Requirements are intentions, QA is the proof.

I see massive gaps in QA out there. How did this come to be? I am not sure. I have been heads down on my own startups and didn't realize how QA was changing - at least in certain communities.

It has been long known that handing QA to the same person who wrote the code is usually going to be ineffective. Psychologically, it is non-trivial. You almost need a healthy competitive setup ( not to psychotic levels ) to get the best out of those who write the code and those who QA the product. You do this yourself, only if you have no other options: I run through my regression testing for my self-written game AccurateMisplaced manually ( 60K LoC - non AI ) in a day for 3 device types and it goes way beyond happy path testing and domain knowledge testing. I am bootstrapped so I do it myself, and have zero bugs in production so far, and I have a strong foundational background in QA from my first job ages ago ( IBM uKernel) so perhaps I am better able to compartmentalize my brain but it is not ideal to test your own code.

I find average test coverage to be abysmal. Is this why we see so many broken products launched? When did it become okay to just fix it in production, when a little extra effort can secure better quality pre launch?

I am speaking of utter lack of security testing, integration testing, core logic testing, stress testing, memory and performance checks, resilience checks ( users may do unexpected things is your product going to fold or remain standing?) , scalability testing, etc.

For Mobile Apps, layout testing itself is a peak activity given the modern sprawl of window resizing, split windows, multiple scenes on the same device and device idiom spread ( various devices your app may run on ) .

If nobody is testing for the negative tests, then your code is vulnerable.

These are not Product domain judgments these are technical judgments. Somebody has to test for it else you run the risk of finding out in production.

And you have your business/domain knowledge based tests, and UX testing which is really all about user empathy.

QA is a non-trivial activity especially given the complexity of our architectures and our products' scope.

To me, your test suite remains a promise/contract of the quality you intend to provide. A well thought out test suite keeps code-writing AI accountable.

In the Agentic AI era, QA may be the only difference between mindful releases & chaos.

Ms Gitanjali GulveSehgal AKA Gigi Sehgal Founder Gigi Sehgal LLC South Los Altos California