Why Some Projects Demand Human Hands: The Case Against AI in Specialized Web Development

Every few months, a job posting surfaces that makes developers pause. This time, it landed on Hacker News: a straightforward request for a web developer to build a math website, with one non-negotiable requirement—absolutely no AI tools. No Copilot suggestions, no ChatGPT rewrites, no AI-generated code. Just human craftsmanship.

At first glance, this might seem like an extreme stance. After all, AI-assisted development has become standard practice across the industry. Tools like GitHub Copilot, Cursor, and ChatGPT have dramatically accelerated how we write code. But here's why this particular request deserves serious consideration.

Math demands precision in ways ordinary web apps don't. When you're building educational software for mathematics, the logic must be airtight. A minor error in how an equation renders or how a proof is structured could confuse students or worse—teach incorrect concepts. Human developers who deeply understand mathematical notation and pedagogy bring something that current AI tools simply cannot replicate: genuine comprehension of the domain.

The "vibe coding" trap. The tech world has embraced "vibe coding"—writing software by describing what you want to an AI and iterating on the results. It's powerful for prototypes and MVPs. But vibe coding tends to produce code that works without understanding why it works. For specialized projects requiring long-term maintainability, educational integrity, and deep integration with mathematical concepts, you need developers who can truly own and explain their codebase.

Trust and accountability matter. When a developer writes every line themselves, the work carries accountability. You know exactly what decisions were made and why. When AI tools are layered in, debugging becomes complex—did the human introduce an error, or did the AI generate faulty logic? For educational platforms where correctness is non-negotiable, this traceability is invaluable.

This isn't an attack on AI tools. They're genuinely useful for many projects. But specialized domains—whether mathematics, medical software, legal platforms, or creative work requiring specific aesthetic vision—sometimes demand human hands that understand both the craft and the context.

The HN post sparked dozens of responses from developers willing to take on exactly this kind of project. It was a reminder that for certain work, the human element isn't a limitation—it's the point.

If you're building something that requires domain expertise and absolute precision, finding a developer who shares your commitment to craft over convenience might be the best decision you make.

What do you think? Is requesting "no AI" a reasonable requirement, or is it fighting progress? We'd love to hear your perspective.