In a quiet corner of the technical archives, a debate from 1978 has resurfaced that cuts directly to the heart of the modern founder’s dilemma. Edsger Dijkstra, a foundational figure in computer science, long ago laid out a scathing critique of "natural language programming." As today’s operators increasingly lean on large language models to bridge the gap between human thought and machine execution, Dijkstra’s warnings from the University of Texas at Austin feel less like historical footnotes and more like urgent dispatches for the current generation of builders.

The core of the argument rests on the fundamental difference between human conversation and technical instruction. For decades, the industry has chased the dream of the "citizen developer"—the idea that if we could just describe what we want in plain English, the friction of software development would vanish. Dijkstra’s insight was that this friction is not a linguistic barrier, but a cognitive one. Natural language is inherently ambiguous, relying on shared context, cultural nuance, and the ability of the listener to fill in the blanks. Programming, by contrast, requires an unforgiving level of specificity. The machine does not possess common sense; it possesses only a set of instructions.

For the founders at the Numerous Times desk, this matters because it changes how we view the tools sitting on our desks. If we assume that the future of building is simply telling a machine what to do in vague terms, we risk losing the discipline required to understand what we are actually building. The move toward formal, symbolic languages wasn't an attempt to make life difficult for humans; it was an attempt to provide a tool sharp enough to carve out complex logic without the messy overlap of everyday speech. When you remove the rigors of a formal language, you don't necessarily make the task easier—you just make the errors harder to spot.

We are currently watching a massive industry pivot toward "prompt engineering" and conversational interfaces for code generation. While these tools are undeniable force multipliers for the experienced operator, Dijkstra’s perspective suggests a ceiling. The real work of a founder isn't just writing lines of code, but the rigorous definition of a problem. As we move forward, the most successful builders will likely be those who treat natural language as a starting point, but maintain the formal discipline to verify exactly what the machine has interpreted. The next phase of development won't be about eliminating the technical barrier, but about ensuring that our intent remains as precise as the logic it seeks to command.