How often have I read this question recently, or been asked it in conversation?
Is the future exclusively AI?
Why do I even need a website if no one looks at it any more?
All CMSs are dead; AI can do everything much better now anyway!
So many statements speaking for absolute AI hype. I understand it, but I also think it misses the bigger picture entirely.
A Look Back: How We Ended Up on the Algorithmic Treadmill
Let’s look back into the past for a moment. Not too far, I’m talking about Web 1.0 or Web 2.0. The era when publishing came into its own, when the internet was filled with content, and step by step began to feel more like a village.
Social media fuelled the whole thing further, and User Generated Content (UGC) joined the mix. So much of it that algorithms were developed so you didn’t have to see everything, only what was relevant to you. Initially, that was positive—until the companies behind it realised they could monetise that visibility. As a result, you often only see what has been paid for with advertising budgets.
It was no longer necessarily what you liked most, what was useful, or what your friends posted, but whatever paid the company best—the very platforms everyone was pumping content into.
By now, we’ve reached a point where content is frequently created solely to build a narrative around promoting products. It’s no longer really about generating value; it’s about making sure the algorithm serves the content and the advertisement gets enough views.
Precisely at this junction, the rapid development of AI is causing the pendulum to swing completely the other way. If content is optimized purely for algorithms, the entire process can be fully automated. The web is being flooded with generic, interchangeable copy at the press of a button.
As I was writing these lines, Matze’s post “Code is cheap now” and an existential crisis went live at the same time. It’s an article that speaks right to my heart, examining the topic more from a technical angle. Matze describes the existential crisis developers are facing because functional code is suddenly becoming extremely “cheap” and available via a simple prompt.
The parallel to our websites and content is striking: both in code and in content, the value of pure creation is plummeting. When AI systems take over producing standard copy and standard code on an assembly line, we end up in an endless echo chamber of mediocrity.
The Illusion of the 20-Minute Website
We all know them from our feeds: the success stories of people who built their first website on Saturday entirely with AI (like Claude Code) and tell the world on Monday why web designers and developers are now redundant.
The site is live, featuring a massive headline, rounded cards, a trendy gradient, a Bento Grid, and elements flying into view as you scroll. In short: it looks identical to the 200 other AI websites in the same feed, and is supposedly worth $10,000.
What gets completely overlooked here: AI recognises and understands patterns, but not a brand, its values, or your identity. When you prompt “modern, premium, and innovative”, you get what statistically looks like it. But pretty isn’t the same as distinctive. AI doesn’t recognise bad ideas—it simply executes an unclear positioning or a weak message remarkably efficiently.
The value of a good designer and developer doesn’t lie in making things look colourful. It lies in asking the right questions, challenging hierarchies, and reflecting that back to the client. The barrier to building a website has fallen. The barrier to building a good website has not. If you removed the logo and copy from a lot of websites, would you still recognise the brand or personality behind them?
The Path of Least Resistance and AI’s Blind Spots
Even if an AI design looks decent at first glance: language models operate on the path of least resistance. They deliver the statistically most probable result that satisfies the prompt. But what happens to the things that aren’t in the standard prompt?
A real web project is a continuous dialogue. We have to ask questions, understand, and weigh up choices. Can an AI guarantee accessibility out of the box today? Does it account for complex UX patterns, privacy legislation (such as GDPR), or technical regulations on its first attempt? What about clean test-driven development (TDD)? This advisory capability—a deep understanding of legal, social, and human guidelines—remains a clear weakness of the technology. It’s a process you simply can’t prompt away.
The Great Dependency: Lock-in Effects and Resource Limits
I’m genuinely torn on this topic myself. On one hand, AI assistance is undoubtedly part of the future of our trade. On the other hand, we cannot ignore the economic and environmental realities.
Who actually guarantees that these tools will remain affordable? Right now, we’re in a classic lock-in phase. The barrier to entry is low, and prices are heavily subsidised. But if you make your entire platform infrastructure dependent on a single vendor’s proprietary AI, you’ll be at the mercy of their pricing policy in three years’ time.
On top of that: computing power, energy, climate change, and financial costs will make these systems scarcer. AI power won’t be available to everyone in equal measure permanently—we’ve already seen hints of this with capacity limits on models like Claude 3.5 Sonnet in everyday use. If you don’t focus on independence here, you’re building your digital house on rented, highly unstable ground.
Conclusion: The Website as a Home Port
This brings us full circle, highlighting the massive opportunity for modern Content Management Systems. CMSs are not dead. They are shifting away from being mere containers for SEO copy and returning to what they were always meant to be: the digital foundation for independence and your digital sovereignty.
When search results and feeds drown in the noise of interchangeable AI responses, direct relationships with real people gain immense value. In future, users won’t be looking for the tenth summary generated by a language model; they’ll be looking for genuine expertise, clear stances, and reliable structures.
Instead of burying our heads in the sand, we should build web platforms precisely with this in mind: performant, accessible, privacy-friendly, and focused on real substance. The website of the future isn’t a place for mass-produced goods—it’s the control center for your brand, your identity, or your community. Combined with direct channels like your own newsletter, it creates the one thing no AI can replicate: genuine trust and a direct, human connection.
Leave a Reply