Comment by BananaPelican
4 days ago
I have a project that I'm hoping to launch on show HN in the next few days which was built entirely with the help of AI agents.
It's taken me about month; currently at ~500 commits. I've been obsessed with this problem for ~6 weeks and have made an enormous amount of progress, but admittedly I'm not an expert in the domain.
Being intentionally vague, because I don't want to tip my hand until it's ready. The problem is related to an existing open source tool in a particular scientific niche which flatly does not work on an important modern platform. My project, an open source repo, brings this important legacy tool to this modern platform and also offers a highly engaging visual demo that is of general interest, even to a layperson not interested in programming or this particular scientific niche.
I genuinely believe I have something valuable to offer to this niche scientific community, but also as a general interest and curiosity to HN for the programming aspects (I put a lot of thought into the architecture) as well as the visual aspects (I put a lot of thought into the design and aesthetics).
Do you have any advice on how to present this work in a compelling way to people who understandably feels as burned out on AI slop as you do?
Just my opinion, but if you present it in a way that first explains the problematic, then explain what other similar tool fail to solve, and have a genuine understanding of the tool in the sense that you can understand and answer questions to generate a discussion, it doesn't matter much how it was coded. The "slop" part of the AI really comes down to having a vague idea for a tools, barely doing any research if the problem has been solved and generating a tool that nobody asked for and that you barely know, there is not much room for an interesting discussion.