This argument is predicated on what might become an outdated idea of software as an asset. If I can quickly generate software from natural language to solve a very specific problem, that software isn't worth maintaining, let alone publishing or selling. Its value to people who aren't me is low, and its defensibility against being copied by someone else with an adequate coding agent is even lower.
> If I can quickly generate software from natural language to solve a very specific problem
This isn't likely to happen -- if the problem is very specific, you won't be able to sufficiently express it in natural language. We invented programming languages precisely because natural languages are completely unsuited for the task of precisely specifying a problem.
So how are you going to express the specificity of the problem to the LLM in natural language? Try, and you'll discover their shortcomings for yourself. Then you'll reinvent programming languages.
This is a bit backwards. Formal systems engineering is rooted in defining problems in natural language. The programming languages are normally used for defining the solutions, not the problems. Even if you're taking about TDD, the executable tests are still derived from natural language test case specifications.
