A JS framework called Torpor: https://torpor.dev (https://github.com/andrewjk/torpor)

Been working on this for a while, with the aim of making something simple and immediately usable. Components are JS functions, containing UI that is (mostly) HTML, with reactivity only done through proxied objects.

To test it out I built a distributed social media/microblog site called Redraft: https://redraft.social (https://github.com/andrewjk/redraft)

Self-hostable or hosted for a fee, all your posts stored in an SQLite file, comment/like/react on other people's Redraft posts via a Web extension (pending approval, but you can use the unpacked version from source).

Both are in a pretty rough state, but usable for the intrepid.

A Parquet file compactor. I have a client whose data lakes are partitioned by date, and obviously they end up with thousands of files all containing single/dozens/thousands of rows.

I’d estimate 30-40% of their S3 bill could be eliminated just by properly compacting and sorting the data. I took it as an opportunity to learn DuckDB, and decided to build a tool that does this. I’ll release it tomorrow or Tuesday as FOSS.

The stuff I'm working on never feels worth sharing, but I am doing a lot of computer stuff lately. It's kind of the year of moving towards declarative setups for me.

- Migrating to Niri on my laptop and re-evaluating my literate config approach, switching from xkb configs to kanata and a few other QOL changes to make my tooling more composable and expressive

- Shoring up my blog / media sharing infrastructure (migrated to a landing page on an s3 bucket, with different prefixes for several different hugo deployments for different purposes, still need to get better about actually posting content)

- Preparing to migrate a bunch of my self-hosted services to a k8s cluster which can can be fully deployed locally for testing and defined in code. All this is managed through argo and testable with localstack and crossplane for some non-local resources

- Attempting (somewhat unsuccessfully) to setup a nixos config for a bunch of services that just don't feel right to run in containerized stack that I want to live in ec2 and have as close to 100% uptime as possible (uptime kuma, soju/weechat relay/bitlbee, conduit, radicale, agate, whatever else I think of that is small and has a built-in nixOS service module. Thinking about some kind of RSS aggregating solution here as well)

- Experimenting with vibecoding by trying to get an LLM to do the legwork to build a TUI interface to ynab using rust (which I don't know how to write)

I'm hoping that by the end of this summer most of the tooling I use for most things will be way more concrete and seamless. I also want to get my workflows down and get on top of converting at least a few the ~100 draft blog posts I have laying around into something I can actually post. Ditto for my photography albums, which are not yet organized into coherent groupings or exported for web.

Syntax highlighting caught my interest after I created a data format. I stumbled upon TextMate grammar bundles which are supported in some editors and created a bundle that works with three of them. The gnome text editor uses a different language definition format for which I created a grammar file as well. [1]

To highlight the syntax in the browser I checked out the CodeMirror project that uses Lezer grammars. It is very flexible and allowed me to implement additional features like custom folding. [2]

I would also like to create a grammar for tree-sitter, finish the Java implementation and documentation of the ESON parser before I try to implement it in other languages.

[1] https://gitlab.com/marc-bruenisholz/eson-textmate-bundle [2] https://gitlab.com/marc-bruenisholz/eson-lezer-grammar

I am working on Community Notes for Bluesky: https://github.com/johnwarden/open-community-notes

My Carnatic Raga classifier is progressing very well. I am now training a classifier to identify 142 ragas.

A bit of background: I have been working on a Raga classifier since November of last year - I started with just 2 ragas and a couple megabytes of audio. After experimenting with a lot of different ideas and Neural Net Architectures, I finally landed on one that could scale. I increases to 4 ragas, then 12, then 25 and then to 65.

All the training is done locally on my desktop (RTX4080, AMD 7950X, 64G RAM). My goal is to make an app for fast inferencing (preferably CPU) and to get this app in the hands of enthusiasts so that I can get some real data on its efficacy. If that goal is hit, then my plan is to iterate and keep increasing the raga count on the model and eventually release to the public. As long as I can get the model to either run locally or for very cheap on server, I hope to not charge for this.

It has been an amazing learning experience. The first time I got a carnatic singer to sing and the model nailed almost all ragas was the highest high I've felt in a while.

Not a technical project in the typical sense, but I finally started working on a satirical faux-exposé series about my years working in the industry.

If anyone's interested in that kind of thing:

https://massiveimpassivity.substack.com/p/softcore-how-nobod...

We're building Zigpoll (https://www.zigpoll.com), a survey platform focused on zero-party data collection — think post-purchase attribution, customer feedback, and segmentation — all done directly on your site without relying on third-party cookies or offsite links.

We initially built it for Shopify, but now it’s fully embeddable, supports headless implementations, and integrates with tools like Klaviyo, Zapier, n8n, and Snowflake. One thing we’re especially proud of is how fast and unobtrusive it is: polls load async, don’t block rendering, and are optimized for mobile and low-latency responses.

From a tech angle:

Frontend is all React, optionally SSR-safe.

Backend is Node.js + Postgres, with a heavy focus on queueing + caching for real-time response pipelines.

API-first design (public API just launched: apidocs.zigpoll.com).

We recently open-sourced our n8n integration too.

If you're a dev working on ecom, SaaS, or even internal tooling and need a non-annoying way to collect structured feedback, happy to chat or get you set up. Feedback welcome — especially critical stuff. Always looking to improve.

I'm building a budget app for my wife and myself.

Basic goals:

- Web based for zero update latency

- Have it work offline

- Automatically import transactions from my banks

- No running/hosting cost

- Secure

Tools used so far:

- InstantDB for the datastore, providing the offline capability too

- A gmail account that automatically gets forwarded bank alerts for purchases

- Gitlab.com w/scehduled pipelines for cron based email-syncing

- Netlify for the free hosting

- InstantDB magic codes / email links for securing the data

I'm at the point where I can track and categorize purchases, including split transactions.

Next steps:

- Add in date ranges for reporting / data views; e.g. show expenses incurred in a one month period instead of for all time.

- Add in planned / project transactions for month forecasting

- Statement import & import reconciliation and statement reconciliation

- Scrape company specific digital reciept emails (like Amazon) to autopopulate more transaction data

And that'll be the end of the stuff I can do for free. I think I will add features that require money and/or dedicated hardware though:

- OCRing receipts -> autopopulated transaction data / description

- Using chatgpt to suggest categorizations

- Scrape extra data from my bank sites, like physical addresses of entities involved in charges.

Insurance is negative NPV. Trying to make it NPV neutral by giving people tools to self-insure. Starting with an app that lets you self-insure your phone with friends and family.

https://apps.apple.com/us/app/open-insure-self-insurance/id6...