Erlang Meets Idris: Cure Programming Language | Hacker News Reader

Erlang Meets Idris: Cure Programming Language

cure-lang.org

46 pointsdelitrem33 comments

This is 100% LLM generated; website, documentation and tutorials. There is no link to downloads or a repository. No way to use anything.

Why should anyone care about this?

I was about to say the same thing haha

There is a github repos and that + code looks also LLM generated to me. Not necessarily bad, if it works for what was intended that is; I just don't have time/patience to try it because of how lazy their web page is. I mean LLMs can DEFINITELY make a lot better pages than this; this what you get if you do it one-shot and publish.

The idea may be good and result may be functional but regarding adoption, especially for the domain it targets, for someone to depend on what appears to be vibe coded project is irrational.

> This is 100% LLM generated

Who knows, may be you are right here. I actually thought so at first, but knowing the author personally (he is my former colleague, I had the pleasure of working with him in the same team about 17-18 years ago), his extraordinary abilities and his writing style even before the widespread use of AI, I had my doubts.

Emojis make it look LLM af.

Emojis at the end of a statement online are a generational thing, not an AI thing.

Replying to an email inline rather than at the top marks you out as of a certain generation. Using text emojis rather than finding the graphical emoji does too.

Everyone needs to relax about AI generation anyway (did you learn something useful or not? If you did, does it matter if it was AI generated as a site?), but saying "this is what people under 30 frequently do, so it must be fake", is just this weird vibe spreading everywhere I don't get at all.

emoji at the end of a statement are not the same thing as emoji adorning or replacing every heading

I'm talking about this kinda style...

* <Arrow hitting target emoji> 15 compiled libraries!

* <green tick> Works on my machine

* <red cross> No ARM support.

None of which are at the end of a statement. So, I'm not sure who you're replying to.

Incidentally, I recently reviewed a PR heavily written by Cursor that had statements like this.

    logger.info("<magnifying glass emoji> DEBUG: {actual message")

And then CursorBot reviewed it and flagged the emojis as indicative of "debugging statements not suitable for production".

Which made me laugh, loudly, and only somewhat sadly, Cursor added the emojis, Cursor then flagged them as not appropriate in prod code.

But CursorBot missed the obvious problem with

    logger.info("DEBUG: ...")

It isn't a generational thing. The choice of emoji is a generational thing, but people of all ages do it. AI most certainly does not use emoji in the same way a young person does (unless you encourage it to, but even then it comes across as cringeworthy). If anything it's closer to how a middle-aged person uses them.

I'd also say the use of text emoticons has all but died out in anything other than ironic usage, or in situations where it's difficult to use unicode emoji (e.g. games or this very site)

When text is very obviously generated by AI it communicates to the reader that there is nothing of value to be read. It always writes in the same vapid, overly enthusiastic, overly verbose way. It's grating and generally conveys very little information per word. It's a cliché at this point, but if nobody bothered to write it then why would I bother to read it?

There is a 0% chance that the vast majority of this site and the repo that was linked elsewhere was written by a human. I would have zero confidence in anything about this language, and frankly your former colleague should be embarrassed about putting this out.

Edit: I just noticed in another comment: "Perfect for : Trading systems, industrial control, Medical devices, aerospace applications". I'd go further than embarrassed, and say this person should be ashamed of themself and take this down.

I'm sorry but this has all the earmarks of being AI generated, at the very least the website and all the project documentation; and "Trust me I worked with him 2 decades ago, is a very poor argument to inspire confidence"

The emoji list is so in your face I'm leaning towards it being a parody or some kind of art piece.

> Perfect for : Trading systems, industrial control, Medical devices, aerospace applications

Regulator, here is some code in an unknown and poorly documented language with no operational experience. The compiler was written using AI and no one has audited it.

That seems like an excellent idea to me.

This is the GitHub repo: https://github.com/am-kantox/cure-lang

  Compilation Performance
  
    Small files (<100 lines): <1 second
    Medium projects (1K-10K lines): 5-30 seconds
    Large projects (100K+ lines): 30-300 seconds with incremental compilation

Love that there's an upper limit on compilation time. No matter how large your project gets, it will never take more than five minutes to compile (incrementally).

Also it’s not possible to write programs that have between 100 and 1000 lines.

Super exciting. Can't wait to use this in production. Imagine, using AI to write with a language built with AI, building AI products that AI people use.

Curious what the E, e, e, L and G stand for in the logo.

My money's on L = LLM, G = Generated

This is not a real language, it's pure LLM slop.

Just look at the so-called sort example from the repo:

    def sort(list: List(T)): List(T) where Ord(T) =
        match list do
            [] -> []
            [pivot | rest] -> sort(rest)
        end

that will achieve incredible performance on the right array

It only works correctly on an empty array, on which any sorting algorithm is fast.

https://cure-lang.org/examples/ gives a 404

The project looks very young. I do like the goals of the project though, and I like that it's on the BEAM.

It absolutely drives me nuts when people spend so much time building something but make it difficult to show you what they’ve built.

A short code snippet (with syntax highlighting thank you) should be the first thing on your page.

I do not have to scroll through a huge wall of text (probably AI generated), 2 images (definitely AI generated), miss it, start clicking links, still not find it, hit the back button, scroll through the slop again, etc.

I want to see the thing, I don’t care about what you have to say about the thing until I can get a sense of the thing.

> when people spend so much time building something

I do not think that much human time was spent on this actually.

I would like to see some interesting code examples showcasing the main features.

How does this compare to Gleam, in terms of goals, features, etc.?

> A strongly-typed, dependently-typed programming language that brings mathematical correctness guarantees

Regarding goals, from a quick check on both, the essential difference is Cure has dependent types with SMT-backed validation. So, as mentioned in homepage, is oriented towards domains requiring correctness over convenience, whereas Gleam targets general development. (Beyond goals for anyone that hasn't heard Gleam before, Cure appeared out of nowhere recently and seems like AI slop, Gleam exists for few years and people are using it to make actual projects.)

Everything smells of AI here, is it the world's first slop language?

please keep the erlang ecosystem out of the llm griftosphere. jesus christ.

Reply on news.ycombinator.com