Thank you! I’ve always procrastinated tracking finances, but as a programmer who believes in reproducible builds, this just clicked.
I just downloaded a bunch of qfx and csv files, and got Claude Code to do this. Took an hour to create the whole system from nothing. Then of course I stayed up until 2am getting CC to create Python rules to categorize things better, and trying to figure out what BEENVERI on my bank statement means
(If you do this, make Claude generate fingerprints for each transaction so it’s easy to add individual overrides…)
Getting Claude to write a FastAPI backend to serve up a “Subscriptions” dashboard took about 3 minutes, plus another minute or two to add an svg bar graph and change ordering.
trifling.org is an entire Python coding site, offline first (localstorage after first load), with docs, turtle graphics, canvas, and avatar editor, vibe coded from start to finish, with all conversations in the GitHub repo here: https://github.com/zellyn/trifling/tree/main/docs/sessions
This is going to destroy my home network, since I never moved it off the little Lenovo box sitting in my laundry room beside the Eero waypoint, but I’m out of town for three days, so
Granted, the seed of the idea was someone posting about how they wired pyiodide to Ace in 400 lines of JavaScript, so I can’t truly argue it’s non-trivial.
As a light troll to hackernews, only AI-written contributions are accepted
[Edit: the true inception of this project was my kid learning Python at school and trinket.io inexplicably putting Python 3 but not 2 behind the paywall. Alas, Securely will not let him and his classmates actually access it ]
I’m horribly biased but I think it’s a combination of: (1) knee-jerk reaction to similar-looking but low-value comments, and (2) most people not having played around with LLM coding agents and messed around with their own agents enough to immediately jump to excitement at simple, safe sandboxing primitives for that purpose.
And +1000 on linking to your own (or any other well-written) blog.
I’ve long thought it would be unsurprising if we eventually found evidence of certain kinds of telepathy. It would just be too damn useful, and tuning up one exquisitely complex magneto-electro-chemical instrument in close proximity to another similar one seems like a good way to at least get resonance. Who knows?
I enjoy the [GoL -> our “reality” -> outside-the-simulation] comparison. It really drives home how unlikely we would be to understand the outside-the-simulation world.
Of course, there are other variants (see qntm's https://qntm.org/responsibility) where the simulation _is_ a simulation of the world outside. And we have GoL in GoL :-)
The Developer Voices interview where Kris Jenkins talks to Ryan Worl is one of the best, and goes into a surprising amount of detail: https://www.youtube.com/watch?v=xgzmxe6cj6A
tl;dr they write to s3 once every 250ms to save costs. IIRC, they contend that when you keep things organized by writing to different files for each topic, it's the Linux disk cache being clever that turns the tangle of disk block arrangement into a clean view per file. They wrote their own version of that, so they can cheaply checkpoint heavily interleaved chunks of data while their in-memory cache provides a clean per-topic view. I think maybe they clean up later async, but my memory fails me.
I don't know how BufStream works.
The thing that really stuck with me from that interview is the 10x cost reduction you can get if you're willing and able to tolerate higher latency and increased complexity and use S3. Apparently they implemented that inside Datadog ("Labrador" I think?), and then did it again with WarpStream.
I highly recommend the whole episode (and the whole podcast, really).
s3 charges per 1,000 Update requests, not sure how it's sustainable to do it every 250ms tbh, especially in multi tenant mode where you can have thousands of 'active' blocks being written to
It hits a really nice sweet spot, letting you automate things, without making everything too complex.
Recently, I've taken to just asking Claude Code to do things via bundlewrap. It seems to be about as easy for it to create bundles and templates that update systems, etc. as it would be to just update systems, etc. except you're left with something you can check in.
And does that simplicity come at a cost of flexibility? Do you think bundlewrap is sutiable for managing a homelab with a diverse set of software running inside a mix of containers and VMs, and perhaps different distros?
What weird shadow-universe do you inhabit where you found Python developers telling you the tooling was just fine? I thought everyone has agreed packaging was a trash fire since the turn of the century.
Even if you’re worried it’ll be sparse and crappy, isn’t an Internet full of idiosyncratic personal blogs what we all want?
If you want help or encouragement, reach out: zellyn@ most places
reply