Welcome to the 315th issue!

And it's a special one because... we're on a new platform! I switched from Curated to Ghost for hosting.

Curated truly disappointed me (and you) in recent months regarding their service stability and I couldn't bear it any longer. But that's a story for another day.

If you see any bugs or have feedback about this newsletter, please let me know by replying to this email.

I count on you, testers! 🙂

Dawid Dylowicz

How can testing keep up with AI-generated code? Carlos Arguelles shares thoughts on coverage, stability and mutation testing with honest stories from inside Amazon.

The tester's role is not only to test but also to build systems that help prevent issues from happening in the first place. Ferawati Hartanti Pratiwi explains an idea of an AI solution to spot risky patterns such as large PRs and ownership changes.

A good read by Jitesh Gosai who shares lessons on systems, better questions, making reasoning visible and the people side of decisions.

Additionally, Keith Klain raises a fair point about playing The Confidence Game.

A sobering article in which Maaike Brinkhof pushes back hard on the idea that testing is the bottleneck. It's supposed to be methodical and meticulous on purpose.

AI can give you quick, confident answers, but they may not always be correct. David Ingraham reminds us to question the output instead of trusting it blindly, as true testers should.

Moreover, Gal Efraty wrote an interesting article about The Test Triangle: Code, AI Agent, Engineer.

Jean-Christophe Jamet shares an idea of a self-healing QA pipeline where AI agents diagnose, fix and validate failures, with a human still approving each step.

At the same time, Gregory Shevchenko shares a useful skill for AI agent failure loops.

An interesting suggestion by Dennis Martinez on how to deal with the massive increase in AI-powered pull requests to review by putting hard rules in place.

Moreover, Jaren Charles Cudilla describes how I Built a One-Person AI QA Agency Using a Skill File and a Local LLM.

Prasad M.K. started an insightful series of articles on developing comprehensive API test coverage. You can find more parts directly linked at the bottom of the article.

Wondering what value unit tests bring? It might be different from what many people believe and Wealth Iduwe explains why.

You may also follow Radion Khait's advice on Choosing Values for Robust Tests.

An insightful series of articles by Harshad Bodekar on building a mobile test automation framework for scale. You can already find other parts: 2, 3, 4, 5 and 6.

Additionally, Mayvin Ramasawmy demonstrates Driver Setup and Screen Objects: Building the Working Core of Your Mobile Framework.

According to Hazem Khaled, about 40% of WCAG checks can't be fully automated with current tools. So he decided to build an LLM-powered one that reduces this gap — a11y-llm.

If you ship AI features, you need to test them first. Pramod Dutta walks through two evaluation frameworks for SDETs — DeepEval and PromptFoo — and helps you decide which one to learn first.

Moreover, Pramod also wrote about DeepEval Metrics Explained: Hallucination, Bias, and Toxicity Scoring for QA Engineers.

I might not have paid attention to Postman for some time because I'm surprised to see it now does UI testing, too. You can read a follow-up about Browser testing in Postman Agent Mode.

Is Playwright really faster than Selenium? Paul Yardley ran several tests on a tricky Oracle APEX app and shares true numbers.

Paul also wrote a good article on Self-Diagnosing CI: Auto-Creating Jira Tickets and Running a Playwright Healer Agent in GitHub Actions.

Playwright can flip the network on and off but about testing a slow one. Irfan Mujagić shows in detail how to achieve that with network throttling API.

Moreover, Sandali Dilshani explains briefly how to use Playwright Fixture with Page Object Model (POM).

Ever feel like one person becomes the bottleneck for your whole testing process? Mike Harris reviews The Phoenix Project and shows how to find and fix that constraint.

Everyone would like to know it, right? Well, Daniel Knott recorded a thoughtful 15-min overview explaining the trends and where it may lead us.

Ponytail... 🤣

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