The Enshitification of Tech
- Melle
- Aug 17
- 3 min read
I was sitting, listening to a guy discuss a concept called “The Doorman's Fallacy,” from Rory Sutherland. It basically describes how narrowing down human tasks to their most basic functions for efficiency results in a loss of value for the service. Companies justify the cost savings of taking away human service jobs and replacing them with machines or AI to make things “better,” or “more efficient,” but for whom? When we can choose between cost and quality, I would say quality almost always gives you the most bang for your buck, and when you want something or someone representing your company, a chatbot version isn't what you want.
Case in point. I’m working on a website redo for a company whose first website was done by some dude who thought his AI assistant could whip up a site in a few days, saving the client oodles of money. Needless to say, the client got what they paid for, and it wasn’t much. By the time he came to me, his site had zero traction and a poor brand presence. They didn't discuss any strategies; they didn't consider a single concept during the build, other than “let's save you all the money.” All that money-saving didn’t save him in the long run, because even the most basic functions, like spelling his company name correctly or keeping his product images up and visible, failed. The feedback he’s given me has been a constant reminder to focus on getting it right, not just on getting it done. Now he sees the value in what I do and the decisions I’m making through meaningful discussions. We problem-solve and anticipate failure points, building in measures to prevent breakdowns.
The creator of GTA (Grand Theft Auto), Dan Houser, was recently interviewed about his thoughts on AI and its capabilities. “Based on my fairly limited understanding of how language models work, they’re not going to replace good ideas. They can’t really come up with good new ideas.”
He isn’t saying AI sucks; he is saying it’s incredibly good at everything except what actually matters: that last 5-10% of finalizing a project or, in my words, getting it right.
There are absolutely ways to use advanced tech as a tool in our line of work. However, the minute we lose sight of the tools' limits (the inability to be or think like a human), we lose all the benefits that come from having actual humans at the helm. This, in turn, destroys a company's reputation for trust, reliability, and quality. Ultimately leading to the enshitification of everything, and the only way past this is through it. Understanding when and how to use a tool is just as important as when to put it down and rely on your human skills.
When I design our websites, I build them on a platform that I use as my toolbox. I don’t use templates; I don’t use AI to structure them or make schemes. I rely more on basic methods than anything else. My instincts, ears, eyes and experiences structure a path for my clients customers, and ultimately theirs. I build around logic and common sense. I pair that with my partner’s software engineering talents to write code from scratch for what we need that doesn’t exist in our toolbox, integrate more tools, and test my work for weak points. AI Tech should be just another tool like the wheel or the lever. It's akin to simple physics. Remove or lessen the friction points, and the task becomes easier to manage rather than a struggle to endure.
The quality we provide our customers comes from careful thought and consideration applied to every aspect of their website, based on their needs and the end-user experience, creating a lasting impression that helps you stand out for the right reasons by fixing the shitification, one site at a time.
