Now, however, we know how that played out in the case of assembly language. The fact of the matter is that only a very tiny fraction of software engineers give the structure of the compiled assembly code even passing thought. Our ability to generate assembly code is so great that we don't care about the end result. We only care about its properties...i.e. that it runs efficiently enough and does what we want. I could easily see the AI software development revolution ending up the same way. Does it really matter if the code generated by AI agents is DRY and has good design if we can easily recreate it from scratch in a matter of minutes/hours? As much as I love the craft and process of creating a beautiful codebase, I think we have to seriously consider and plan for a future where that approach is dramatically less efficient than other AI-enabled approaches.

Something I haven't seen talked about, though, is how powerful the type system is for constraining LLMs when using them to generate code. I was recently trying to get LLMs to generate code for a pretty vague and complex task in Haskell. I wasn't having much luck until I defined a very clear set of types and organized them into a very clear and constrained interface that I asked the LLM to code to. Then the results were much better!

Sure, you can use these same techniques in less strongly typed languages like Rust, and you can probably also use a similar approach in dynamically typed languages, but Haskell's pure functions allow you to create much stronger guard rails constraining what kinds of code the LLM can write.

https://www.ioccc.org/authors.html#Fabrice_Bellard

https://www.youtube.com/watch?v=j0XmixCsWjs

That being said, I think you're right about some of the challenges that an effort like this will encounter.