alt Hacker NewsAsk HN: What are you working on? (November 2026)
What are you working on? What have you been curious about lately?
Comments
Fast and dense stock market related visualizations, example:
https://chartrow.com/quote/msft/trends
And showing groups of stocks on a single page like this:
https://chartrow.com/top-performers#view=charts&p=ytd
https://chartrow.com/top-performers#view=heatmap&p=ytd
https://chartrow.com/top-performers#view=calendar&p=ytd
It's interesting to see some AI stocks reverse their trends in October 2026:
https://chartrow.com/top-performers#view=calendar&p=ytd&in=n...
I work on a suite of open source public transit software projects, collectively called OneBusAway (or OBA). The software is used by millions of transit riders every day around the world, including in New York City, the Seattle area, San Diego, and Washington, D.C.
University of California San Diego has deployed our Wayfinder and Waystation web apps for their 35,000+ students!
We're getting ready to launch our Vehicle Positions software, which replaces expensive, proprietary vehicle location systems with free, open source software that runs on commodity smartphones.
Our next-generation server software, Maglev, is adding support for newer transit data formats, like GTFS-Flex so that we can show demand responsive transit systems on the map.
We are always looking for more developers, biz dev, ux designers, product managers, and more to help out!
Open volunteer positions: https://ossvolunteers.com/organizations/open-transit-softwar...
Our software: https://github.com/OneBusAway/
More about us: https://opentransitsoftwarefoundation.org
Writing a Newspeak implementation of the PostgreSQL TCP protocol!. Talking to a GnuCash database from Chrome web browser. Uses my Newspeak TCP library. Both are bare minimum, probably very broken, but I am able to parse a Postgres result set from a simple query.
Here is a one minute video where I print the result data to the console from a Newspeak workspace.
A singing training app for bad singers. You sing the pitch line and get a score. Let's you sing to any original song by listening along and not playing the song itself.
Recently got some good feedback on HN https://news.ycombinator.com/item?id=50195488 and am following up with Linux and Windows versions and using Fable 5.1 to internationalise the app and landing page which explains the app.
I'm still working on https://openaltfinder.com, a directory of open source alternatives for proprietary software.
Trying to find out why Google has completely de-indexed the website, but other than "low quality content", I'm not getting any answers. Even Bing brings in more traffic...
I have been working on building an agent and workflow builder that converts chats with claude, chatgpt, gemini or any other chat agent into a workflow or autonomous agent. Its live at toolscaled.com
Also built a free apache 2 agent/ project management system: https://github.com/mhalstrom/maximizepm
Essentially I found the key thing with modern ai assisted development the key thing is to manage your own time/ parallelize without having to switch between chat session as much as possible.
I’m reverse engineering and replacing the entire stack on Samsung A54 (Exynos 1380, Mali-G68), started with booting to a customized linux kernel (7.3) w/o Android layer at all (Wayland and WPE WebKit only), built a custom WLBT thin firmware to replace the original fat blob (using mac80211 as the foundation), figured out how to exploit USB control transfer and got ACE in the PBL, reconstructed (with byte-identical verification) their LK-based S-Boot, uH and harx, patched mali/panfrost, optimized kernels for this device (finally got STT/TTS running on GPU faster than on CPU), figured out how to utilize NPU, built a custom toolchain for it. I’m trying to properly own this device (and giving it to Claude on the lowest possible level, but still running pKVM and getting rid of original hypervisors), it’s the goal number one at this point, no binary blobs, ability to change sw and control hw on-the-fly. Obviously with AI usage (Opus 5.5/4.8, GLM 5.3, Kimi 3.1), there’s a lot of work.
A series of actually hard puzzle books, so far sudoku, killer sudoku and some fiendish wordsearches that only use 5 letters per grid.
https://extremepuzzlebooks.com
A sample wordsearch is available as a pdf here: https://extremepuzzlebooks.com/assets/sample/extreme-wordsea...
I'm quite proud of this wordsearch concept as I've not seen it done before.
All the logic puzzles have an online solution to show step by step how you can complete them.
https://extremepuzzlebooks.com/walkthroughs
I've been working on the software to generate, test and verify these for a couple of years and finally built up enough puzzle stock that meets the difficulty requirements to properly launch.
Now the work is on every hackers favourite part... Marketing!
I burn most of my leftover tokens on Ultima Online stuff nowadays.
Started with a funny experiment, to have a CLI that can spin up a server (ServUO, Sphere, whatever), then deploy it somewhere (local, VPS, or Railway). Then saw interesting stuff that people did on top of existing ClassicUO, and decided to extend CLI's capabilities to client area too. So, I was able to deploy a server and a web-client (that you can open a webpage and play the game without the actual client installed on your machine). This was a significant step because running UO on macOS is always a struggle. Then I decided to add player-like bots so that the server doesn't feel so lonely. Started with some local decision trees, when it failed I hooked it up with Jev and Clef. Also added Haiku for speech generation.
While doing this, I realized I need to find some game content (gumps, tile or item data, animations, etc). Similar to UO client, running UOFiddler on macOS is also another struggle. So I decided to port it to web.
If you're into Ultima Online, these two are open source, you might want to take a look at them:
For the last 3 years, with a small team has been building https://kube-dc.com/ - a platform for building enterprise grade private and public clouds. Its like VMware + OpenShift + Railway in a single product: completely multi-tenant, built entirely on the K8s API (we built over 46 custom controllers in Go, with UI in Radix UI), with managed services like nested K8s, Postgres, Kafka, Valkey, and GitHub runners. Not the charts :) real services. Also we can do vlans BGP, shared GPU and all fancy stuff required.
Now we don't deploy this to private companies except some gov projects. But we are engaging existing IaaS and dedicated servers providers. Later hope we would fight with Broadcom and IBM :)
We previously ran a services company that built over 100 private clouds, then decided to turn that experience into a product with an aggressive level of automation to let DevOps, SRE, and platform teams hit a 1:300 ratio per developer (now dev + AI agent teams).
We are currently deployed across 12 datacenters and raising our seed round this month.
Regarding been curious, today I have created my first explainable video with Remotion +claude so if you want to see open edit results check https://promo-review-shalb-promo.stage.kube-dc.com/
Would be glad if you comment. Thanks!
My 10+ year old SaaS for test prep: https://crucialexams.com/
I recently rebuilt my COVID project, a Steam/online version of the drinking game Kings Cup: https://store.steampowered.com/app/1506610/Kings_Cup_The_onl...
A friend and I have also started working on a restarting/brewery manager game that is not released yet.
Recently worked on https://fastsleep.app
Use this to quickly get sleepy without taking any medicinal pill or long form meditation type things (It is as short as 7-minute session).
Easy to use. Simply close your eyes and imagine the micro-scenarios you hear. Like if you hear "moonlight on a white flower", imagine that scenario until you hear the next one.
You may also use it to break nightly doomscrolling habit. Try this tonight! (and any suggestion to improve it?)
Independent Vet - a website and app that helps pet parents find locally owned (and corporate/PE owned) veterinary practices nearby. It came out of a bad experience we had with our dog 3 years ago at a practice that had been recently acquired by private equity. The sites now get about 30k visitors a day and continuing to build in the evenings and weekends with Claude/Codex has been a lot of fun.
A month and half ago launched Maxxmod, a Chrome extension that improves the YouTube experience by removing distractions (hiding Shorts and cleaning up search results) and adding new tools and shortcuts (volume booster, screenshot tool, quick access to Picture-in-Picture and playback speed controls). I’ve been adding features since then.
Last week, YouTube introduced GIFs in comments, so added an option to stop them from autoplaying and play them on demand (a play button).
Disclaimer: I posted a Show HN when launched the 1.0, but it didn’t get traction: https://news.ycombinator.com/item?id=49829812
I've been deep in the weeds on two wildly unrelated projects.
Project #1: I've been working on migrating an application [1] from GCP (where I pay $200/mo to host it) to a physical server (PowerEdge R7625) I've racked in a colo near me, running on my own IPs/ASN [2] and all that jazz. I've been dragging my feet on the this as there's just a lot of moving pieces, but I got a good chunk of the last few hard bits done today.
Project #2: Building a DIY bright-ass lamp, like Brighter [3] or Rhythm [4], but for a few hundred bucks instead of $1-2k. I've specced out the parts, designed the hardware, and built the firmware. Later today I'll do my first LED test against a bench supply, to get some hard numbers for if my cooling solution (a chunky aluminum heatsink and some Noctua PC fans) is sufficient for the full 500+ W final version.
As far as curiosities go, I've been doing a lot of Go (the board game, not the language) puzzles, and reading about sailing (mostly the book from the Hundredrabbits folk [5])
[2] https://bgp.tools/as/55078
I live in San Francisco and we get some beautiful sunsets here! Over the past few weeks, I cobbled together a system of alerts for when the sunset is predicted [1] to be particularly outstanding, resulting in this page that displays the prediction and also shows the current view from several cameras around the city.
[1]: Predictions are LLM-powered largely to more reliably consume the free-text descriptions in traditional (e.g., NWS) weather forecasts of predicted high and low cloud cover around sunset.
Over the last few months, I decided to learn more math and dive deeper into LLMs, RL, and robotics.
I ended up building a personal agent that asks me questions every day, emails me interesting reads in these areas, and helps me progress in my learning.
It started as a web app, but I have just ported it to a Mac desktop app.
It works like Conductor, but for your learning side quests: https://miletus.app
Drive.py Manage files across your connected and disconnected drives
A complete implementation of React in Godot:
Because I can. Well, let's be honest: because LLMs can. There's nothing particularly novel about this. It's just wiring a bunch of libraries together. Certainly would have been time consuming, but not difficult, strictly speaking, since all the heavy lifting is done by the libraries utilised.
Likely going to migrate https://breaka.club HUDs and editor UI to React since Godot control nodes are by no means impossible to work with, but somewhat cumbersome.
EDIT: Oh, will be open source like a lot of Breaka Club's other libraries very soon. Still optimising.
Recovering from pushing out an article on coroutines [1]; coordinating getting the house in order for the workweek [2]; and drafting a package to make literate programming in org-mode a bit more to my liking.
---
[1] https://www.reindeereffect.com/0004
[2] Parts of the house smell of slowly cooking meal components (e.g., stock and beans), while others smell of cleaning products.
https://github.com/mquan/netflix_challenge
I'm revisiting the Netflix Prize challenge with LLM. So far I've hit the 9% improvement mark within a couple weeks. I seem to have hit a wall in the last week or so with no further progress. I'm hopeful to hit the 10% goal with one of the many experimentations though.
My partner and I are both software devs. She quit her job a year ago, and has been working on a few micro-businesses to try a) learn some new skills and b) diversify our sources of income. So far this year she has built:
* https://unofax.com - A simple a la carte fax service
* https://unopost.com.au - A way to send physical mail online (and avoid standing in line at the post office)
* https://snailgram.com.au - Custom greeting cards, with the message inside written by a 3D printer holding a pen.
There are a handful of others too, but these are the ones that are live so far. I've contributed between 5-40% to these projects.
I built an app to help me develop healthy habits and routines on my computer. IT blocks sites/apps based on a schedule and tracks my activity for later analysis.
I had simpler versions of this earlier but never something comprehensive. It was also my first AI assisted development project. I've been using it personally for around a year. I polished it and released it as an open source project last month. It's specific to Linux (which is my primary platform)
The website is at https://glockerapp.com/
I'd posted this earlier on HN hoping to get some feedback but didn't get any traction https://news.ycombinator.com/item?id=50349646
I'm working on my first hardware device, based on ESP32, it's an AI employee in a cube.
A few Apple container plugins.
https://robertderose.github.io/container-krun-runtime/
It’s an alternative VM engine that uses libkrun. The main purpose is that libkrun supports returning memory back to macOS when the VM releases it. It also has better performance in some areas, but it’s only been tested on my M1 laptop and I haven’t added nested support yet because I only have an M1
And https://robertderose.github.io/container-compose-plugin/
It’s a full docker system running inside Apple Container’s machines with automatic localhost port forwarding, automatic shutdown if nothing is running, and exposes a tcp socket so you can use most tools by just exporting your DOCKER_HOST environment variable.
I have a few other projects but they are still in earlier stages
Given the popularity of decision models I tried to create a standard text classification pipeline that runs entirely in the browser using webgpu/wasm. Downloads an embedding model, trains the classifier (a Linear SVM for now, whose hyperparams are determined by gridsearch), calibrates the classifier so that reliable confidence scores may be produced. This is in Python that is run using Pyodide. You can paste your own data (no uploads supported) to train a model on it. WIP, and mostly generated using GPT Sol. Testing a few things out. Will have a blog post soon.
Cloud save sync between a real Super Nintendo and emulators.
I’ve just got back into hobby electronics after a fifteen year diversion into software. I set up my workbench this week and this is my first project, something I first imagined building twenty years ago!
I spent the weekend wiring the spare controller port to a microcontroller to act as WiFi bridge. I have a flash cart, so I patch the reset vector to a small save sync routine that talks to the microcontroller.
It was especially magical having Claude read from a logic analyser and flash firmware tweaks iteratively. I still got to wield the soldering iron at least.
A suite of high-dopamine games which use smart exercise bikes as the control interface.
When playing Zwift, Rouvy, etc, I find that I get too competitive and ruin longer workouts by trying to keep up with or sprint past people who are much faster than me over long distances.
This is bad for building stamina and endurance, but I get bored, and it's hard to fight boredom and physical strain both at once.
I do not get bored sitting in front of Balatro, Raccoin, Tetris - anything with bright colours and satisfying noises and combos.
The first finished prototype game is very simple: you sort coloured marbles. Higher watt output from the bike means marbles are blown, by wind, upwards as they roll across the screen. Lower watt output means they fall. Steady watt output means they move horizontally. Marble goes into correct hole -> satisfying noise. It rewards steady watt output over time, with minor adjustments, which is what I'm trying to work on right now.
I was pretty amazed by how engaging I found it (total iPad baby brain). The miles really fly by - I get tired before I get bored.
I think that next I'll either do a tower defence game, or a less-dopamine-rewarding game in the style of Kerbal Space Program. Taking your actual weight and watt output, putting a bike with wings into a physics engine, and the gameplay loop being unlocking increasingly large airplane pieces, and designing - then flying - increasingly ridiculous airborne bicycles seems like a compelling idea, though less good for fitness.
The control surface being so limited - essentially one measurement which can either go up or down - is a really fun constraint for creativity.
For Halloween I made a VR survival horror game for meta quest 3/3s.
- https://sxp.studio/aftercorp
The cool thing about that game is that it uses a realtime environment mesh reconstruction algorithm that allows me to create an approximation of the world 3D geometry and project a dummy “flashlight” with hand tracking. I think I’ve only seen one other game do that technique on quest (some laser tag game).
It runs at 60fps with a tight GPU budget. Consequently it’s not the most beautiful game, but it’s actually pretty fun, immersive and can scale to all sorts of environments. I’ll be porting it to iOS next (handheld AR style).
ArtCraft and the Craft Apps. https://getartcraft.com
We're going to open source everything. Software, SaaS, infrastructure, weights, platforms, protocols.
We're going to bridge local and cloud compute in a seamless way.
We're building an app factory for high throughout development. Reverse engineering tools for non-technical people. We're going to make high quality libraries that others can remix.
We're going to make all of this agentic and AI/CLI first class.
We're going to open source everything and make the world hacker-first again, like Web 1.0 with P2P and RSS and APIs and IRC.
Join us!
I'm building an open-source "spatial workspace for directing AI coding agents like an engineering director".
I got overwhelmed by walls of text of parallel Claude Code sessions. This limited the sessions I could steer in parallel to maybe 5 or 6. I knew there had to be a better way, and I found one: A canvas per project. Work laid out as cards. An agent in a worktree per card. Chat condensed to WhatsApp-style back and forth, not a litany of tool calls. Result is usually a narrated demo video that I can comment on by voice to have the agent go another round. Remote-control from my phone. Remote workers to scale beyond my machine. Everyone I know who tried it has stuck with it so far.
As a heavy LLM user in my daily work, I've been working on optimizing agent memory on complex projects. That resulted in this simple Rust-based tool called Wayfinder: https://github.com/zaterka/wayfinder
I've been trying to improve agent knowledge over a codebase, and existing approaches weren't giving the result I wanted, and I wanted to keep it simple. In the end, I added a semantic layer on top of an LLM Wiki and it saved 30% of token usage in multiple benchmarks I ran. It also adds some nice features to deal with concurring agents updating the shared memory, to ensure all information there is reliable and fresh.
It's published to PyPi, so an "uv tool install wayfinder-memory" already does all the work, and it's compatible with almost all Harnesses.
I’m working on https://greenflagged.dev
I originally thought it would be useful to have a package registry that helps prevent supply chain attacks. It reads and judges all of the changes to popular NPM packages. While building I realized that it’s a package manager that agents can settle package versions with as old versions have new vulnerabilities found.
I’ve caught several supply chain attacks since I started monitoring, and am building an automatic response pipeline to validate and publish new attacks.
I’ve also been working on python support and hope to have that released soon.
I recently made a mobile app for practicing singing: https://singtrace.karamagic.com
The free version you can record and keep an archive of your daily practices of each song, while a paid version does some analysis of your vocals.
I’ve been working since the start of the year on a minimalist web reader (https://papercove.app) that I use every day to follow RSS feeds, newsletters and read-later, and that helps keep things from piling up.
I thought I’d get where I wanted much faster, but it turns out I ran into quite a few challenges: it’s quick to build something simple, but it takes much longer to get a reader that’s robust and pleasant to use (extraction, parsing, security, infrastructure, etc.).
On the tech side, I’m using my usual stack, Laravel and Livewire, with some Python for extraction and Go for the email infrastructure. This month I’m working on the beta release, since several people around me were already using it, so I’d like to gather more feedback.
For 4+ years I've been working on a game programming language which makes your game multiplayer automatically! Every program written in Easel is guaranteed deterministic and snapshottable, which it uses to automatically perform rollback netcode. The rollback is incremental, it automatically stores deltas of only what has changed using a copy-on-write mechanism, so it's really efficient for games where your players are walking around large worlds which are mostly sleeping except when the player is nearby.
The idea is that I hope teenagers who love to play multiplayer games, would find this to be the easiest way to code a multiplayer game by far since the multiplayer is automatic. Learning to code could be much more fun because they can make games that involve their friends, classmates, brothers/sisters, parents, or just anyone online. Maybe especially more important for raw coding to be fun and engaging in the age of AI.
This month I'm making the multiplayer even more efficient so people really can just code large multiplayer games without worrying about it, which I think is important for beginner coders.
A proportional width font terminal. "i" and "w" shouldn't take up the same amount of horizontal space. We used to have to do that because computers were slow and we counted cycles and every bits mattered. We live in a land of plenty now though so we can spend cycles doing math so that ls -l/top/etc table output is still aligned but "i" and "w" can take up the appropriate amount of horizontal space.
Teaching myself how to build a Spiking Graph Neural Network in Zig, to run on my Brainchip Akida AKD1500 neuromorphic accelerators (and Efinix Titanium Ti60 FPGAs).
Good times, but lots of foundational math I hadn't previously been exposed to.
Some reading material if anyone else is interested in that:
Spiking Graph Convolutional Networks (2022): https://doi.org/10.24963/ijcai.2022/338
A Tutorial on Spectral Clustering (2007): https://doi.org/10.1007/s11222-007-9033-z
Semi-Supervised Clustering with Graph Convolutional Networks (2017): https://arxiv.org/abs/1609.02907
Spectral and Algebraic Graph Theory (2025, draft book): http://cs-www.cs.yale.edu/homes/spielman/sagt/sagt.pdf
https://opentaint.org/blog/sast-is-dead/
A data-flow-based, lightweight formal verification engine for application security. Basically, it's SAST++:
1. Deep enough that you really have some formal guarantees about the codebase (like zero findings actually means something, not just that the tool is broken). 2. Fast enough that you can shift the formal verification phase left and apply it to every feature during development (for example, during agentic development sessions to get fixes done within the same feature context). 3. Expressive and customisable enough that you can write specs for most vulnerability types from the OWASP Top 10 (even Broken Access Control and business logic flaws). Skills for writing specs included.
Here is the repo: https://github.com/seqra/opentaint/
I'm making my Ancient Earth globe better. It uses plate tectonic data to show you where your city was on the globe hundreds of millions of years ago.
Recently added new high-res maps, a timeline, and country names and borders.
For example, see Pangaea:
https://dinosaurpictures.org/ancient-earth?borders=on&names=...
I made a browser-based recreation of the popular 1999 MMORPG "EverQuest" - and it plays itself! https://idlequest.net/
Have 70+ people online today playing it, it's been fun to see it get so popular recently (for a free niche audience hobby project).
Also have some other cool browser-based 3d game projects listed here: https://brynnbateman.com/
https://github.com/Vermeulen-Solutions/baseline
Astro site starter template meant to be used together with agents.
Building my an airflow pipeline focused on collecting news.
Partly because I want one, but also because it makes for a good learning platform. e.g. initially do some of the processing with a stronger llm...save result and then attempt to train a SLM or similar
- https://probl.dev: A "probabilistic" programming language where we can do things like `if 45% { x } else { y }`. It works by creating many weighted worlds, that are then merged when identical. Operators like `observe` can remove possibilities. There is also a sampling mode that instead run many simulations to get an approximate result.
- https://github.com/garlab/gloss | `brew install garlab/tap/gloss`: render markdown, sources, images, chess and go (game) records in the terminal. It's built for libghostty but fallback on other terms.
Building a native macos app with a suite of completely local AI tools. The first version is out and have a long roadmap in front of me. Would love to hear some feedback.
https://www.intelligrade.de/en
A SaaS for teachers and schools that helps them create, conduct and grade exams and tests (with safe exam browser integration to prevent cheating). Grading happens automatically for almost all segment types, for free form text we have an optional AI integration. I recently added an MCP server so teachers can bring their own LLM with fairly fine grained permission controls for it. It’s hosted in the EU with options for schools to self host with a license.
Currently we are in a pilot phase and are yet to make our first sale but we are reaching out to schools and gave out some pro accounts for free for people to try out and feedback has been great.
We are just two people, a teacher who was dissatisfied with the tooling in the field and me, who engineered a solution that we believe is genuinely better than what’s out there.
I’m still working on https://feedsync.net, a tool to push RSS feed updates to Discord.
It’s been getting a decent number of sign ups despite me not really posting it anywhere or doing any marketing.
Ultimately a fun project, and useful for myself and a few friends, so I’m happy others are even interested in trying it.
I’ve started to regret using Phoenix LiveView, as it’s only deployed in one region, so high latency far from there. I’m exploring a SPA as a better alternative.
Otherwise, I’ve been playing around with a Lua engine to build feeds for websites that don’t support them, using AI to build the Lua as a once-off. It works reasonably well, but isn’t ready for prime time just yet.
I’ve been working on Beacon, which just got approved on the App Store! It’s a mobile app to instantly answer “which friend is free to talk right now?” and summon them in a single tap.
Send your beacon to a group, and everyone gets a gentle ping. The first person who answers gets connected for a 1-1 call, and for everyone else the signal drops silently. It does not log any missed calls so there’s neither pressure to answer nor pressure to return calls.
Overall it works really well! It’s best if you're able to join with 4+ friends/family you’d like to chat with more. I have dozens of connections in the app now and it feels like absolute magic to me. I tap the button to send my beacon to everyone and get a random friend on the line.
I was laid off a few weeks ago. Then I learned my dad’s carpet cleaning company has been struggling, so I started learning about Google Local Service Ads to help him out. THEN I learned his old Windows desktop CRM was discontinued and one OS update away from no longer working. I wanted a way to track every dollar spent to determine return on ad spend, but the Google LSA dashboard left a lot to be desired. And existing SaaS CRMs didn’t have great integrations to Google Ads API. So I built Crewbird, a Field Service Management and CRM for service based businesses with Google LSA integration where my dad is quickly approaching 3x ROAS!
I’m still looking for work in the mean time but I’m proud of what I’ve achieved so far.
https://getcrewbird.com/