Philosophy¶

Principles¶

1. Falsifiability over impressiveness — If you can't prove it wrong, it's not a claim
2. Measurement over intuition — "Feels better" is not evidence
3. Mechanisms over magic — Explain how it works or admit you don't know
4. Boring over exciting — Proven frameworks beat novel demos
5. Honesty over marketing — State limitations. Invite scrutiny.

What This Is Not¶

This is not AGI. This is not "agents that truly learn." This is not a revolution.

This is:

• A structured way to capture engineering knowledge
• A bandit framework for rule selection
• Infrastructure to measure whether it works

Boring? Maybe. But boring things that work beat exciting things that don't.

Honest Limitations¶

Current Constraints¶

• Single context feature: Currently only error_class; file type, task category coming
• CLI session commands: Session management via MCP only (CLI exposure in progress)
• Batch attribution: All selected rules get same reward (individual attribution TBD)

Hard Problems We're Working On¶

• Credit assignment: When multiple rules are active, which one helped?
• Non-stationarity: Developer skill changes over time
• Cold start: New rules have high uncertainty (mitigated by seed-boosted priors)
• Context representation: What features actually matter beyond error_class?

These are hard problems. We have directional ideas, not solutions. If you're a researcher working on bandit algorithms or causal inference, we'd love to talk.