It is a misconception that I do not use AI for development. On the contrary, I have studied AI security extensively and analyzed its pitfalls to produce the most secure and efficient solutions possible for my clients. My approach is not to reject AI outright, but to use it safely, deliberately, and intelligently—with full awareness of its risks.
Over the last few years, AI has dramatically changed the software landscape. In many ways, AI has rendered general-purpose frameworks increasingly redundant and made package management systems dangerously vulnerable.
The original purpose of general-purpose frameworks was to separate application layers—frontend, backend, and database—into logical components, often through object-oriented design patterns such as MVC (Model-View-Controller). These patterns were considered the gold standard for web, app, and even game development for over two decades. They helped teams divide responsibilities, improve code organization, and enable modularity.
Java was the language that popularized object-oriented programming (OOP), and later frameworks such as Spring, Laravel, Django, and Rails extended this paradigm into the web domain. As a web developer, my experience has primarily been with MVC frameworks, AI-assisted development, and the evolution of general-purpose frameworks over time.
However, I have never been a strict believer in OOP or MVC design patterns. In practice, these frameworks—though elegant in theory—often led to unnecessary abstraction, slower performance, and bloated production cycles.
Why I Still Favor WordPress Over AI “Vibe Coders”
I built a large part of my career on WordPress because it allowed me to rapidly develop complex, functional, and affordable websites by leveraging a vast ecosystem of plugins and minimal custom scripting. While many developers historically dismissed PHP and WordPress as “unscalable” or “unprofessional,” the numbers have consistently proven otherwise:
- WordPress powers over 43% of all websites and more than 73% of all server-side applications that use PHP.
- Facebook itself began on PHP and still maintains portions of its codebase in it.
- Among the top 1,000 most visited websites in the world, the majority are still powered by PHP. The nearest competitor, Node.js, holds only about 5% of the market share.
While critics have argued that WordPress is unsuitable for enterprise development, I and countless others have disproven that claim by building everything from ERP systems to fleet management platforms entirely within WordPress—at a fraction of the cost and development time of other frameworks.
The truth is that WordPress continues to evolve. Its simplicity, scalability, and extensive plugin ecosystem keep it ahead of the curve, especially when compared to AI-based “Vibe Coders” or closed app-building systems like Cursor.
Critical Arguments
1. Package Poisoning
One of the greatest security threats in the AI era is package poisoning—the insertion of malicious code into commonly used software libraries. LLMs (Large Language Models) often recommend or automatically integrate packages into your projects that have been compromised, either intentionally or through upstream dependencies.
AI-powered “Vibe Coding” tools such as Cursor can directly install and configure these malicious packages into your production environment, creating hidden backdoors, trojans, or remote-access exploits. Since AIs are trained on publicly available data, they are easily manipulated by data-poisoning techniques that teach them to trust and recommend these malicious libraries.
Package managers such as Composer (PHP) and npm (Node.js) have become particularly vulnerable. As a result, it is now considered best practice to avoid automated dependency installation whenever possible and manually verify every package before integration.
2. The Decline of Framework Necessity
General-purpose frameworks were originally designed to make development faster, safer, and more maintainable. In reality, they often became bloated and counterproductive. Layers of abstraction introduced unnecessary complexity, performance overhead, and dependency chains that now act as points of failure.
AI development has accelerated this decline. With “Vibe Coding” tools, developers can now generate fully functional scripts, modules, and APIs that run natively without the structural baggage of frameworks.
When used correctly—and by trained engineers who understand the underlying language and architecture—AI-assisted coding can produce lightweight, efficient code that performs better than framework-dependent solutions.
The guiding principle for modern developers should be:
Keep your code as close to native as possible.
Minimize layers, dependencies, and abstraction to maximize performance and control.
3. The Real Role of Vibe Coding
Vibe coding—using AI models to generate or assemble functional applications—has its place, but it must be approached with extreme caution. As someone who has studied AI systems design, cybersecurity, and AI exploitation, I only use vibe coding in controlled, auditable environments.
Laymen should never use Vibe Coders or AI-based app builders unsupervised. They lack the technical understanding to evaluate the generated code for vulnerabilities, inefficiencies, or hidden logic. In many cases, these users unknowingly introduce dangerous dependencies or expose their infrastructure to exploitation.
General-purpose frameworks, while far from perfect, were designed with some security and modularity in mind. AI-generated systems, however, can skip critical safeguards entirely.
4. Why WordPress Still Wins
WordPress remains one of the safest, most efficient, and flexible platforms available when used properly. Its ecosystem includes over 100,000 plugins, both free and premium, which provide nearly every conceivable functionality without reliance on external AI builders.
Vibe coding tools require highly specific technical prompts and strict specification design to be effective and safe. Most users cannot produce these accurately, leading to unstable or exploitable code.
In contrast, WordPress offers:
- Speed and simplicity – Fast deployment and rapid customization.
- Full code ownership – Complete control over your infrastructure.
- Security through transparency – Source code is auditable, not hidden behind AI layers.
- Extensibility – Thousands of plugins, APIs, and documentation.
While companies like Microsoft and Google are aggressively promoting AI app builders as “the future,” they are, in reality, Trojan horses—tools designed to centralize control over your data, intellectual property, and creative output.
Conclusion
AI is powerful, but it must be treated as a tool, not a replacement for human logic and craftsmanship. Developers who understand AI’s vulnerabilities can use it safely for code generation, optimization, and idea prototyping. However, uncritical adoption of AI-based app builders and frameworks can lead to catastrophic consequences for data integrity, security, and ownership.
WordPress, despite its critics, remains one of the last bastions of developer autonomy—a system where you control your code, your data, and your destiny.
The goal is not to abandon AI but to master it responsibly—to leverage its strengths while maintaining human oversight and security discipline.
