I've been chewing on this while building Ulsa, so I wanted to open it up here before we launch.
If you work in a fast-moving field, staying current is basically a second job. For years my routine was checking ten different sources every morning, newsletters, X, a pile of news sites, and I still felt like I was missing the things that mattered for my work while drowning in things that didn't.
I don t know if I can be convinced that AI doesn t intentionally sabotage code. Whether using Claude, Gemini, or ChatGPT, every vibe coder has hit that wall where the model seems to actively work against you on a bad day.
You know the drill:
Ridiculous function names: maybeSetConnectionState? Maybe? Come on.
Instant drift: Completely ignoring agreed-upon boundaries right after clarifying them.
Convoluted architecture: Shoving complex files into straightforward process chains, making them brutally difficult to navigate.
The CLI nightmare: Watching a command overwrite your entire codebase, followed by the AI gaslighting you about how to find a folder it just erased.
Hi Hunters, my name is Alfred and i'm launching on ProductHunt 17th of april.
I know that it's a bit close to the deadline but will anyone be the hunter for my project? Project link is: https://rainz.net and the link to the launch is: https://www.producthunt.com/prod...
Retrace records every LLM and tool call in an agent run, lets you replay it step by step like a video, and lets you fork any step to change the input and see how the run plays out differently. I built it because debugging an agent that failed in prod usually means re-running the whole thing and guessing, and I wanted to reproduce the exact run and change one step instead. Docs Playground
Just wanted to start a thread to learn from and connect with more founders and builders in the Product Hunt community. I always enjoy discovering what people are building, so I thought I'd start.
We're building Myspec. The idea came from a problem we kept running into ourselves: AI can generate code incredibly fast, but it can't fix unclear thinking. We found ourselves constantly rewriting prompts, discovering missing requirements halfway through development, and realising everyone had a slightly different understanding of the product. So instead of asking AI to write better code, we built MySpec to help founders and developers turn rough ideas into structured, build-ready specifications before development even begins.
We launched MySpec on Product Hunt about two months ago, and we've been continuously improving it based on feedback from our early users. If you're curious, we'd love for you to check it out and share your honest thoughts: https://www.producthunt.com/prod...