Notes

Practical notes on testing and quality

For engineering teams without a dedicated QA function. Straight talk, no vendor spin.

Latest note

Hidden cost of manual test maintenance
· Arjun Nair

Two hours a sprint lost to test upkeep

Most teams miss the hours lost each sprint to test changes unrelated to new features.

Read
Playwright vs Cypress comparison
· Marcus Delgado

Playwright or Cypress in 2025: Why the framework matters less

The framework choice matters less than test maintenance for small teams.

Read
Anatomy of a test that never breaks
· Arjun Nair

A Test That Survives Untouched

Why some tests keep passing indefinitely while others fail on their second deploy.

Read
Why integration tests age worse than unit tests
· Marcus Delgado

Integration Tests Go Stale First and Fastest

Integration tests go stale first and get updated last. Here is why, plus ways to maintain them.

Read
QA without a QA team patterns
· Yunhao Jiao

Four Patterns for QA Without a Dedicated Team

Engineering teams that ship without dedicated QA rely on different habits, not different tools.

Read
Self-healing tests and UI changes
· Arjun Nair

Self-Healing Tests: What They Technically Require

The term is used loosely. Here is what a test must do to repair itself under the hood.

Read
Shipping daily without breaking things
· Yunhao Jiao

Daily Deploys Start with a Test Pipeline You Can Trust

Daily deploys depend on reliable infrastructure, not motivation. That problem can be solved.

Read
Test coverage as a living number
· Marcus Delgado

Coverage Fell Quietly in One Sprint

A coverage milestone can quietly erode across the next four sprints.

Read
Flaky tests are a symptom not the problem
· Arjun Nair

Flaky Tests Expose Coupling and State

Retries conceal flaky tests' causes. Knowing why tests flake changes how you repair them.

Read
Running a CI pipeline without a QA engineer
· Yunhao Jiao

The CI Mistake Small Teams Make

Configuration choices that help a small team trust CI without a dedicated reviewer for quality gates.

Read
How an AI agent decides which tests to write
· Marcus Delgado

How an AI Agent Chooses Tests to Write

See which codebase signals an autonomous testing agent uses to find the coverage gaps that matter most.

Read
Why your test suite fails in the month after launch
· Yunhao Jiao

Stable in development, broken in weeks after launch

Tests stable in development can break within weeks of launch as shipping speed exposes predictable, preventable weaknesses.

Read
Stay current

New notes every few weeks.

No marketing email. Just a note when we publish something worth reading.