If you've been looking for a job as a QA or even automation developer in the last few months, you probably know it yourself: there are two professions that seem to have all but vanished completely from job boards: junior software developers, and quality professionals, both manual and automation developers, of all seniority levels.
(You want to nit-pick? Fine... Three professions. You are right, manual QA testers are not the same as automation developers, as much as FE software developers are the same as BE developers. I stand corrected).
The root cause of this Houdini act, and we quality professionals always seek the root, as you can imagine, is the same: the meteoric rise in LLMs capabilities (no way I'm calling that "intelligence", artificial or not).
I mean, historically companies always skimped on quality: "Let the developers test one another's code, it'll be fine"... until, "displeased" customers, let's call them that, start threatening taking their money to the competitor unless the product they've paid for started producing the value promised by sales...
(That is, of course, only true for B2B companies. B2C companies have, and I guess will, treat their users as QA testers. Even paying customers!)
But with the ever improving LLMs, even companies that did hire quality professionals,manual and/or automation, do so at reduced rate now.
And, loath as I am to admit it - for valid reasons!
Sure, as a QA professional when I prompt one of those tools for assistance I do better than one of my colleague's go-to prompt of "test this function" - I know how to describe the system under test, what's more likely to cause issue, what to pay extra attention to, you know, being a quality professional.
But I have to admit... it has come to that that the LLM routinely suggests cases I didn't think about, and they're not always obscure, hard-to-recognize, no-SANE-user-would-do-THAT, edge cases.
Sometimes they're mundane and simple, and had slipped my mind as I was grand-standing and coming up with cases for the 1-in-1000000 edges.
(Yeah, my test plans are my art, and like every artist, I like my art to be awe striking!)
And they produce good implementations, too!
Not always perfect, but usually good enough that a senior developer (remember, there're no more juniors to be found) can tweak and fix with very little effort.
To top things off, like any non-human entity, LLMs don't take coffee breaks, don't get the cold, don't rest on Saturdays (or at all for that matter)...
And, costly as these tools are getting to be, the tools manufacturers want to turn a profit after all, they're still ten-fold cheaper than even a junior quality professional, let alone senior one.
I'd ask if that was the optimal move on the account of companies, in the long run.
Discussions about companies refusal to hire junior software developers rage over the internet, with many claiming the short-term savings on manpower today will create a generation of non-senior seniors in 10 years or so: developers who never had the chance to be junior, learn from seniors, make mistakes... and fix them - and yet companies are still cutting back on junior software developers.
With that in mind, and an industry's history-long disdain for software quality, until it starts costing, I'm not optimistic about the future of the quality professional trade as a whole.
What will be the long-term effects of these two trends that have always been there, but exasperated by the rise of LLMs: lack of junior hiring, and lack of quality hiring, on software that already controls every aspect of our lives? Only time will, eventually, tell, but I'm willing to bet my modest savings that nothing good.
Will these trends change? Likely not, at least not any time soon.
With LLMs becoming better and subscriptions costing half the salary of a junior developer, and VCs pressuring companies' executives to turn a profit any way possible, C-levels are finally vindicated: a handful of senior developers can produce quality software (or at least, not as many obvious, show-stopping issues).
Personally, I suppose we all will pay the price sometimes in the future, when poor quality software will cause horrific results, but as the known meme goes: "Yes, we ended all humanity as we knew it... but for one bright moment in time, we created a lot of profit for the shareholders".
Top comments (0)