What are you doing this week?
15 points by caius
15 points by caius
What are you doing this week? Feel free to share!
Keep in mind it’s OK to do nothing at all, too.
Getting married! 🥳
I had fun creating a table plan that's easy to query for both main use cases, "What's my table?" then "Who else is at my table?". With relations between people and tables represtented as strings. (picture)
I’d like to continue my development streak of >150 days on Hister. My plan for this week is to add content import features for several bookmark managers.
My initial targets are:
I’d appreciate recommendations for other services to support.
Wow! Thats so exciting! Loving that every feature that I wish you added, you’re working on. Adding Hoarder will be amazing.
Potential suggestions: paperless-ngx, RSS feeds (with the option to delete items once read from feed reader,
That's awesome. But I do wonder what's the likelihood of you getting banned or heavily rate limited by attempting to crawl a bunch of websites you visit often? Especially now that AI companies are crawling en masse and scummy companies are selling unconsenting user traffic for those that want to bypass firewalls.
I'm guessing trying to achieve the sweetspot of how much you can do before that happens is a very fine line? I'd be interested to hear your thoughts on that.
Hister by default does not perform any scraping/crawling. The primary source of data is the browser extension which uses the content of the rendered browser tabs. This way no need to create extra HTTP requests or fight against bot detection. If your browser renders a page, Hister can index it.
Oh, so it's a search engine for your browser history. I hadn't visited your project's website and only read the readme, I thought you would have to manually set up every page you want indexed. That's a pretty cool project.
Recovering from a nasty fall off my bicycle last week... I believe I may have cracked one of my ribs, because my chest still hurts badly.
On topic-wise, I'm tweaking my website. I realized I'm not too fond of its current design and if I'm to shake it up, I might as well try to do it in a way that passes most of the nuHTML and/or Dillo checks.
Work: need to build some new images, fix a script that broke, continue overseeing training of new vendor. Gotta write some Terraform and work on some observability dashboards, too.
IRL: Went to the doctor last week and learned that my blood pressure is a little high, also saw that my weight reached a number that I always said would be my signal that I need to get on top of my physical health, so I'm joining a gym. Going to set an exercise plan and try to maintain it for a few weeks to see how it goes and then try to dive into preparing for a 70.3 triathlon.
I just finished a recent blog post on AI usage, so I'm back to working on my conference talk on Ada. I'm focusing on correctness techniques you can try out easily and borrow for other languages or tools.
Lately I've been working on my own implementation of the Nand2Tetris suite that is used to follow the course and the book. I used it a few years back with the java version and then the web version and both suck in different ways. So I'm taking the matter into my own hands and writing a C++20 alternative that uses Dear ImGUI.
My goal is to provide better error messages, have good performance and be very easy to use. I have implemented the assembler, disassembler, hack emulator and am working through getting all the other tools in there slowly.
Trying out Kimi K3 vs Claude Opus 4.8 (Fable is too token-hungry for me). It's pretty good. I got it to build a new feature for Pagecord and gave Claude the same prompt, and the features were almost identical. This was quite unexpected to be honest, but almost everything down to the method names, model names, etc were the same! It's certainly slower than Claude, and did make a couple of sloppy mistakes... but still, for $39/mo vs $90 I'll probably switch for a month. Well, as soon as it's not "sold out" which it is right now!
I have done some research on agentic workflows and caching that I will be publishing hopefully tomorrow. To help people building agentic harnesses save some money and time. I'll try to submit a PR to pi.dev as part of the core or maybe as an extension (will need to talk to that team about what makes the most sense).
Outside of that I have more updates to Thetix to deploy (https://thetix.ai). Better quality for about the same price. I also ship updates pretty much everyday to Abject (https://abject.world) which is a research project that I ended up using daily.
Trying to set up WSL2 to run PI with paseo.sh frontend in Docker containers, that consumes the LMStudio Gemma (running Windows) and open router models.
On paper (aka Gemini output about the setup) it should be a breeze, but hitting some Windows config issues all the time. The flip side, learning a lot about Windows and WSL2 configs.
Hopefully I can get this running on the side so I can try out some simple automation with only electricity cost.
It's the first week of summer vacation for me and thus also the first of 4 weeks of writing software for money.
In my free time, I'm working on a small SSG that renders jj repos to a website.
Last week before my vacation, trying to remain productive at work. Doing a deep dive into Scheme macros for my own projects, it's fun but I'm having some trouble keeping track of what happens when.
Returning back to work from a full week in Missouri, where I went to attend Software Should Work!
Aside from my day job, I'm working on my side project I call Popsicle Boat, a little Elixir/Phoenix social site I run solo, but the part that's been fun this week is playing with using it as the comment section for my own static blog.
Each blog post links to a discussion thread hosted there, there are no dependencies to import, just set up an account and a space for your blog and add a few lines to wire it up.
It's super fun to work on, I don't expect many users or to change the world or anything but I am truly proud of it. The functionality has all been there for a lot of years, mostly at this point I'm just polishing.
My blog is jva.lol if anyone wants to poke at the comment thing.
I recently restarted an OMAD diet/routine after a long break. I'm hoping to do a longer fast this week - probably 72 hours to start with, and then we'll see how long I can go on. Having been into intermittent fasting for a decade, I've never done longer fasts - hoping to change that now.
Apparently I’m going to be doing some research into vestibular migraines and heartburn and potential comorbidities. Had my first migraine in a decade last night, followed by waking up at 3am with brutal heartburn that’s only starting to fade now, four hours later. Haven’t figured out any potential triggers yet, yesterday was a very ordinary Sunday.
This week I’m updating WordJoy, a Chinese-character puzzle game, based on player feedback.
The current demo is a compact solitaire-style puzzle: component cards are dealt into seven stacks, and only the exposed top cards can be moved. A component can stack only onto one with exactly one more stroke, so you need to untangle the layout and uncover the highlighted target components. Release them in the required order to complete two Chinese characters and clear the level.
The demo currently includes six levels. I’d especially appreciate feedback on rule clarity and whether the character-building goal works for players who don’t read Chinese. Demo: https://chunq.itch.io/wordjoy
Getting everything done before I can enable my OoO and spend two weeks swimming in Austrian mountain lakes.
Had to travel to the office in Tennessee this week. Will be attending my company's leadership cohort. I need to decide if I am going to talk to my boss' boss about how or lack of planning and focus over the last 6 years is costing us millions in VMWare mistakes or just wait for my next job offer to hit my inbox and take it.