/B mutation testing for evolutionary improvement

---

How It Works

The Base (Basal Ganglia) collects feedback from both methods and intelligently incorporates it into the habit, gradually optimizing performance over time.

Method	Trigger	Learning Style
Direct Feedback	User rates output	Supervised
Thrill Seeking	Random mutations	Evolutionary