Making a photo-based calorie tracker accurate.
Ask HN: What are you working on? (October 2025)
281–290 of 1001 posts
Re: Ask HN: What are you working on? (October 2025)
#282[1] https://www.robinsloan.com/notes/home-cooked-app/ [2] https://booplet.com/blog/anyone-can-cook
Re: Ask HN: What are you working on? (October 2025)
#283I found a neat way to do high-quality "semantic soft joins" using embedding vectors[1] and the Hungarian algorithm[2] and I'm turning it into an open source Python package: https://github.com/olooney/jellyjoin It hits a sweet spot by being easier to use than record linkage[3][4] while still giving really good matches, so I think there's something there that might gain traction. [1]: https://platform.openai.com/docs/g…
Cool project! I see you saved a spot to show how to use it with an alternative embedding model. It would be nice to be able to use the library without an OpenAI api key. Might even make sense to vendor a basic open source model in your package so it can work out of the box without remote dependencies.
Re: Ask HN: What are you working on? (October 2025)
#284The current challenge is the display. I’ve struggled to learn about this part more than any other. After studying DVI and LVDS, and after trying to figure out what MIPI/DSI is all about, I think parallel RGB is the path forward, so I’ve just designed a test PCB for that, and ordered it from JLCPCB’s PCBA service.
Re: Ask HN: What are you working on? (October 2025)
#285Re: Ask HN: What are you working on? (October 2025)
#286It's largely finished and functional, and I'm now focused on polish and adding additional builtin functions to expand its capabilities. I've been integrating different geometry libraries and kernels as well as writing some of my own.
I've been stress-testing it by building out different scenes from movies or little pieces of buildings on Google Maps street view - finding the sharp edges and missing pieces in the tool.
My hope is for Geotoy to be a relatively easy-to-learn tool and I've invested significantly in good docs, tutorials, and other resources. Now my goal is to ensure it's something worth using for other people.
Re: Ask HN: What are you working on? (October 2025)
#287I wanted to build my own speech-to-text transcription program [1] for Discord, similar to how zoom or google hangouts works. I built it so that I can record my group's DND sessions and build applications / tools for VTTs (Virtual TableTop gaming). It can process a set of 3-hour audio files in ~20 mins. I recorded a demo video of how it works here: https://www.youtube.com/watch?v=v0KZGyJARts&t=300s [1] https://github.…
I have found a hack. If you wait long enough, someone will build what you wanted to build :) Thanks for building this. I am trying to set it up but facing this issu > `torch` (v2.3.1) only has wheels for the following platforms: `manylinux1_x86_64`, `manylinux2014_aarch64`, `macosx_11_0_arm64`, `win_amd64`
Re: Ask HN: What are you working on? (October 2025)
#288Microlandia, the brutally honest city builder. Posting this for a second time, because i’ve been working super hard on a steam release. last month’s “what are you working on” thread impulsed me to upload this game to itch and 1 month later, i’ve got a small community, lots of feedback and iterations. It brought a whole new life to a project that was on the verge of abandoning. So, I’m really grateful for this thread.…
Brutally honest? I hope it shows the huge amount of land needed for parking lots :P I heard that the SimCity devs have had to fudge that out for gameplay's sake ever since the oldest versions
Re: Ask HN: What are you working on? (October 2025)
#289The Pain Point: If you are analyzing a large YouTube channel (e.g., for language study, competitive analysis, or data modeling), you often need the subtitle files for 50, 100, or more videos. The current process is agonizing: copy-paste URL, click, download, repeat dozens of times. It's a massive time sink.
My Solution: YTVidHub is designed around bulk processing. The core feature is a clean interface where you can paste dozens of YouTube URLs at once, and the system intelligently extracts all available subtitles (including auto-generated ones) and packages them into a single, organized ZIP file for one-click download.
Target Users: Academic researchers needing data sets, content creators doing competitive keyword analysis, and language learners building large vocabulary corpora.
The architecture challenge right now is optimizing the backend queuing system for high-volume, concurrent requests to ensure we can handle large batches quickly and reliably without hitting rate limits.
It's still pre-launch, but I'd love any feedback on this specific problem space. Is this a pain point you've encountered? What's your current workaround?