Live data from Hacker News

Ask HN: What Are You Working On? (March 2026)

news.ycombinator.com

731–740 of 1001 posts

Re: Ask HN: What Are You Working On? (March 2026)

#734
post #716

This month I'm recovering from major surgery. I'm okay, it's something I've been waiting to get done for a very long time. I'm just bored in recovery and running out of my science YouTube backlog.

Some fun channels if you are bored, and I hope you feel better soon =3

https://www.youtube.com/@project-326/videos

https://www.youtube.com/@Asianometry/videos

https://www.youtube.com/@PhysicsExplainedVideos/videos

https://www.youtube.com/@TwoMinutePapers/videos

https://www.youtube.com/@DrBenMiles/videos

https://www.youtube.com/@professorjenniferhaslersci8510/vide...

https://www.youtube.com/@RobertFeranec/videos

https://www.youtube.com/@AdvancedTinkering/videos

Re: Ask HN: What Are You Working On? (March 2026)

#735
I am building npm for LLM Models, you can now install run and ship AI Models.

It’s a CLI tool that makes open-source LLMs installable like packages.

llmpm install llama3

llmpm run llama3

You can also package models with your projects so others can reproduce the same setup easily.

Website: https://llmpm.co

GitHub: https://github.com/llmpm/llmpm-dev

Re: Ask HN: What Are You Working On? (March 2026)

#736
I'm building a proxy that sits between Claude Code and the model to visualize the full interaction loop. It shows prompts, tool calls, and responses in real time so you can explore how the coding agent is reasoning and operating. Project: https://github.com/tndata/CodingAgentExplorer

Re: Ask HN: What Are You Working On? (March 2026)

#737
I'm writing a print farm manager for my wife, who has 8 bambu printers. I've put one of them in lan mode for testing, but I'm pretty close to replicating the important handy (Bambu's cloud app) features, at which point I'll be able to start adding the really nice.

She sells a product with 16 different printed parts, and she prints the parts in bulk batches across 7 different pause points, some of which have pause points for embedding magnets.

The idea is to integrate inventory management and print scheduling into the tool, which will be nice.

I have working so far: * Pulling camera images * Pulling the currently printing file, including the preview image (rendered in bambu studio and bundled with the print; standard for bambu studio), and the pause points * A dashboard with projected timing information * Notifications about jobs starting, stopping & pausing * Remote printer control

Next on the list: * Delayed printing - schedule a print to start in the night. Mostly useful so that if there's a pause point we don't leave a print paused for hours on end. * Print queueing - manually build a list of prints so that after switching plates we can just "next print" for a printer * Print scheduling - select a quantity of print files or groups of files to print, and have it schedule the prints, including projected switch times, to maximize printer utilization by avoiding jobs ending at night * Tracking magnet & filament usage, and integrating BoM and production quantity tracking.

I've been mostly AI coding this, but I've go in to make it extract out components, etc. And I lay down and enforce the DB schema. I've had to ask it to back out a few things entirely. And I've had to give it the Bambu API docs I found github. But it's been going pretty well.

Re: Ask HN: What Are You Working On? (March 2026)

#738
Nothing revolutionary, but there is a small organisation called The Himalayan Database [1] who have recorded (in great detail) climbing expeditions to peaks in the area. The data is available in a downloadable format, but it is a little awkward to browse. I have been working in collaboration with them to build a website for making the database more accessible https://himalayan-database.climbing-history.org/

[1] https://www.himalayandatabase.com/

Re: Ask HN: What Are You Working On? (March 2026)

#739
I'm building a personal habit tracker, solely for my own use. It is intended to be pretty basic - just a html file, with data saved by vanilla JS in browser's local storage. Currently about 50% of the work is done by AI (Cursor).

After adding a couple of extra features and having a "finished" tracker, I will try re-implementing this tracker in React, Svelte, Vue, Preact and some others.

My goal for this project is twofold: to get familiar with these frameworks and to practice using AI as a personal tutor (leading my way and answering my questions).

I've tried learning React, Laravel, etc before, but I've used them to build a fresh project from scratch and I've always got stuck early on due to the lack of knowledge/understanding.

I hope that re-implementing something that I already know and understand fairly well would make my learning process much more effective.

Post reply on HN