Wait wait - your argument is that lower-level languages, having too many footguns, will be written "perfectly" by LLMs? Seems to assume a LOT of growth in the quality of LLM output.
Also, how does this perspective help for (A) existing applications, or (B) frontend applications, which are almost never provided via low-level lanaguages (think web apps using JS, or Android apps using Java/Kotlin)?
Also, how does this perspective help for (A) existing applications, or (B) frontend applications, which are almost never provided via low-level lanaguages (think web apps using JS, or Android apps using Java/Kotlin)?