Show HN: Haystack – Review the PRs that need human attention

Haystack showed up on Hacker News promising a very tidy future: software changes get checked automatically from the moment they’re submitted, the system learns from repeated mistakes, and only the truly questionable stuff gets kicked to a real human. In plain English, it’s selling the dream of a robot reviewer that gets smarter over time and catches problems before they become office drama.

But the real action was in the comments, where the crowd immediately split into three classic internet camps: the skeptics, the nostalgic exes, and the branding police. One of the sharpest reactions basically asked, is this actually new, or is it just the same artificial intelligence tools everyone already has with a fancier wrapper? That’s the core challenge hanging over the launch: people like the idea, but they want proof it does more than tell a chatbot to wave down a human when things get weird.

Then came a surprise emotional subplot. One commenter mourned Haystack’s old canvas-based editor like a cancelled cult TV show, saying it was a joy to use and much better than ordinary tab-based tools. And finally, in the most Hacker News twist possible, someone popped in to note that “Haystack” is already used by several dozen things. So yes: the product promises calmer reviews, but the community reaction was a delicious mix of curiosity, side-eye, and “sorry, we liked the old you better.”