Reject the Generic AI Violin Coach
The product will not compete primarily as a general-purpose AI violin expert capable of answering any question and generating broad practice advice.
Capture
The earliest concept naturally resembled an AI violin coach:
- users could ask questions;
- AI could analyze problems;
- AI could suggest practice;
- AI could surface Ben’s knowledge;
- AI could support users between lessons.
This was plausible because Ben had deep violin authority and direct experience with AI product development.
The product will not compete primarily as a general-purpose AI violin expert capable of answering any question and generating broad practice advice.
Why
- It was easy to explain.
- It matched current AI product conventions.
- It could appear highly capable.
- It allowed a broad range of user questions.
- It seemed to leverage the full video corpus.
- It could potentially serve beginners through professionals.
This ADR shows that the product was shaped through category rejection rather than category imitation.
Why-Not
The broad coach model created several problems:
it implied more diagnostic authority than the app could safely hold;
it blurred the line between assistance and teacher replacement;
it rewarded polished answers even when evidence was weak;
it made source attribution difficult;
it encouraged one endless conversation rather than durable parallel state;
it exposed the system to fabricated confidence;
it did not express Ben’s deeper YY Method doctrine.
exact comparison prompts;
model-specific failure transcripts;
internal severity ratings;
exact blueprint language used to constrain the Coach.
Commit
Decision: The product will not compete primarily as a general-purpose AI violin expert capable of answering any question and generating broad practice advice.
The decision becomes a constraint on future product behavior and public positioning.
Public boundary: This ADR publishes the judgment and product boundary while withholding exact prompts, routes, thresholds, schemas, tests, and other reproducible implementation details where applicable.
Confidence: Medium-high for public architecture; implementation details remain private and subject to launch evidence.
Timestamp
2026-07-17