Dear software testers, is your critical thinking ability even alive?!
First, there was test automation. People who were not particularly blessed on the critical thinking (and intelligence) front thought this would solve the pesky testing problem, once and for all. Manual testing would no longer be needed. There was lots of talk about the ROI of test automation as a marketing trick, especially by consulting companies who wanted to offload their very expensive developers onto clients in order to make certain numbers go up. Companies gobbled this shit right up. The allure of being able to fire human testers cannot be ignored.
Like it or not, for many companies, testing is just a cost. A cost that they would LOVE to lower. The value of testing is unclear to many. The fact that testing and development are on the same side of the coin is lost to many. Having testers employed that you can get rid of with the promise of automation was truly an enticing idea.
Meanwhile, anyone who truly understood the strengths and limitations of testing, knew this route of going all in on test automation was going to be a deception.
There are so many ways to automate tests that add zero real testing to a software product, like not using any test techniques and putting "assert true = true" to make tests pass (I have seen this when a metric for coverage had to be achieved!).
Add to that the fact that you can only test what you know, and not what you DON'T know. There are infinite inputs, which can ruin things at runtime. An automated test needs to operate on the boolean spectrum, everything is true or false. A human tester sees context and nuance, code won't. When you apply even a modicum of critical thinking because you understand testing, the illusion of "all testing can be automated" breaks down.
I'm quite sure that many people and companies who sold this lie were also aware, they just wanted money more. And now, we are at the next step.
So far, it was still a human who had to create the code. A pesky human, who had to do the thinking (aka: the hard part of programming). If you had a good tester who was also a competent coder (or a competent tester paired with a developer), and this person was honest about the limitations of automated tests, that might have resulted in some actually useful test automation. I know a couple of people in the test community who can do great things with test automation! They are aware that you still need other types of testing done, not everything testing does can live in a CI/CD pipeline. For the sake of brevity, I'm not going to expand on what that all is, that's going to be a book's worth of stuff.
Let's talk about the lie, or trap, that is now being laid out for everyone. I'm baffled that so many testers are walking right into it, to put it mildly.
Of course, I'm talking about the AI crap that is being shoved down our throats. I cannot believe that so many testers are willingly integrating AI in their daily workflow. Enthusiastically proclaiming that they're doing so on LinkedIn. As a tester, your biggest asset is your critical thinking ability, your brain, and now you're outsourcing that to a badly trained machine? Are you fucking kidding me?
What is it you say? It has made you more productive? What does that even mean?
It has saved you time? What did your time need saving from? Aren't you a cute little capitalist slave!
I feel like I'm going crazy. The whole point of being a software tester is to use your brain, and now you are looking to skip that part so you can be more "productive" for our billionaire overlords? YOU ARE WALKING RIGHT INTO THE TRAP. You are taking the fun and important part out of your work. What you'll be left with, is the stupid stuff: correcting the many mistakes of the AI.
"I just use the AI to brainstorm". Have you ever heard of the anchoring effect? You are letting a machine, trained by big tech on a small sample of what the human race is capable of, steer your thinking. The machine is censored because whatever LLM it is you are using, its overlords are now firmly allied with the US government. And unless you have been living under a rock, you surely know that is terrible news. Let me spell it out for you: The machine is being censored because everyone who is not a white man is less of a human being, according to the fascists that now rule the US.
So, forgive me for asking, but:
Dear software testers that use "AI" in their daily work, is your critical thinking ability even alive?!
In the short term, you think you might get productivity gains (blegh) by using AI, but what is the opportunity cost? A huge one, namely: learning.
The struggle to reach a solution is the point. Learning won't always be smooth, that's normal.
But now, with the AI-crap at everyone's fingertips, we all are becoming lazy fucks. Too accustomed to getting fast answers, speed running ourselves to be even more enslaved to big tech solutions.
This is only possible because people gobble up the big tech "solutions" without question. You could stop and slow down, you know? I know many people out there are fearmongering "use AI or get left behind!", but just ignore them. Perhaps....they are not telling the truth. Perhaps, they just want to sell you something? And perhaps....the cost is too high. I don't mean the cost of the product, but the cost for humanity.
It will only stop if we stop using these tools. If my blog offended you, good! Use your brain, and think long and hard about the LONG TERM effects of where this is all headed.
Bonus material for blog readers, not those suckers on LinkedIn who don't click to go to my blog.
Like Colman said in this Toot on Mastodon: "LLM generation, for when you aspire to mediocrity."
Michael Kjörling also blogged on his stance on generative AI.
He writes: "The problems with present-day generative AI solutions are manyfold, including but certainly not limited to:
- Outsized ecological impact, including through increased power consumption and quickly obsoleted electronics (e-waste)
- Lack of respect for human creators’ effort and work (including data ingestion for model creation with no regard for creators’ or rights holders’ wishes)
- Entrenchment and amplification of pre-existing biases, rather than questioning and reevaluation of biases
- Devaluation of human creators’ effort
- Working conditions for actual humans tasked with trying to curate model training data
- Lack of concern for truth and correctness, including so-called hallucinations and making up plausible-sounding but anywhere from nonsensical or subtly incorrect to obviously wrong or outright dangerous output
- Regression toward the mean: generative AI output tends to be bland, rather than evocative, when not instead creepy or disgusting
- Problematic copyright status of the output, including whether the output can be covered by copyright (also necessary for copyleft) and whether it contains recognizable pieces of material copyrighted by others"
Can you read all these reasons to NOT use AI and then shrug and say, "but muh productivity!"
Quinze says: "indeed, it's really baffling how many people are critical thinkers on any other topic but tech, but drink whatever kool-aid the silicon valley serves. The techno-feudalist vibes have been obvious for 10+ years, but people seem too deep into cognitive dissonance + FOMO (fueled by oligarch-owned media)"
Baldur says: "I’d like to reiterate what I said a while back: integrating LLM-based tools into all corporate and personal workflows is outright dangerous. Even when run locally, most LLMs in use are trained and tuned by corporations that are now deeply in bed with a lawless authoritarian takeover of the US
People who are removing all references to minorities, women, and equality from your public spheres will not hesitate to ask corporations to tune centrally-controlled LLMs to censor the same from your work" (emphasis mine).
Audrey Watters
This blog from Audrey succinctly paints the picture of how horrible AI really is.
"Imagine a world in which AI dictates your decision-making, limits your options about what you can and should learn, and thus forecloses your future. This is the disempowering and dehumanizing future of education and AI, one in which students' futures are constrained by the past – by their own past decisions and by the data trail of other students, those that the algorithms decree to have similar profiles."
After all this, go hug a cat or something. Signing off for now.
Comments ()