alt Hacker NewsAsk HN: What are you working on? (January 2026)
What are you working on? Any new ideas that you're thinking about?
Comments
Porting WebKit to Fil-C
Current status: JavaScriptCore builds in the JSCOnly config and panics on start.
JSC has extensive use of uintptr_t as a carrier of pointers and that's what most of the panics are about
I have been porting over my favorite Erlang library "Bitcask" to Rust.
https://git.sr.ht/~tombert/feocask
Called "feocask" cuz feo means "ugly" in Spanish and FeO to mean Iron Oxide for Rust. I thought it was funny.
I will admit that I had help from Codex, but I did write most of it myself, and I think the design is coming out kind of neat. I have a very strict "no lock" policy [1], including lockfiles, and this should still be safe to use across any number of threads, at the cost of N^2 reconciliation to the number of threads and a lot more drive space.
I like my design; I have an excuse to use Vector Clocks and Hybrid Logical Clocks and I think it might actually be useful for something some day. I'd like to eventually write something that goes a bit beyond getting parity with bitcask and optionally have the ability to automatically distribute across multiple nodes, but I'm still trying to think of a good design for that, because my current design depends heavily the atomicity of POSIX filesystem commands, and introducing the network introduces latency that would likely greatly degrade performance.
[1] At least no explicit locks. I am using Tokio channels and they are probably using locks in some spots behind the scenes.
I am working on a local RAG LLM designed for lower end PC's - ability for people to try out searching their own documents, seeing it was such a learning curve to get to this stage - hoping others can learn from my mistakes.
https://github.com/acutesoftware/lifepim-ai-core
Only been public a few days, so please let me know if there are glaring issues.
I'm working on https://wireplug.org: A simple, free, and open source connectivity coordinator for WireGuard. Basically a way to keep WireGuard tunnels connected while moving between different access points. It handles (basic) NAT traversal and works with the in-kernel WireGuard driver on Linux and OpenBSD. You can find the technical details at https://wireplug.org
During the holidays, I’ve experimented with some ideas. I wondered if it was possible to make money in 2025 using simple, nano banana wrappers, thr answer is yes!
for example, RecolorLife.com and Headshoti.com generate around $800 USD.
Now I will expand for real estate.
I think frontier models are getting to the point where we can start to reach higher trust agentic workflows.
As a hardcore AI chat user, I'm often frustrated with the single-agent workflow, where a single context window is used for even very long conversations. If I want to change the topic, open a thread, or go on a tangent, I often end up compromising the main thread and I'm forced to copy context over if I want to dive into something.
To solve this, I'm working on a collaborative AI agent orchestrator that models the solution as a group chat with humans and AI agents, including an agent orchestrator.
You can spawn participating agents with the orchestrator who will decisively route messages to the existing agents, or spawn new agents if needed. Also, you can open agent details and send messages directly to existing agents, similar to threads in slack.
So far, I have MCP integrations working with Linear and GitHub, but plan to add many more.
I've been working on this just over 2 weeks, making heavy use of 4+ concurrent Claude Code agents. This would have been impossible otherwise.
If you're interested, feel free to DM on X.
Nowhere near as impressive as some of the things here but I’ve been slowly growing out a daily puzzle website https://regularly.co - I built the first few games for my wife, no subscriptions, no pay to play, no sign up needed.
Im launching a new word game next week which I’m super excited about. If you do play it and have any feedback do shoot me message!
To relax and reset at the end of the day, my wife plays this simple tetrisy game - fill rows and cols to make points and try not to die. But the game and many of its ilk are super frustrating. They have ads and arbitrarily make it impossible at some point to keep going (in order to show more ads I assume.) So I combined my love of roguelike deck-builders and made a game for her. It's the same basic premise of filling rows and columns but it has a deck you improve over several boards of ever-increasing complexity. And it has cards with special powers (of course.) It turned out to be really fun both to build and play. I'm still working out a few kinks (I built it in Love2D but running inside iOS is a challenge...) But it's live in the App Store. Check it out if you're interested and hmu if you'd like a free offer code - the game is fully playable without paying, but keeps some options behind an in app purchase. https://tetranea.net/
EDIT: It's only on iPhone and iPad for now. Android version coming soon.
Over the last two years I spent a surprising amount of time analyzing XSD files to understand regulatory changes in financial reporting. It was repetitive, painful, and unnecessarily manual.
In the past two months I built https://xsdviewer.com to make working with XML Schemas simpler: visual structure, navigation, diffing, and faster understanding of what actually changed between versions.
Right now I am iterating on new features and performance improvements. If you regularly work with XSDs or XML based standards and have ideas, pain points, or feature requests, I would love to hear them.
Random assortment of projects as time allows with the $JOB.
- Prototyping a cute little SSH-based sorta-BBS, inspired by the Spring '83 protocol, but terminal-centric rather than web-based. It's called Winter '78, and if we get another Great Blizzard this year, I'll be able to make some progress on it!
- Another prototype, for an experimental HPC-ish batch system. Using distributed Erlang for the control plane, and doing a lot of the heavy lifting with systemd transient units. Very much inspired by HTCondor as well as Joyent's (RIP to a real one) Manta.
A hobby project I started putting together late last year; a little spot on the internet for prayer and reflection.
https://dugnad.stavanger-digital.no/
A pro bono tech consultancy for local non profits. The idea is to help them use tech to better deliver on their mission.
Hello, I am new and this is my first post.
At Sylvester I stopped smoking.for the beginning and hard times I searched for a deflection and started with my own quitesmoke app - purio
I released it some days before a reached a workable state.
Here the result, if you want to check:
Android: https://play.google.com/store/apps/details?id=io.codingplant...
Feedback is welcome
Free to play FPV racing game with a global leaderboard that can run on integrated grapics in the web browser. Focus is on realistic physics (aero, PIDs and filtering, propwash, signal noise).
Continuing to work on Minifeed (https://minifeed.net/), a directory, reader, and search engine for personal blogs. The indexing & searching backend in Typesense, and I'm moving from their paid managed cloud service[0] to a self-hosted VM. It was very easy to start with the reasonably priced managed service, but with the number of feeds/posts growing, I have decided to self-host it. I'm also using Typesense Dashboard, a nice visual tool to do basic administration [1].
Overall, Minifeed keeps chugging along, fetching new posts every day from almost 2k feeds. I'm hoping to find some nice and ethical monetization strategy for it this year.
Acme [1], a text editor for programmers. It’s a very unusual tool and I have collected hundreds of notes (as well as dozens of bugs and inconsistencies) from the past two years of using it. Now I feel fully assimilated, I do not take as much notes and it feels the right time to start organizing the knowledge, to put ideas together.
I think that Acme is very underrated in its domain (a tool for experts). The coverage is minuscule. That’s why I am thinking about blog posts, maybe video tutorials. I do not know what should it be exactly.
I do not have any time estimates. I have a very demanding main job (I work as a software developer) and young family that I need to take care of. This project demands focused efforts and selectivity that I can barely satisfy. But I wouldn’t give up, the thing is totally worth it. If it’s going to take years, so be it.
If you have any comments, send me a letter at [email protected]. Have a good day.
https://fuzzygraph.com - aspires to be the most beautiful equation graphing app
I built an in-browser data explorer for datasets too large for spreadsheets but too small to justify spinning up SQL. It runs entirely client-side using DuckDB-WASM, no server uploads accounts or anything.
The core idea is a visual DAG where each transformation (filter, join, aggregate, pivot) creates a view node. Nothing materializes until you need it, DuckDB executes the full chain on demand so you can build deep pipelines without copying data at each step.
Input files can be CSV/Parquet/Excel (Excel might not work great). There's a SQL editor with schema-aware autocomplete, pivot tables with drill-down to underlying rows, and sessions can be exported as files or shareable URLs (the entire pipeline gets encoded in the hash).
Sharing can be granular and you can choose not to embed the files or if files are too big they will not be embedded and the user when opening the link will have to upload the files to restore the session.
The part I find most useful: you can replay pipelines on new data. Share your January analysis, and a colleague runs the same transformations on February's data with schema validation.
Privacy-first since files never leave your browser, it's a static website actually. I will open source soon, and make it probably MIT licensed.
Also it's a WIP and so it may be buggy (there's not even images on the homepage) https://repere.ai
I kept making a hacky eval tool to try and compare the outputs of different models, model configs, prompt versions, etc and finally rolled it up into a web app + downloadable app (kind of like Postman or Insomnia, but for AI).
Free, holds your keys in localstorage and makes direct calls to the APIs (unless there's a CORS issue), at https://evvl.ai if you want to try.
I'm working on a complex garden planner to keep track of gardens, orchards, and landscaping projects. Meant for very large and very small scales. Public alpha at the end of this month, I hope.
You can see what it looks like right now here: https://x.com/simonsarris/status/2010359423806615907
It is based on a garden designer I made for myself to keep track of my rose garden (I have over 100 roses) and orchard (I have about 15 trees): https://garden.simonsarris.com/
However, my version was very specific to my needs, so this general version requires a lot more work to get it usable for a lot of people.
Working on finally releasing the programming language I’ve been working on for quite a while.
I’m setting up the basic site, which is not a huge deal, but I’ve been inspired by more recent language designers having a streaming presence, so I am working through test runs of streaming my development.
I hope to start with demos of the basic language features and then move on to streaming both a reimplementation of my compiler and on a Rocq implementation of the syntax and semantics of the language for proof work.
The language has a rather small niche at first glance, so I’m hoping to use the streaming as a way to explore areas of appeal and maybe draw some interest. A low level concurrent and parallel ‘functional’ language with very non-traditional syntax and a modal, dependent type theory is not going to appeal to everyone, but hopefully I can find some interest eventually, even if just to hang out on chat and talk about the subject.
Building MatGoat (https://matgoat.com) - management software for Brazilian Jiu-Jitsu and other martial arts' academies.
I train BJJ and kept hearing the same complaints from academy owners regarding attendance tracking, comms, missing payments, etc.
So I'm building a tool for student tracking with belts progression, automated payments, attendance-based promotion criteria, and a tablet check-in system.
Focusing on Spanish-speaking markets first since it's completely underserved. Currently onboarding early academies, and will market it in the US/UK soon.
I’m building AgentOS [1], trying to experiment where agent substrates/sandboxes will head next. It's a deterministic, event-sourced runtime where an “agent world” is replayable from its log, heavy logic runs in sandboxed WASM modules, and every real-world side effect (HTTP, LLM calls, code compilations, etc.) is explicitly capability-gated and recorded as signed receipts. It ensures that upgrades and automations are auditable, reversible, and composable. The fun bit is a small typed control-plane intermediate representation (AIR) that lets the system treat its own schemas/modules/plans/policies as data and evolve via a governed loop (propose > shadow-run > approve > apply), kind of “Lisp machine vibes” but aimed at agents that need reliable self-modification rather than ambient scripts.
I've created a programming language and game for (my) kids called Stacky Bird. You learn Stack Based programming with Flappy Bird. Inspired by my old HP calculators, the game 2048 and Flappy Bird. In levels you earn a new instruction and at the end you have a complete instruction set, and can solve complicated puzzles. With new instructions you can even go back and solve previous levels more elegantly.
It's not mobile friendly yet, but maybe that'll be a next weekend project. At least you can view a video of a level on mobile.
My kids are loving it.
https://game.stackybird.com/ and the source https://github.com/jtwaleson/stacky-bird
Recently have been focusing on a personal assistant type of thing.
I've been building it with the agent sdk and any time I want an additional skill, I create it
Examples: parse this pdf containing my credit card bill and add all transactions
Given it has a db, I've been using it to save notes, ideas etc.
Been fun
Previously, I made a live win probability model for the 5v5 arcade game Killer Queen Arcade from their game events API.
Now I am trying to use that model to make:
1. A post game instant replay that shows the most important/pivotal moments from the most recently finished game. Some arcades have a seperate display for observers, it could work well there, or as good filler between matches on twitch streams.
2. A personalized per tournament/yearly highlights recap.
If it works well, it might be a kind of tool that generalizes well for summarizing long twitch streams for Youtube.
I'm working on BacklogAI, an autopilot engineer for SaaS teams.
You connect GitHub, CI, Sentry, and Linear, and it takes tickets all the way to production. Claude writes the changes, BacklogAI handles tests, migrations, feature flags, staged rollouts, etc...
It’s clearing months of backlog work in hours, and a couple teams I’m working with have already stopped hiring because it’s cheaper than adding more developers. It's crystal clear that developers and designers won't be needed in a couple of months because claude increases productivity by at least 120x and one or two PMs can do pretty much everything.
My stack is Claude, v0, nextjs, shadcn, clerk, supabase, vercel.
A specialized programming language for 3D geometry generation + manipulation called Geoscript as well as a Shadertoy-inspired web app for building stuff with it: https://3d.ameo.design/geotoy
There have been lots of cool technical challenges through the whole process of building this, and a very nice variety of different kinds of work.
I'm working towards using the outputs from this language to build out levels and assets for a browser-based game I've been dabbling with over the past few years.
Find local businesses with no websites (or check bad ones with built in SEO tools).
Build contracts, create billable invoices and track tasks for clients with a lightweight, web dev focused CRM.
Essentially, an all in one platform for web developers to find and manage their clients
I'm still working on my Web Server Library .NET Core
I'm rewriting from scratch : https://simplew.net/v26/
CentralCI (https://centralci.com) - Fully-managed Concourse CI as a service.
Got tired of helping enterprises run Concourse themselves, so we productized it. We've deployed and maintained Concourse for Starbucks, Accenture, Sky UK, and others over the years—CentralCI packages that operational knowledge into a SaaS.
Why Concourse over GitHub Actions?
* fly execute lets you test pipelines locally before pushing. No more "commit and pray" * fly intercept drops you into a running container to debug failures * Resource-based triggers can monitor anything—not just git pushes * Full pipeline visualization from dev to prod in one view * Workers you actually control (no arbitrary cache limits or runner queues)
What we handle:
* Dedicated Concourse instances on high-spec hardware (Ice Lake Xeon, DDR5, NVMe) * Worker scaling without the Kubernetes complexity * SOC compliance, auditing, AWS PrivateLink for enterprise * 24/7 support from people who've been running Concourse in production for years
The pitch is simple: Concourse is the right tool for complex CI/CD, but running it is a pain. We make it not a pain.
A transformer-based (but not LLM) chess model that plays like a human. The site right now is very rudimentary - no saving games, reviewing games, etc., just playing.
It uses three models: * A move model for what move to make * A clock model for how long to 'think' (inference takes milliseconds, the thinking time is just emulated based on the output of the clock model) * A winner model that predicts the likelihood of each game outcome (white win / black win / draw). If you've seen eval bars when watching chess games online, this isn't quite the same. It's a percentage based outcome, rather than number of centipawns advantage that the usual eval bars use.
Right now it has a model trained on 1700-1800 rating level games from Lichess. You can turn it up and down past that, but I'm working on training models on a wide variety of other rating ranges.
If you're really into computer chess, this is similar to MAIA, but with some extra models and very slightly higher move prediction accuracy compared to the published results of the MAIA-2 paper
I'm working on https://fullremote.it, which started as a side project and is now turning into a real business.
It's an "italian-oriented" curated remote-jobs and remote-work community, currently ~10k subscribers across newsletter and Telegram. What began as "let's share good remote roles" is evolving into paid job postings, sponsorships, and coaching for companies and devs.
Do you know you can hire remotely in Italy?
Spending the past few weeks building a memory layer for AI Agent's. Sync's with your business apps or remembers preferences in a consumer setting.
Planning to support a self hosted version soon -- if you'd like to give it a try ping me,.
An eBPF-based Wireguard mesh VPN in Rust using Aya.
The idea is to rely on kernel Wireguard, and process packets in kernel space (via eBPF) for maximum performance and minimal CPU overhead. I plan to use egress and ingress TC to “apply” the policy at both sides. XDP is faster, but only works on ingress, which is not sufficient for a mesh VPN imo.
Netbird already exists in this space, so this is mostly a learning exercise, and maybe a reference implementation for those learning eBPF in Rust.
Goals and constraints:
1. Single digit CPU overhead for multi Gbps bandwidth (probably a bit too ambitious, but we’ll see)
2. Linux only
3. No hole punching or complex NAT handling
4. Basic policy language for L3 and L4 traffic (L7 requires punting packets to a userspace proxy)
Still experimenting with different ways to make learning easier using LLMs.
I put together Codose as a tool where you paste a link to an Exercism or LeetCode problem, and it spins up a code editor with an AI tutor that walks you through the solution step by step, with mini lessons along the way when you need them.
You can try it without signing up but I’m on the Google AI Studio free tier right now, so I’m not sure how many uses can it handle
MemoryPlugin (https://www.memoryplugin.com)
Long term memory for dozens of AI tools, designed for power users who want more control and flexibility than native memory systems and who do not want to be locked into any one platform. You can also have the system remember your entire chat history going back years and use this information to help you better in new chats, it sometimes makes chats 10x more useful when I say something like: “Using recall tool, do 10+ calls for 1000 tokens context each to learn about my interests, strengths, curiosities, what I’ve tried in the past, what worked, what didn’t, etc and suggest a new hobby I would enjoy”.
Without long term recall, AI is a super intelligence in your hands that uses the knowledge of the world to give you generic, nearly useless advice because of how generic it is. With long term memory, you have a super-intelligence that knows YOU. This is what MemoryPlugin solves for.
I am building an analytics calculator/toy that slowly turned into something useful.
It started as curiosity. I wanted to see if I could express business logic as simple choices and let the numbers fall out on their own.
The app is built interactively in Streamlit. I do not sit down with a spec or a backlog. I add one small idea at a time, refresh the page, react to what looks wrong, then adjust. It feels closer to sketching than programming. The interface tells me what the logic should be next.
Underneath there is a growing pile of rules about the business I am in.
I do not write code in the traditional sense.
I have never coded before and are solving my own problem done by one, this sure feels like magic!!
Article about Norway problem in YAML. Draft is ready (link: https://lab174.com/blog/202601-yaml-norway ), beta readers welcome, please let me know if you have any feedback - thanks!
Also, my iOS apps; they're free and with no ads:
- Nonoverse (link: https://apps.apple.com/us/app/nonoverse-nonogram-puzzles/id6... ), a game about nonograms (image logic puzzles), now has 200+ levels.
- Polygen (link: https://apps.apple.com/us/app/polygen-create-polygon-art/id8... ), an app for generating low poly wallpapers and digital art, recently updated for latest iOS devices.
I am working on a very simple browser extension that allows you to take notes right on youtube. It is open source, offline and everything is locally stored.
Here are the links.
Chrome store: https://chromewebstore.google.com/detail/video-notes/phgnkid...
Firefox store: https://addons.mozilla.org/en-US/firefox/addon/video-notes-f...
I actually would love some feedbacks and suggestions from someone here :)
Recently, I've been very motivated to make one niche crafted service after another. For myself, family and friends. But struggled to find a compelling hosting solution for projects that only has and will have only a single user for years. I bought the cheapest mac mini M4 on sale, put it in basement and started working on some cli+daemon to help me automate all things around it. The biggest risk is security, so probably gonna rely a lot on Cloudflare at start.
I am writing a Context Retrieval and Context Minimization program for Rust code. I am trying to automate my own workflow in how I use LLMs.
The program is gonna do, what I am currently doing by hand, opening files and copying function/method signatures usually, from files all over the place.
The key here is to fetch into the context window only what is needed for one-question/one-answer and no more, hence the Context Minimization. Context fetched is gonna be specified by the programmer, for example sig(function) fetches only the signature, while @function captures the whole body. sig(Struct) is gonna fetch the fields and signatures of all of it's methods. sig(Trait) similarly.
In my view, giving the A.I. more information than needed, only confuses it and accuracy degrades. It is also slower and more expensive but that's a side effect.
The project is in early stages, for the moment it calls ast-grep under the hood, but eventually, if it works as it is supposed to, I plan to move to tree-sitter queries.
If there is a similar project somewhere I would appreciate a pointer to it, but I am not interested in implementations of agents. My program does not give the A.I. a whole view of the codebase, only the necessary points specified by the programmer.
Still working on a system for allowing Open(Python)SCAD to model cuts for a CNC using G-code, and just added support for 3D Printing via Full-control G-code:
https://github.com/WillAdams/gcodepreview
Hopefully I can restore the OpenSCAD interface layer and get it working with OpenSCAD Graph Editor:
https://github.com/derkork/openscad-graph-editor
again. Having trouble finding FCG examples which do more than move in straight lines though...
You know how Spotify (or any other service) will just silently remove tracks from your playlist because they lost the rights? No notification, a beloved song just disappeared from your life.
Building a webapp to keep track of your playlists and notify you when a song disappears
I've been working on some electronics stuff more recently, and I designed custom PCBs which read the voltage and internal resistance of the 12V NiMH batteries used by robots in FIRST Tech Challenge. It's not always easy to know how charged or healthy the batteries are for a match, or when it's time to get rid of them. They're pretty small and are easy to assemble. Rather that having a screen (which would take precious flash and pins on the ATTiny412), it just uses a WS2812B RGB LED.
I wrote the firmware in Arduino, which was a great learning experience because I typically work with CircuitPython or Go, where I'm less constrained.
I'm working on a CUDA implementation of Forman-Ricci curvature-based clustering (I checked online and saw that there is currently no GPU/CUDA implementation, so I thought why not do it). Hopefully it would help with my CV.
Github link: https://github.com/dangmanhtruong1995/Ricci-curvature-cluste...
I'm working on a native MacOS Postgres client: http://github.com/ravelantunes/Searchlight I've been slowly chipping at it over the last couple of years as I find time, and I like it enough to already use it as my primary client (although still buggy here and there).
Another itch of my own to scratch but thought I'd see if I can make some side income as well.
It's a web based bookmark manager with extensions for Firefox and Chrome.
You can easily import and export bookmarks so you have all your data whenever you need.
One main thing I really like that I think makes it stand out is the ability to export the contents of a bookmarked page to an epub file to read later on your Kindle or other e-reader device.
Looking for any constructive feedback on this! Thanks.
Pretty cool ideas here already!
I'm working on a subscription-based short-form video site called NICKEL[1]. I felt gross about using YouTube but wanted to share my gaming clips, so I made my own thing. Then I thought about making it sustainable so here we are. I'll have an update to the mailing list out in a few hours, I'm "building in public."
My feature-complete deadline is April 15th and I think I'll make it. If you want to check out the UI, visit the explore[2] page. I have it setup to redirect to a public video while I work on the intended UI (a design challenge I've never tried before but we've all seen). I'm thoroughly enjoying figuring out how streaming video works and how best to optimize things.
---
[1]: https://nickel.video
I'm in the final stretch of self-publishing my debut sci-fi novel Sector 36. Been working at it for almost 2 years. I'm waiting for the final print proofs to make sure things look good before the "official" 1/31 launch date.
I vibe coded the book website over the holiday break - https://sector36.space/
I've been serializing chapters on Substack - https://sector36.substack.com/
Built a tool that helps me re-organize my furniture on a new apartment layout (had to move a couple of times recently).
I wrapped this up a while ago, but sharing because a few friends found it useful.
https://furnimapper.pages.dev/