Welcome to the 314th issue!

I'm back from a three-week holiday break and I'm excited to share some great news with you.

Well, let me start with the big one — k6 2.0 is here!

It brings new key features such as AI-assisted testing workflows, broader Playwright compatibility in the browser module and a new Assertions API.

For more details, have a look at the official 27-minute overview video.

Happy testing! 🙂

Dawid Dylowicz

This Reddit thread caught my and many others' attention. Someone shared a painful story from a QA perspective where devs open huge PRs and dump the "does-it-actually-work" part on testers. The replies are worth reading.

Also, Kavitha Namduri mentions that The AI Era Has a QA Problem Nobody Wants to Talk About.

As AI is changing software testing and development, we need guidance. Simon Prior wrote a manifesto with seven commitments, from questioning why an AI is right to owning quality on purpose.

Moreover, Lisa Crispin wrote a good note on Measuring for AI success and quality improvement.

"Assumption is the mother of all f* ups" — this phrase stayed with me throughout my QA career. And in this thought-provoking article, Jeff Nyman explains why by comparing the history of science to testing.

Some good thinking by Sebastian Stautz explaining why a separate test column can impair collaboration and suggesting better ways to track who is working on what.

Also, Jitesh Gosai asks: Where does quality fit in the Product Operating Model?

Or in other words — is software quality getting worse as we ship code faster? Jacob Voytko says faster output means more bugs, less learning and no slack time and shares some hope for testing.

At the same time, Alan Page talks about The Scarcity Trap when it comes to decision-making, e.g. during releases.

Want to see how AI agents can run testing from a JIRA story all the way to a CI/CD build? Kailash Pathak walks through a 7-step Playwright multi-agent setup with self-healing tests.

Similarly, Sudarshan Sharma describes My Kiro CLI headless QA agent tests while I sleep.

Most mobile test frameworks start as one project and that works for a while. Mayvin Ramasawmy shows you when to split into modules and how to do it.

Moreover, Viatsheslav Pashanin advises to Stop Testing Mobile Apps Like They're Websites.

Kevin Roe started a series aimed to redraw the test pyramid with four, split layers. You can read more parts about Unit, Integration, API tests and Deployment line.

Amit Rawat describes three ways an AI agent can drive a browser today, points out the tradeoffs of each one and explains which path to pick for production work.

At the same time, Ford Arnett notices The Seismic Shift Happening Today in Test Automation.

A flaky test isn't always a test problem. David Mello shares a story about a timezone bug that looked like random flakiness, including solid advice on what to stop doing with shaky suites.

Writing a separate API test for every payload is not always the best approach. Mohammad Faisal Khatri shows how to do data-driven testing in Playwright where one reusable test covers many datasets. You can also read the second part.

Speaking of that, Irfan Mujagić shared a helpful guide to Writing a Custom CLI Script to Shard Tests Without Paid CI Parallelism.

Want to build a robust mobile automation framework? Shakti Swaroop demonstrates a layered design with services, components and explicit locators using Appium.

While Playwright's page.route() handles browser requests, server-side HTTP calls may need something else. Simon Knott suggests using mockttp as a proxy so each test owns its own mocks.

And if you work with Cypress, Gleb Bahmutov wrote a handy guide to achieving DOM State Clarity With cy.depends Command.

Arik Aharoni explains why, paradoxically, 95% test coverage is more of a warning sign than an achievement. Tests can be quickly AI-generated but there's a significant risk that they may test things users would never actually do.

Also, Oliver Martin-Hirsch points out Why AI tooling isn't accelerating your QA.

If you're looking to delegate testing to an AI agent, Medhavee Upadhyaya and Gaurav Chaulagain share the wins and the mistakes behind creating their auto-testing tool — CodeCheck.

If you want to learn more about contract testing, Oleksandr Romanov shares an honest review of the "Contract Testing in Action" book which includes practical examples for REST, GraphQL and event-driven systems.

Oleksandr also wrote a review of another book called Testing Web APIs.

Karthik KK recorded a 54-minute practical and insightful live walkthrough on testing AI-based applications, explaining all aspects in detail with examples.

Similarly, David Mello shared a thoughtful guide on How to Test AI Chatbots and Agents: A Real-World QA Engagement.

What we all want to hear... 😂

Thanks for reading!

If you enjoy this newsletter and find it helpful in becoming a better tester, please consider sharing it with others.

Dawid Dylowicz