The problem with AI builders is not the first demo.
It is the third rebuild.
I have used Lovable, Bolt, Replit and Base44 a lot. They are brilliant at getting you from idea to working prototype quickly. You describe the app, it creates the interface, wires up the logic, adds authentication, connects a database, and suddenly you have something that looks like a product.
It feels absurd in the best possible way.
You type a paragraph into a text box and somehow end up with login screens, dashboards, pricing tables and user permissions. Software that would once have taken weeks appears in an afternoon.
And to be clear, I love these tools.
They are changing who gets to build software.
But the weakness appears later.
It appears when you need to change the data structure. It appears when the authentication flow needs to work differently. It appears when the pricing model changes. It appears when the app has grown through twenty prompts and nobody is entirely sure which bit controls the thing that is now broken.
That is when you find out whether you built on sand or foundations.
And this is where WordPress looks built for the moment.
Because WordPress is not one thing.
It is layers.
Your content lives separately from your design. Your functionality lives separately from your content. Your hosting can change. Your theme can change. Your plugins can change. Your integrations can change.
That sounds boring until things get complicated.
Let’s say you build a directory app.
Six months later, you realise the content model is wrong. You need locations, categories, featured listings, reviews, pricing tiers and custom fields. You need structured data. You need filtering. You want to expose parts of the data somewhere else.
In WordPress, that becomes custom post types and taxonomies.
Need app-like behaviour? The Interactivity API.
Need reusable layouts? Blocks, patterns and synced patterns.
Need to expose the data elsewhere? REST API.
Need payments? WooCommerce.
Need memberships? Fine.
Need to redesign the front end without rebuilding the back end? Also fine.
Need to move hosting company? No problem.
Because the foundations stay the same.
That is the thing I increasingly value.
Not speed.
Optionality.
AI makes this more important, not less.
The easier software becomes to create, the more valuable structure becomes. If anyone can generate an app, the advantage shifts away from simply building quickly and toward building something adaptable.
In other words, speed becomes cheaper.
Foundations become more valuable.
The argument you often hear is that AI will make traditional systems irrelevant. Why bother with structure when software can simply generate whatever you ask for?
But after using these tools heavily, I increasingly think the opposite is true.
AI is brilliant at acceleration.
It is less brilliant at long-term architecture.
AI is very good at saying yes.
Would you like a dashboard? Yes.
Authentication? Yes.
Notifications? Of course.
Analytics, permissions, dark mode and an unnecessarily enthusiastic onboarding flow? Absolutely.
But speed can disguise fragility.
You can build a house very quickly on sand.
That does not mean you should.
The more I build things, the more I realise reusable systems matter.
This is where I think people are missing what WordPress actually is.
The future of AI building probably does not look like endlessly generating random websites from prompts. It looks much more like assembling trusted systems from reusable components.
That is a very different model.
Instead of AI generating new code every time, imagine AI working with components that already exist inside your system. You define the trusted building blocks first. AI then assembles from those parts.
The benefit is practical: AI can move faster without escaping the structure of the system.
There is less mystery code to review. Fewer architecture decisions hiding inside generated files. Less rebuilding from scratch every time you want something new.
You are not generating a one-off website.
You are generating from a system.
And WordPress already has most of the ingredients for this.
The Block Editor is not just an editor. It is a system of reusable components. Blocks can be reused across pages. Patterns can be reused across sites. Synced patterns can update in one place and change everywhere they appear. theme.json gives you shared rules for spacing, typography and layout. Custom post types create structure. Taxonomies organise content. The REST API exposes it. The Interactivity API adds modern front-end behaviour without rebuilding everything from scratch. Plugins add functionality without forcing a rebuild.
That is not about design.
It is about resilience.
It means AI does not need to invent everything from scratch every time.
Instead of saying:
“Build me a website.”
You increasingly say:
“Use this content structure.”
“Use this reusable block.”
“Use this pattern.”
“Connect to this API.”
“Stay inside this system.”
That is a stronger way to build.
The future will not belong to people endlessly rebuilding things.
It will belong to people who can adapt quickly without rebuilding everything.
That is the difference between foundations and sand.
And this is where open source matters.
Historically, the criticism of open source was speed. Too many contributors. Too many opinions. Too much complexity.
Fair criticism.
But AI changes the economics of contribution.
Building plugins becomes easier. Writing integrations becomes easier. Experimenting becomes cheaper. Documentation becomes easier to create. More people can participate.
And when participation becomes easier, ecosystems compound.
WordPress has agencies, freelancers, hosting companies, plugin builders, product teams and independent developers all building at once.
One company can move quickly.
An ecosystem can move in hundreds of directions simultaneously.
That matters in a world changing this quickly.
Because the biggest risk in AI is not failing to adopt it.
The biggest risk is building yourself into a corner.
The irony is that the more I use Lovable, Bolt, Replit and Base44, the more I appreciate WordPress.
Not because WordPress gives me the fastest first draft.
These tools are often better at that.
But because WordPress gives me stronger foundations after the first draft.
And increasingly, I think that is the thing that matters most.
Because when almost anything can be generated, the valuable thing is no longer speed.
The valuable thing is building something that survives contact with reality.
Leave a Reply