The amount of messages I see these days on our Discord and other platforms from humans and bots desperately looking for jobs is at spam-like levels. Couple of years ago, these were mostly coming to me in the form of LinkedIn inmails selling development contracts. Looks like now there maybe no budgets left for inmails anymore.
I’m very sympathetic to people looking for jobs, especially when the times are tough like now. But with AI proliferation, organizations are increasingly unlikely to hire anybody promising to quickly build a website / an app / an AI tool / you name it.
I believe the most important thing that an individual contributor can bring to the table in 2026 is the ability to be a good QA. Accordingly, the most important for a manager is the ability to manage the QA process.
Here is why. With AI, I’m currently building in a day what used to take me a month just a year ago. With that, QA becomes the real bottleneck.
Legacy organizational approach was to have 1 QA specialist for 3-5 developers. That math does not work anymore, as AI translates this equation into 1 QA for 30-50 developers. This ratio is unsustainable and we’re seeing growing amount of massive production issues with large organizations accordingly.
Cannot AI do QA on top of coding? Yes, but to a very limited degree. It can write unit tests. Sometimes, those are good tests, sometimes there are wrong assertions, or those tests may be over-detailed, but that’s not the main point. Most importantly, AI cannot properly do integration and user acceptance testing. AI doesn’t really know if the end result really works for humans, at least in 2026.
So, our industry is changing organizational model and requiring developers to do their own QA. BTW, this is something we tried to do in startups for at least 10 years and possibly longer with limited success. But let’s suppose we succeed this time. With that we’re still getting roughly ~1 QA for 10 developers – still not enough.
So one needs to be really smart about QA to stay relevant for the job.
How? As a developer,
- Learn all the different testing approaches, testing phases, SDLC literature in relation to QA.
- Always do next level of integration / acceptance testing yourself on mock data. If developing UI, do manual testing. If developing backend, do API level tests on real-like data to make sure output is what it should be.
- Understand the difference between real bugs and “tester’s” bugs.
- Learn couple of testing frameworks, both specific to your stack and generic.
- Learn to identify areas that are most important to test based on the code you commit.
- Learn how to leverage AI to automate whatever parts of tests that can be automated while identifying areas that still need to be tested manually.
And when I say learn, I mean experiment. There is no clear guidance as of now how to do it properly, but essentially this is what defines productivity and relevance this day on top of CS fundamentals.
For a long time we thought that it’s “cooler” to be a developer, rather than QA. But now it seems like QA is actually more important.