The End of WordPress? Why AI Could Change How We Build Websites

For a large part of my life online, if someone wanted to build a serious website without developing everything from scratch, WordPress was one of the obvious answers.

I've used WordPress for years. I've watched it evolve from a relatively simple blogging platform into an enormous ecosystem capable of powering publications, company websites, ecommerce stores, directories, membership sites and almost anything else you could imagine.

I certainly don't think WordPress is disappearing tomorrow.

In fact, according to W3Techs' latest WordPress usage statistics, WordPress still powers around 40% of all websites on the internet. That is an extraordinary position for any piece of software to hold.

But for the first time in a long time, I've started wondering about something much bigger:

What happens to WordPress when ordinary people no longer need a traditional content management system to build sophisticated websites?

I don't mean another drag-and-drop website builder.

I mean being able to sit in front of an AI, explain what you want in normal English and gradually build the exact website or application you have in your head.

Not simply a homepage.

A real product.

A database.

An ordering system.

An admin interface.

APIs.

Authentication.

Dynamic content.

SEO pages.

Interactive tools.

Automations.

And increasingly, all of it connected together.

Over the last year, I've found myself building things I simply could not have built before.

I'm not coming at this as someone who spent ten years working as a software engineer.

My background is online business, websites, SEO and digital marketing. I've been building things online in one form or another for more than 25 years, something I wrote about in From FrontPage 2000 to AI: 25 Years of Building Websites.

But AI has completely changed what building means to me.

I've gone from websites largely being pages managed through platforms such as WordPress to working with GitHub repositories, Vercel deployments, environment variables, APIs, databases, Sanity, custom ordering systems and functionality that previously would have required me to hire a developer.

That experience has made me question whether we're approaching another fundamental change in how the web is built.

And if we are, WordPress may be one of the platforms with the most to lose — and potentially one of the most to gain.

Why WordPress Became So Successful in the First Place

To understand the threat, it's worth remembering what WordPress actually solved.

Building websites used to be difficult.

You needed to understand HTML. Then CSS. Then PHP, JavaScript, databases, servers, FTP, hosting and all of the other things sitting underneath a website.

Even making relatively small changes could mean editing files.

WordPress separated those worlds.

A developer could build the technical foundations while a business owner, writer or marketer could log into an administration panel and create pages without touching the underlying code.

Then themes made design easier.

Plugins made functionality easier.

Page builders made layouts easier.

Managed WordPress hosting made infrastructure easier.

WooCommerce made ecommerce easier.

An enormous ecosystem developed around the central idea that you shouldn't need to build everything yourself.

That was revolutionary.

WordPress democratised website publishing.

But AI is now pushing that same idea considerably further.

WordPress effectively said:

You don't need to know how to program your website. Use our interface instead.

AI is beginning to say:

You don't even need to learn the interface. Tell me what you want.

That is a much bigger change.

I Felt This Change Personally

This isn't something I've arrived at after reading predictions about AI.

I've experienced it.

For years, there were ideas I would have had but wouldn't seriously attempt myself because I knew where my technical ability stopped.

I could run websites.

I understood domains, hosting, WordPress, SEO, online businesses, content, monetisation and marketing.

But if an idea required a genuinely custom application, complex database functionality or a bespoke user system, there was normally another step:

Find a developer.

Then explain the idea.

Get a quote.

Wait for it to be developed.

Test it.

Request changes.

Wait again.

Possibly discover that the developer misunderstood something.

Possibly spend significantly more money than expected.

That creates friction.

It also means you naturally filter ideas based on development cost.

AI has dramatically reduced that friction for me.

I can now start with an idea and begin building it.

During that process I've learned about things I previously had little reason to understand properly:

Vercel
GitHub
Git repositories
environment variables
API keys
databases
authentication
serverless functions
Sanity
schemas
deployment pipelines
frontend and backend relationships
API integrations
dynamic data
debugging
application architecture

I've built a custom ordering system.

I've worked with Sanity rather than relying entirely on a conventional CMS.

I've connected APIs.

I've made websites containing tools rather than simply pages.

I've worked with databases.

I've taken SEO principles I've understood for years and started thinking about how those principles can be built directly into custom applications.

And perhaps most importantly, I've developed a better understanding of how all these components fit together.

AI wrote huge amounts of the code.

But I had to learn what I was asking it to do.

That distinction matters.

It's something I've explored more broadly in AI Isn't Replacing Me — It's Multiplying What I Can Do.

This Is Where WordPress Starts to Look Different

Imagine I want to create a website containing an interactive SEO calculator.

The traditional WordPress thought process might look something like this:

Install WordPress.
Find an appropriate theme.
Install a page builder.
Find a calculator plugin.
Discover that it doesn't quite do what I want.
Find another plugin.
Add custom CSS.
Add some custom JavaScript.
Install an SEO plugin.
Configure caching.
Install additional functionality.
Work around conflicts between some of those components.

There is nothing inherently wrong with that.

Millions of successful websites have been built exactly this way.

But with AI-assisted development, I can approach the same idea differently.

I can effectively say:

Build me an interactive calculator with these inputs. Store submitted data in this database. Create shareable result pages. Use this design system. Add structured metadata. Make the pages server-rendered for SEO. Add an admin interface where I can see submissions. Deploy it through Vercel.

Then I can iterate.

Add this.

Remove that.

This doesn't work on mobile.

Change the database structure.

Add authentication.

Create another tool using the same design system.

Pull this data from an API.

Make these URLs crawlable.

Fix the error.

That is fundamentally different from searching through a plugin marketplace for functionality somebody else decided to create.

Instead of finding software that approximately matches what I want, I'm increasingly able to create software that matches what I want.

That could be one of the biggest challenges WordPress has faced.

The Plugin Model Could Be Particularly Vulnerable

WordPress's plugin ecosystem is one of its greatest strengths.

It's also one of the areas AI could disrupt most heavily.

There are plugins for practically everything.

Forms.

SEO.

Redirects.

Tables.

Calculators.

Membership systems.

Image optimisation.

Directories.

Custom fields.

Analytics.

Popups.

Bookings.

Reviews.

Security.

Backups.

Social sharing.

Schema.

A Plugin Isn't Just Code

This is where some of the more extreme predictions about AI get carried away.

A mature WordPress plugin isn't valuable simply because somebody wrote PHP code.

A good commercial plugin may include:

years of testing;
thousands of edge cases;
security patches;
browser compatibility work;
integrations;
accessibility improvements;
documentation;
customer support;
updates when WordPress changes;
compatibility with other plugins;
payment handling;
regulatory requirements;
performance optimisation.

An AI might replicate the visible functionality of a plugin surprisingly quickly.

That does not automatically mean it has recreated everything that makes the plugin dependable.

This becomes particularly important when money, personal information or business-critical systems are involved.

I would be much happier experimenting with AI-generated code for a small interactive content tool than blindly trusting it to create a payment system handling thousands of transactions.

That difference matters.

AI makes code cheaper.

It doesn't make risk disappear.

Even the Person Who Popularised 'Vibe Coding' Made That Distinction

The term vibe coding was popularised by AI researcher and former OpenAI founding member Andrej Karpathy in early 2025.

His original description captured the feeling remarkably well: using AI so heavily that you can almost "forget that the code even exists."

Embedded post

View post on X →

The important part is that Karpathy wasn't claiming this was automatically the correct approach for every production system.

There is a significant difference between:

"Can AI make this work?"

and:

"Would I trust this to run an important business?"

Developer and Django co-creator Simon Willison makes a similar distinction.

His rule for serious AI-assisted development is simple:

"I won't commit any code to my repository if I couldn't explain exactly what it does to somebody else."

His full piece, Not all AI-assisted programming is vibe coding, is worth reading because it separates rapid AI experimentation from responsible production software development.

That resonates with my own experience.

The more I've built, the more I've realised that some technical knowledge still matters.

Possibly more than before.

AI Is Teaching Me Development While Simultaneously Making Development Easier

There is a strange contradiction here.

AI reduces how much programming knowledge I need to create something.

But by creating more things, I've ended up learning considerably more about programming and infrastructure.

Before using Vercel, I didn't particularly need to understand its deployment model.

Before connecting APIs, I didn't need to think much about API authentication.

Before using databases directly, I didn't need to understand schemas and data relationships in the same way.

Before working with GitHub regularly, commits and repositories weren't a central part of how I worked.

Before building custom applications, environment variables weren't something I was constantly thinking about.

AI allowed me to enter these areas without having years of formal training first.

Then experience started filling in the gaps.

I think millions of people are going to follow some version of that path.

They won't necessarily become conventional developers.

They may not be able to sit down and write an application from scratch without AI.

But they'll understand enough to direct AI.

That's a new category of builder.

And This Is What Worries Me Slightly

I am proud of what I've learned.

Genuinely.

There are things I've built recently that I would previously have considered completely outside my abilities.

And there's something incredibly satisfying about that.

But it creates an obvious question.

How long is this an advantage?

Today, knowing how Vercel works, understanding environment variables, connecting APIs, using GitHub and building database-backed applications with AI still feels reasonably advanced compared with the average website owner.

What happens when AI becomes another ten times better?

Will someone with absolutely no technical knowledge simply say:

Create a full SaaS platform for estate agents. Add authentication, subscription billing, a CRM, AI property descriptions, an admin dashboard, automated emails, a PostgreSQL database, analytics and SEO landing pages. Deploy it and buy me an appropriate domain.

And twenty minutes later it's live?

I think we're moving towards some version of that.

That's slightly uncomfortable when you've just spent a lot of time learning how everything fits together.

But I've come to a different conclusion about what that means.

Knowing how to physically create the software may become less valuable. Knowing what software should exist becomes more valuable.

The Bottleneck Moves Upwards

This may be the most important consequence of AI development.

Historically:

Idea → developer → code → product

The developer was a major bottleneck.

Increasingly:

Idea → AI → product

So where does the bottleneck go?

It moves.

Towards:

identifying worthwhile problems;
understanding users;
product judgment;
differentiation;
design;
trust;
distribution;
marketing;
SEO;
customer acquisition;
data;
brand;
reputation;
business models;
execution.

AI might allow ten million people to build SaaS applications.

It doesn't mean ten million people suddenly know which SaaS application customers want.

Canva made graphic design accessible.

That didn't make everybody an exceptional designer.

Shopify made ecommerce infrastructure accessible.

That didn't make every Shopify owner a successful retailer.

WordPress made publishing websites accessible.

That didn't make every WordPress site successful.

AI development will follow the same pattern.

The baseline rises.

The differentiator moves somewhere else.

Chris Lattner, creator of LLVM and Swift, put this particularly well when discussing AI-generated software:

"As writing code is becoming easier, designing software becomes more important than ever."

His wider analysis of what AI-generated software means for the future of engineering argues that implementation is increasingly being automated while judgment, abstraction and stewardship become more important.

That's very close to where I've landed myself.

So Does This Mean WordPress Is Finished?

No.

And anyone claiming WordPress is about to disappear is ignoring the enormous advantages it still has.

At the time of writing, WordPress remains responsible for roughly 40% of websites globally.

That means an enormous number of:

businesses;
developers;
agencies;
publishers;
plugins;
themes;
hosts;
integrations;
ecommerce stores;
contributors;
documentation pages;
tutorials;
communities.

Technology doesn't normally vanish simply because something better appears.

Legacy and network effects matter enormously.

Businesses aren't going to rebuild functioning WordPress sites simply because someone demonstrates an impressive AI-generated Next.js application on X.

Nor should they.

If a plumbing company needs five pages, a contact form and an occasional blog post, WordPress may still be a perfectly sensible solution.

There is no prize for using the most complicated technology possible.

Sometimes boring technology is exactly what you want.

WordPress Has Another Huge Advantage: It Can Use AI Too

Perhaps the most interesting part of this entire argument is that WordPress isn't standing still.

WordPress 7.0, released in 2026, introduced AI infrastructure directly into WordPress core, including an AI client designed to connect WordPress with generative models.

WordPress.com has gone further.

Its AI website builder now allows somebody to describe a website conversationally and have WordPress generate the initial design, layout, images and content without requiring traditional coding or even the usual template-selection process.

That is fascinating.

Because WordPress understands the threat.

AI doesn't necessarily replace WordPress.

AI could become the new interface to WordPress.

Instead of:

Dashboard → Appearance → Themes → Editor → Plugins → Settings

the interface becomes:

"Make the homepage more modern."

"Create a landing page for this service."

"Add a booking feature."

"Show testimonials underneath this section."

"Create an SEO page targeting this service in Birmingham."

"Change every CTA button across the website."

WordPress.com is already moving in this direction with AI-assisted site creation and editing.

If WordPress can successfully hide more of its historical complexity behind an intelligent conversational layer, it might actually become easier to use rather than obsolete.

WordPress Could Become an Operating System for AI-Built Websites

This is one possible future I find particularly interesting.

Perhaps people stop thinking of WordPress primarily as a visual CMS.

Instead, WordPress becomes infrastructure.

AI sits above it.

The user talks to an agent.

The agent manipulates WordPress.

WordPress handles:

users;
permissions;
content;
media;
revisions;
ecommerce;
plugins;
APIs;
publishing;
data.

The human barely sees the traditional dashboard.

In that scenario, WordPress survives because its enormous ecosystem becomes something AI can operate.

WordPress itself is already exploring this direction through AI functionality, its Abilities API, connectors and work designed to let AI systems interact with WordPress capabilities in structured ways.

That's a much more compelling future than simply bolting a text generator onto the WordPress editor.

But Headless CMS Platforms Suddenly Make Much More Sense to Non-Developers

This is where my experience with Sanity changed my thinking.

A few years ago, if someone had told me:

Use a headless CMS and create a separate frontend application deployed through a cloud platform.

my first reaction probably would have been:

Why?

WordPress already gives me the frontend and the CMS.

A headless system creates more development work.

And historically, that was true.

You would need somebody capable of creating that frontend.

AI dramatically changes the calculation.

If AI can create much of the frontend for me, separating the content system from the presentation layer becomes far less intimidating.

Sanity can manage structured content.

The frontend can be built specifically for the project.

APIs can connect additional services.

A database can handle application data.

Vercel can handle deployment.

GitHub can hold the code.

AI can help connect everything.

Suddenly I'm no longer asking:

"Which WordPress theme is closest to what I want?"

I'm asking:

"What architecture makes the most sense for what I'm building?"

That's a much healthier question.

Websites Are Becoming Software

This is another reason I think traditional CMS platforms face pressure.

The distinction between a "website" and a "web application" is becoming increasingly meaningless.

A modern website might include:

calculators;
AI assistants;
recommendation engines;
customer dashboards;
personalised content;
interactive maps;
account systems;
databases;
document analysers;
live external data;
API integrations;
generators;
comparison engines;
automated workflows.

I've written much more about this shift in Why AI Is Changing How We Build Websites.

The future website won't necessarily just tell you things.

It will do things for you.

That's important because WordPress originated in a publishing world.

Its core mental model is content.

Posts.

Pages.

Authors.

Categories.

Media.

That model remains excellent for publishing.

But an increasing number of websites need an application model rather than simply a publishing model.

AI makes those custom applications much easier to build.

SEO Is Another Reason I'm Interested in This

My background in SEO makes the development side particularly interesting.

One thing I've realised is that custom development allows SEO to become part of the product itself.

Instead of simply creating another 2,000-word article, a website can create something genuinely useful.

An interactive calculator.

A database.

A comparison tool.

A generator.

A searchable resource.

An industry benchmark.

A free utility.

A personalised report.

Those things can produce links, searches, repeat usage and brand recognition in ways another generic blog article may struggle to achieve.

AI makes them much cheaper to build.

That's huge.

The internet is also changing beyond conventional Google search, something I explored in The Internet Has Changed More in the Last Three Years Than the Previous Fifteen.

We're now building for humans, search engines, AI search interfaces, social discovery and potentially autonomous agents.

The humble website is becoming considerably more complicated.

Paradoxically, AI is making that complexity easier for individuals to manage.

The £5,000 Website Is Probably Under More Threat Than WordPress Itself

If I were looking for the immediate casualty, I wouldn't actually pick WordPress.

I'd pick the traditional basic website development model.

The familiar proposition:

Five-page business website — £3,000.

Homepage.

About.

Services.

Blog.

Contact.

Built from a theme or design system.

That business model is going to become increasingly difficult to defend.

Not immediately.

There will still be businesses that want an agency to manage everything.

People pay for reliability, expertise and convenience, not merely code.

But the production cost is heading down rapidly.

If somebody can describe their company to AI and receive a genuinely attractive, responsive, SEO-friendly website within an hour, the perceived value of manually constructing that same site falls.

Agencies will need to move higher up the value chain.

Strategy.

Positioning.

Conversion.

Content.

SEO.

Integrations.

Automation.

Custom tools.

Analytics.

Distribution.

Maintenance.

Security.

Business outcomes.

The website itself becomes less of the product.

The result the website produces becomes the product.

WordPress Agencies May Actually Become More Powerful

This is why I don't think AI automatically destroys agencies either.

Imagine an experienced WordPress developer who already understands:

PHP;
JavaScript;
databases;
security;
hosting;
caching;
WooCommerce;
APIs;
WordPress internals.

Now give that person extremely capable coding agents.

They may not become redundant.

They may become ridiculously productive.

One person might produce what previously required three.

An agency might handle far more custom development without increasing headcount.

Plugins could be customised much more deeply.

Technical problems could be diagnosed faster.

Migration work could become easier.

Old code could be modernised.

Tests could be generated.

Documentation could be improved.

Mullenweg himself has argued that AI can give motivated technical people significantly more leverage rather than simply making agencies obsolete.

That seems plausible to me.

The pressure will be greatest on people whose entire value was performing repeatable production tasks.

AI Doesn't Eliminate Technical Knowledge — It Changes Which Knowledge Matters

This is something I've become increasingly convinced about.

I don't need to memorise every command.

I don't need to remember exact syntax.

AI can do that extremely well.

What helps me enormously is understanding concepts.

Knowing what an API is.

Knowing why an API key shouldn't appear publicly.

Knowing what an environment variable is.

Knowing that a frontend and database shouldn't necessarily communicate however the AI first suggests.

Knowing what GitHub is doing.

Knowing where Vercel fits.

Knowing the difference between content stored in Sanity and application data stored in a database.

Knowing why authentication matters.

Knowing why a page might not be indexable.

Knowing enough to recognise when something doesn't look right.

AI can increasingly handle the how.

Humans still benefit enormously from understanding the what, where, when and why.

Research into vibe coding has reached a similar conclusion. A 2026 literature review found that development work shifts away from manually writing every line and towards specification, supervision and validation, while evidence for long-term maintainability remains much less certain than the obvious gains in prototyping speed.

That sounds very close to what I've experienced.

The Real Danger Is False Confidence

The scary scenario isn't necessarily somebody knowing nothing about development.

It's somebody knowing nothing about development while believing AI means they no longer need to know anything.

AI is incredibly convincing.

It will happily tell you something has been fixed.

Sometimes it has.

Sometimes it hasn't.

It can create something visually impressive that has terrible architecture underneath.

It can expose credentials.

It can create security vulnerabilities.

It can duplicate logic.

It can misunderstand a database.

It can fix one bug and create another.

It can build an application that works beautifully for five users and collapses with 5,000.

The interface becoming simple doesn't mean the underlying system became simple.

Cars became much easier to drive without becoming mechanically simple.

Computers became much easier to operate while becoming vastly more technically sophisticated.

Web development may follow the same route.

The complexity doesn't disappear.

It moves underneath the interface.

So What Happens When Anyone Can Build an Advanced Website?

This is the question I'm most interested in.

I suspect we're heading towards a world where building a technically impressive website stops being impressive by itself.

That sounds strange today.

But imagine everyone can build:

a custom SaaS;
an AI assistant;
a directory;
a marketplace;
a dashboard;
a recommendation engine;
a booking platform;
an ecommerce site;
a mobile application.

The existence of the software stops being the achievement.

Then the questions become:

Does anyone care?

Is it actually useful?

Can people find it?

Do they trust it?

Will they pay for it?

Will they return?

Can you operate it profitably?

Can you improve it faster than competitors?

This is why I'm increasingly less worried that AI will make the things I've learned worthless.

The technical knowledge I've developed isn't valuable because I can type certain commands.

It's valuable because I now have a better mental model of what can be built.

And AI keeps increasing that possibility space.

WordPress's Biggest Threat Is the Blank Canvas Becoming Less Scary

For years, WordPress had a powerful advantage:

Starting from nothing was hard.

So you didn't start from nothing.

You started from WordPress.

You chose an existing theme.

You installed existing plugins.

You used established patterns.

AI changes the economics of starting from nothing.

The blank repository isn't quite as intimidating when an intelligent coding system is sitting next to you.

A custom frontend isn't as intimidating.

A database isn't as intimidating.

An API isn't as intimidating.

GitHub isn't as intimidating.

Deployment isn't as intimidating.

That is the structural threat.

Not that AI builds a better blogging platform.

AI makes it possible to bypass the blogging-platform model entirely.

My Prediction: WordPress Won't Die — But Its Role Will Change

I don't think we're witnessing the end of WordPress.

At least not anytime soon.

Its installed base is far too large.

Its ecosystem is far too mature.

Its community is far too extensive.

Its use cases are far too broad.

And importantly, WordPress itself is adopting AI aggressively.

What I think we could be witnessing is the end of an era in which WordPress is the automatic answer to:

"I need a website but can't develop one myself."

That was one of its greatest advantages.

AI weakens it.

The decision increasingly becomes:

Do I actually need a CMS like WordPress for this?

Sometimes the answer will absolutely be yes.

For a traditional content-heavy website, WordPress may remain one of the best options available.

For WooCommerce businesses, publishers and organisations embedded deeply into the ecosystem, abandoning WordPress because AI is fashionable would often make little sense.

But for a new project?

Especially something involving custom functionality?

I'm personally much more willing to look elsewhere than I would have been a few years ago.

That's entirely because AI has expanded what I can build.

WordPress May Have to Become Invisible

The winning version of WordPress in an AI-first web may actually be one users interact with less.

AI could become the interface.

WordPress becomes the infrastructure underneath.

The average business owner doesn't need to know what block, plugin, database table or template controls something.

They simply say what they want.

The agent works out how to accomplish it safely.

If WordPress successfully makes that transition, its massive ecosystem becomes an extraordinary advantage.

If it doesn't, newer platforms built from the beginning around AI agents, APIs and structured data could gradually take more new projects.

That battle may be much more important than whether WordPress adds another feature to the block editor.

The Future Probably Isn't WordPress Versus AI

It's tempting to frame technology shifts as battles.

WordPress versus AI.

Developers versus AI.

No-code versus code.

Traditional CMS versus headless CMS.

I increasingly think that's the wrong way to look at it.

AI becomes a layer across everything.

WordPress uses AI.

Sanity uses AI.

Developers use AI.

Designers use AI.

SEO professionals use AI.

Entrepreneurs use AI.

Hosting platforms use AI.

GitHub uses AI.

The underlying technologies continue to exist.

What changes is how humans interact with them.

And that may ultimately be the biggest shift of all.

Twenty-Five Years Online and This Feels Different

I've seen plenty of things described as revolutionary.

Some were.

Many weren't.

I've seen website builders, CMS platforms, mobile apps, social media, cloud computing, no-code tools, cryptocurrencies, NFTs and countless new platforms arrive with predictions that everything before them was finished.

Usually reality is less dramatic.

Old technologies survive.

New technologies become part of the stack.

People adapt.

But AI does feel different to me for one particular reason.

It isn't simply giving me another platform to use.

It's expanding what I'm capable of doing across almost every platform.

That's a much deeper change.

I couldn't previously build some of the things I'm building now.

Not because the technologies didn't exist.

Vercel existed.

GitHub existed.

Databases existed.

APIs existed.

Headless content systems existed.

The barrier was my ability to combine them.

AI lowered that barrier.

That is why I think its impact on website development will be enormous.

The Irony Is That WordPress Started the Same Revolution

There is something wonderfully circular about all of this.

WordPress succeeded because it enabled people to publish online without needing to become programmers.

AI is succeeding because it enables people to create software without necessarily becoming programmers.

Both movements reduce barriers.

Both allow more people to create.

Both threaten established specialists.

Both initially produce plenty of terrible websites.

Both also enable people with genuinely good ideas to build things they previously couldn't.

So perhaps AI isn't philosophically the opposite of WordPress at all.

Perhaps it's the continuation of exactly what WordPress started.

WordPress democratised publishing.

AI may democratise software development.

The question is whether WordPress becomes one of the foundations of that next era — or one of the platforms it eventually leaves behind.

Where I Stand Today

Would I still use WordPress?

Absolutely.

Would I automatically use WordPress for every new website?

No.

That's the change.

Five or ten years ago, WordPress would have been the starting point for many of my ideas.

Today I start with the idea itself.

Then I ask:

What is the best way to build this?

Maybe it's WordPress.

Maybe it's Sanity.

Maybe it's Next.js.

Maybe it needs a database.

Maybe it needs an API.

Maybe it needs almost no CMS at all.

Maybe AI can help me create something far more tailored than any existing template or plugin could provide.

That freedom is exciting.

And slightly frightening.

Because I suspect we're only at the beginning.

The things that feel advanced to me today may feel completely ordinary in a few years.

Someone with no knowledge of GitHub, databases, APIs or deployment may eventually be able to create an application simply by talking to an AI.

But that doesn't necessarily make learning those things pointless.

If anything, I think the advantage shifts from being the person who knows how to type the code to being the person who understands what all the pieces can do together.

And beyond that, to the person who knows what is actually worth building.

That is where I think the real competition is heading.

So no, I don't believe WordPress is about to die.

But I do believe something fundamental is happening to the problem WordPress originally solved.

And if sophisticated custom software becomes as easy to create as a WordPress website once was, the next era of the web could look very different indeed.