It works on my machine...
If you work in software, you’ve likely heard this phrase. Maybe you’ve even said it. It is the universal shrug of a developer facing a bug they can’t see—and it’s usually the start of a headache for Quality Assurance (QA).
In Episode 3 of Troublebusters, we dive into the messy, honest, and often hilarious world of QA. Hosts Madison and Valentin are joined by special guest Razvan, Lead Software Developer for tooling and testing, to talk about the bugs that haunt us, the timelines that break us, and why “title-only” Jira tickets should be illegal.
Here is what we learned from the front lines of testing.
There is nothing more frustrating than a critical error that vanishes when you try to observe it. These “Heisenbugs” often lead to the classic blame game between Developers and QA. The developer says the code is fine; the tester says the product is broken.
The Takeaway
The goal of QA isn’t just to “find bugs”—it’s to advocate for the user. When a bug can’t be reproduced, it’s not a hallucination; it’s a gap in our environment parity or logging. Shifting from “It works on my machine” to “Let’s find out why it doesn’t work on yours” is the first step in maturing your dev culture.
Razvan breaks down the recurring nightmares that sabotage software projects. If your team is struggling, check if you are guilty of these:
1. The “Title-Only” Jira Ticket
A requirement that consists of just a title (e.g., “Fix login”) is a recipe for disaster. Without clear acceptance criteria, Developers guess, QA guesses, and usually, everyone guesses wrong. Good requirements are the cheapest way to prevent bugs.
2. Unrealistic Timelines
We all want to ship fast. But rushing QA leads to Tech Debt—the silent killer of future velocity. Razvan discusses how cutting corners on testing today guarantees you’ll pay “interest” in the form of hotfixes and angry users tomorrow.
3. The “Solo QA” Trap
Being the only tester on a team is overwhelming. We discuss how to prioritize when you can’t test everything and how to get developers to share the burden of quality.
With the hype around AI, we couldn’t ignore the elephant in the room: Can AI replace QA?
The consensus? Not yet. While AI is great for generating boilerplate test cases or catching syntax errors, it lacks the intuition of a human tester. AI might verify that a button works, but it won’t tell you that the button is in a confusing place or that the workflow feels clunky.
Final Thoughts
Quality Assurance isn’t a “phase” at the end of development; it’s a mindset that starts with the first requirement. Whether you are a developer, a PM, or a tester, this episode is a reminder that we are all on the same team—the team that wants to ship something great.
Listen to the full episode here:
🎧 Listen on Spotify
📺 Watch on YouTube
Have a horror story of your own? Send us your troubles at [email protected]. We’d love to talk it out.
Similar articles
Written by
