41–50 of 118 posts
Earlier quoted context omitted.
The best way to get a strong answer is to post boldly incorrect statements. Then I get other people to write my actually-desired arguments for me. I call it, "wetware LLM prompt engineering".
And the wetware often refuses to do that and simply downvotes you. How dare they.
>Common answers to this question will include words like abstraction, performance, convenience, usability etc. The problem with these answers is that apart from the measurable, they are all subjective, aesthetic choices. That just is not true at all. These are all legitimate engineering tradeoffs, which any serious project has to balance. Calling this "aesthetics" is completely dishonest. These aren't arbitrary categ…
Agree, and we actually have both the standards and established methods to conduct representative tradeoff studies. But this knowledge is mostly ignored by CS and programming language designs. Even for Ada, there was little empirical evidence for specific design decision. A systematic survey discovered only 22 randomized controlled trials of textual programming language features conducted between the early 1950s through 2012, across six decades. This staggering scarcity explains why language designers rely heavily on intuition rather than evidence (see e.g. https://www.cs.cmu.edu/~NatProg/programminglanguageusability...).
Programming languages are obsolete in the LLM era. What current generation AI has revealed is that English is actually the ultimate representation of computer programs and systems, which is both sufficiently terse and precise to economically describe the operation of arbitrarily complex programs. There is no reason to study programming languages in 2025, other than as a historical curiosity - the same way one may stu…
Wouldn't LLM need to directly output machine code for that to be true?
Programming languages are obsolete in the LLM era. What current generation AI has revealed is that English is actually the ultimate representation of computer programs and systems, which is both sufficiently terse and precise to economically describe the operation of arbitrarily complex programs. There is no reason to study programming languages in 2025, other than as a historical curiosity - the same way one may stu…
> Programming languages are obsolete in the LLM era. Wouldn't LLM need to directly output machine code for that to be true?
[0] https://modelcontextprotocol.io/specification/2025-06-18/ser...
I still like Olin Shiver's take on this: https://www.ccs.neu.edu/home/shivers/papers/why-teach-pl.pdf
Earlier quoted context omitted.
Many new languages are still recycling ideas from the 1970's research labs. Outside affine types, all the praise for Rust's type system traces back to Standard ML from 1976. The heretic of doing systems programming in GC enabled programming languages (GC in the CS sense, including RC), goes back to research at Xerox PARC, DEC and ETHZ, late 1970's, early 1980's. Other things that we know, like dependent types, effect…
It seams like most things boil down to ideas from the 70's. The internet, distributed computing, AI, etc...
I still like Olin Shiver's take on this: https://www.ccs.neu.edu/home/shivers/papers/why-teach-pl.pdf
I wish even only half the OOP world actually understood it as the above.
(2022) This is particularly important here as the essay makes no mention of LLMs or coding agents (which were still in their infancy in development environments; this article is post original copilot/codex but pre ChatGPT).
No actually. Why is that important? I dont quite see why that is relevant. Could you elaborate?
> Usability is similarly ill-defined and hard to measure. Human factors are very well studied and standardized, and there is a well-established discipline called "Human Factors Engineering", which also provides established test and evaluation methods. Human Factors research is considered solid and well-established because it has been built on rigorous experimental psychology and engineering principles developed over…
(2022) This is particularly important here as the essay makes no mention of LLMs or coding agents (which were still in their infancy in development environments; this article is post original copilot/codex but pre ChatGPT).
Edit: I assume this comment gets downvoted because people don't like where we are heading, not because they really think LLM programming capabilities won't continue to improve at a staggering pace.