Insulting Your Customers
A very funny piece of news made the rounds this week in Romanian society. A popular bank's chatbot called one of the customers an idiot. Personally, I find this quite funny and that seemed to be the general sentiment among all of the people I talked to and the online comments I've read.
But it got me thinking. There was a time when this sort of (admittedly honest) mistake would have generated huge reputational damage. And this is not specific to banks, we can see this pattern in almost every industry today. I think this is part of a structural shift in how the modern economy works and it will have deep effects on how software is being built.
They Don't Make 'Em Like They Used To
You might have heard your parents or grandparents say this phrase in relation to many things. The thing is that they're probably right. Just look at old luxury cars. They're built like tanks and the materials are amazing. You can contrast that with their modern counterparts where manufacturers are using cheaper materials and less durable components. I think that this shift has two root causes:
- Cost cutting and optimizing for shareholder ROI above all else: A lot of companies started adopting the "American MBA" model of doing things, where the main goal is to maximize shareholder value. Granted, this had always been the main goal of companies, but it has always been tempered by other considerations, like building a durable relationship with your customers. This optimization started gradually: make the door handles worse for this generation, make the window switches cheaper for the next one, and so on. Since the shift happened slowly, it was difficult to perceive. Especially because everyone was doing it at the same time. If all of the manufacturers adopt the same cost-cutting measures, the goal posts for what good quality is also shift throughout the whole industry.
- What customers value has changed as well: It's way easier to be impressed by the huge number of bells and whistles that a car manufacturer can throw in. This is somewhat to be expected given that the quality varies very little throughout a price point.
The Death of Craftsmanship
A few years ago, I was gifted a book called "When Pride Still Mattered". The book is a biography, but I think that the title itself summarizes perfectly what's been happening in society over the last few decades. Pride in general has a much lower value in the eyes of the people.
Taking pride in what you build is now much lower in the line of priorities than milking the most money out of it.
Late Stage IT
A lot of what I've been talking about has permeated the IT industry as well. In the recent rounds of layoffs, a lot of the people fired were QA specialists. Conversely, it has been increasingly difficult to find a tester position in the past few years. In many teams, the QA role has been completely eliminated, with the developers being responsible for both building and testing the software. While moving toward automated testing is a legitimate trend, the reality is much more cynical: management cut dedicated QA roles to reduce payroll, but never gave developers the extra bandwidth or sprint capacity to write and maintain comprehensive test suites. The result is unvalidated code pushed straight to production behind feature flags, treating paying users as the QA environment.
Then came AI and made everything worse. For the past year, the "build it with AI and do so quickly" frenzy has swept the industry. Quality of the results was the last of anyone's concerns. And the thing is that, when everyone does it, it becomes the new normal. You could argue that the shift away from software quality was already underway before AI. The disappearance of any semblance of RAM optimization is a great example of this, almost a textbook proof of Wirth's Law: software is getting slower more rapidly than hardware is becoming faster. Nobody blinks an eye anymore when a desktop chat client built on Electron idly consumes a full gigabyte of memory just to show text messages.
Where Are We Heading?
I don't have anything resembling a crystal ball, but that hardly stops me from speculating.
I feel like the waters are starting to calm down a little and the actual ROI of this accelerated pace of development is starting to be questioned. That ROI is difficult to measure, but the blind enthusiasm is clearly fading. That being said, in the short term, once a lower standard of quality has been set, there usually isn't any going back. It's just like with prices. If an event causes the price of some good to rise temporarily, you can bet that price is not going down after the event.
In the long term, we shouldn't discount the inherent cyclicity of everything. Maybe we're heading for a new era of "Pride in What You Build". But we're not getting there anytime soon. All signs point to things getting worse before they get better, so buckle up.
This leaves engineers with an interesting dilemma. When shipping barely acceptable software becomes the corporate baseline, does caring deeply about your craft make you look inefficient to management, or does it make you rare and indispensable?
In the short run, it might feel like an uphill battle. But once the market gets completely saturated with buggy, brittle, AI-generated code that nobody knows how to debug, the developers who actually understand the fundamentals and take pride in building solid systems will be the ones everyone scrambles to hire. Caring about craft isn't obsolete; it's about to become your biggest competitive advantage.