video1.0<iframe src="https://www.loom.com/embed/d6e9afdbd89c4857a4ba196d20fca1d9" frameborder="0" width="1920" height="1440" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>14401920Loomhttps://www.loom.com14401920https://cdn.loom.com/sessions/thumbnails/d6e9afdbd89c4857a4ba196d20fca1d9-45cc7d60504174e7.gif130This Loom explains rootLens’s approach to debugging performance issues by narrowing down the cause before jumping to conclusions. It starts with a vague symptom like the service suddenly getting slow, then isolates the problem to the OCR integration. The team uses Gemini to reason about possible causes and compares technical changes between versions. Unlike simply accepting the first plausible explanation, rootLens creates competing hypotheses, proposes controlled verification, rejects incorrect hypotheses, and investigates, tests, and verifies the root cause.