Hacker News new | past | comments | ask | show | jobs | submit
Gareth321 | on: AI is forcing us to pay off 40 years of technical debt
I note you're downvoted but I strongly agree. Forcing humans to speak machine language is inherently inefficient. Having them speak a language we're already comfortable with is an obvious evolution in abstraction and control. I suspect you've ruffled some feathers because speaking machine language has been a knowledge and skill gate which is rapidly becoming redundant. It doesn't feel good when one's career and skillset become much less valuable. Never-the-less, this technology is not going anywhere. Opus 5.5 already codes better than most of my developers at about 100x the speed. Architecture is still valuable, but I have no doubt that this is also going to be done better by AI soon.
Izmaki | parent | next
Some people react to harsh realities and truth with negativity and the only the only things they have control of here on HN: a downvote. :)

I, personally, never downvote people I disagree with. Who am I if I don't use another perspective as an opportunity to learn and grow? I can tell that many people agree, too, because my "comment score" is still mostly neutral at the time of writing.

tremon | parent
Human speech is inherently inefficient and ambiguous. Humans rely on non-verbal clues and context to distill the exact meaning of phrases, and even then the potential for miscommunication is non-negligible. It is a huge leap to assume that artificial intelligence can succeed in flawlessly divining intent from regular human speech to the point where it can generate debt-free code. And if your solution to that is to say "no, you need to issue specific instructions to guide the machine to the correct solution" then you have not removed the skill requirement, you have merely redefined the skillset.
Gareth321 | root | parent
I think you mistakenly believe that the existing abstraction layers don't already contain ambiguity. It's true that formal languages are built to reduce ambiguity, but there is still much ambiguity. For example, gaps in the specification still leave room for interpretation. [They list some in Cambridge's free C course:](https://www.cl.cam.ac.uk/teaching/1516/CandC++/lecture10.pdf)

- Implementation-defined: the compiler documents its choice, e.g. sizeof(int).

- Unspecified: the compiler picks from options, e.g. argument evaluation order.

- Undefined: no requirements at all, e.g. signed overflow.

The notes take you through compilers deleting an overflow check and a kernel null check because of undefined behaviour.

There are also abstraction leaks. Spolsky’s Law of Leaky Abstractions says all non-trivial abstractions leak. SQL is one example. Logically equivalent queries can differ in speed by thousands of times.

Also, the translator must be trusted. [See this lecture.](https://www.cl.cam.ac.uk/teaching/2223/R209/Reflections-Trus...) It explains how behaviour can live in a compiler binary and appear in no source.

In other words, formal languages are no different to spoken languages. They exist on a spectrum of ambiguity, assumption, and implied meaning, depending on the domain. Human language has the potential to introduce more ambiguity, assumption, and implied meaning, but that's not a predetermined quality. You can be as detailed as you like in your communication with your AI. To a degree of precision equal to or better than a formal language. You reference this in your last sentence. The distinction is fairly obvious: being specific in English has a much lower skill bar than being specific in C.

One of the more interesting topics in research and philosophy right now is this question of alignment. To what degree to these models understand our meaning when we're not extremely detailed. We can see in research that they're becoming extremely competent at inferring meaning. They can often understand our intent with very few words.