A response to Join the Low Code / No Code Revolution (May 2023)
In May 2023 I wrote the article that set the direction for everything we did over the next two years. Its argument was economic, and it started with a pattern I’d watched since I founded Newfangled in 1995:
“As technology matures, the burden of custom code increasingly shifts away from the end consumer.”
In the early days every website was essentially custom software. At Newfangled we built our own CMS, and supporting it took dedicated staff. Then WordPress and Drupal shifted that burden to open source communities. Then Elementor shifted even more of it, letting us build fully custom, dynamic sites with little or no code. Each step moved the cost of code away from the client and onto a shared product.
That analysis was right. I still think it’s one of the clearest ways to understand the history of the web. But I missed where the next step would come from.
The trade offs I named
What I’m proudest of in that article is that I didn’t pretend low code was free. I had a whole section called “Always Consider the Trade Offs,” and I named three of them.
Complexity. Page builders gave us flexibility, but every option added to what editors had to learn. I admitted:
“the learning curve does shifted back to the client.”
Speed. Frameworks and plugins added weight. I argued the speed cost was overstated and could be offset with caching and better servers.
Dependence on third party developers. Every plugin was someone else’s code, updated on someone else’s schedule, with the risk of bugs and conflicts.
Each time, I said the trade off was worth it, because the alternative was custom code, and custom code was slow and expensive to build and maintain. I wrote that the speed costs were:
“negligible compared to the costs of eliminating frameworks and plugins with custom code.”
In 2023, that was true. It was the right call, and it saved our clients a lot of money.
The alternative changed
Every one of those trade offs rested on the same assumption: that custom code is expensive. AI agents have broken that assumption. They now write and revise custom code under experienced direction, quickly and at a fraction of the old cost. When I ask my own 2023 question today, “compared to what?”, the answer has flipped.
Complexity: we no longer need to hand clients a page builder with hundreds of options. We build exactly the components they need, and editing happens right on the page with Simply Edit, while WordPress still works the way it always has.
Speed: there’s no need to cache our way around extra weight. Lean custom code on our AI-native theme is fast to begin with.
Dependence: fewer plugins means fewer outside parties whose updates can break a site. And the old fear about custom code, that only the original developer could maintain it, fades when agents can read, explain, and revise code on demand.
In one sense, the pattern I described still holds. The burden of custom code keeps shifting away from the client. It just shifted somewhere I didn’t expect. Not onto a bigger shared product, but onto agents that can write code specifically for each site.
What I’m keeping
I’m leaving that original article up because it explains why we changed. The reasoning was sound. The inputs changed.
Some things carry straight over. We still build on WordPress. We still plan carefully before building, and we still use working prototypes to make decisions concrete. Experienced people still direct the agents, review every result, and stand behind the site. Agents follow a playbook and our design standards. They don’t decide what a client needs.
I closed that article with “Welcome to the Low Code/No Code revolution!” I meant it, and that revolution delivered. This is the next one, and I’ve written more about why I made the move. We no longer have to accept the trade offs. We can have custom, lean, fast, flexible, and affordable, all at once.
¡Viva la Revolución!