Tell HN: I'm tired of formulaic, "LLM house style" show HN submissions
> Hi HN, I have a history with some sort of problem — the common one which I'll describe here with some flowery prose that possibly makes gratuitous use of "quotation marks". I realized I was going some thing that reflects my fundamental limitations as a human. This demonstrates that something is wrong with our modern world, although the proliferation of AI slop is definitely not that thing.
> [Now something like:]
> To fight back, I built Software Name. It’s not just another program of the sort you might logically expect. It's something else, designed to implement a puffed-up description of solving the problem.
> [or perhaps:]
> I built hxxps://heck.no, an arbitrarily-described tool that helps you deal with the problem, instantly. You can interact with it in an intuitive way, and an AI does something for you.
> How it works is based on a few key principles from the problem domain:
> Formatting for Emphasis: Text that would make more sense as free-form prose is re-worked into a bullet-point list. Since each item includes multiple sentences, however, this doesn't actually reduce the total amount you'll be expected to read.
> Vague Explanation: This list contains exactly three items which seem generally aimed more at justifying how the product works rather than actually clarifying what will happen when you use it.
> Occasional Incoherence: LLM-generated lists of this sort often include items that don't all seem to belong in the same natural category. This confuses the reader and violates https://en.wikipedia.org/wiki/Parallelism_(grammar).
> The Tech (this would probably have some kind of emoji if HN supported them):
> It's built as fully buzzword-compliant software of some sort (which may be described with a word that doesn't quite fit in context), but some aspect of it is cool, interesting Other Buzzword which has some logical implication. Here's another logical implication. This might be several sentences that basically just mean "it's a web page where you can make an account, that is totally not going to crash if I suddenly get lots of business". Also I have no interest in ensuring you can see any content whatsoever without enabling JavaScript and I've probably vibe-coded a few off-the-shelf components together and launched it on Vercel.
> I wanted to share a personal story (and advertise my product) and thought others might find it useful too. I'm here to answer any questions about the problem domain, the tech, or the process of creating this thing. (I will remind you of this in a very formulaic way.)
> Would love to get your feedback!
> (Or perhaps I specifically "would love your feedback" — https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu... — on some pre-selected aspects of the project, rather than soliciting arbitrary critique.)
> Check it out here [if not previously linked]: hxxps://heck.no/
None of this makes me want to try out these projects, nor does it give me the sense as an HN user that I could actually, well, hack or experiment with a cool software toy. It makes me feel like I'm being marketed to and that I'm being asked to sign up for a product or service intended to solve a problem I'm unlikely to have or care about. The UX I expect as a "hacker" is very different from what ordinary people would expect as clients or end users.
I can assure you that when my projects are ready to appear in Show HN, my write-ups will read very differently from this formula.
>The UX I expect as a "hacker" is very different from what ordinary people would expect as clients or end users.
Although this site is called 'Hacker News', it doesn't imply that it is building apps for hackers; I think Show HN can also be about the making of things for non-hackers. This site is run by a VC.
>I can assure you that when my projects are ready to appear in Show HN, my write-ups will read very differently from this formula.
So what is the right way to write about projects? I've thought about writing something up that I'm building and it probably would have been similar to this structure.
This post makes me lean more toward 'why bother' and perhaps the community loses out on things in this way. Or maybe it is good if I don't write it up.
But another thing is that most tech people really undervalue good marketing and end up doing either the bare minimum (e.g., LLM puke) or narrowly targeting themselves as an audience. Either way, they often fail to get their message out to their market.
It's always nice to see something different. Your "Tell HN" (https://news.ycombinator.com/item?id=43496859) had a structure very similar to the one you're tired of (that one is below), so I'm curious what your breaking point was, and what you believe will be more effective.
What is it about LLMs that causes an excess of emoji in their outputs? I've noticed this to be one of the most reliable signals for discrimination.
that's the problem is that everyone wants to use AI, but nobody really does anything useful.
Well I’ve been working on an agentic app with a fun UI (https://aimakememoney.com) to turn complaints into startups ready for their first round of seed funding.
And yes, it was posted to HN. I don't remember the name though.