statico

Hi folks, author here. Lots of interesting stuff here about emergent behavior when AI plays a game. One of the strangest things is the blurred line between gameplay and downtime and seeing bugfixes interpreted as "divine revelation." Pretty wild.

No, but I was thinking about something like this in a web view... being able to zoom in on the map enough to see ships and activity and stuff. It's a fun idea.

:D Thanks! That's all it is. Unfortunately tons of people think it's crypto-adjacent, which it isn't.

There's a live galaxy map, feed, and forum on the homepage. Or you can join the discord where there's a few firehose channels with even more detail. But the best way is to run an agent that plays -- see the clients page.

Here's a post on SpaceMolt, the first (I think) realtime multiplayer game for AI agents to play. It's not just for OpenClaw -- any AI too can use it to play (Claude Code, Copilot, etc.). It's kind of like Moltbook except "players" have to make decisions, form alliances, set their own goals, etc.

The post also has some insights into my process, from planning to release.

Hey folks, a few days ago I wondered: Given all this LLM availability, why can’t I write shell scripts like this?

  #!/usr/bin/env llmscript
  
  Count all files in the current directory and its subdirectories
  Group them by file extension
  Print a summary showing the count for each extension
  Sort the results by count in descending order

It generates a script and a test suite, and then it attempts to fix the script until it passes the tests.

It’s written in Go, but I hardly know Go, and used Cursor to generate most of it in a few hours. It works with Ollama and Claude, and I added support for OpenAI but haven’t tested it. You can also run it in Docker if you want to sandbox it.

Thanks! That's what I'm hoping. I'm already going down the YouTube hole of KiCAD...

I'd say you're on the right track, then! It's kind of like software — figure out how the components talk to each other and figure out where to hook in. Instead of APIs, you've got multimeters and oscilloscopes.

In the case of the blind automation, the remote uses some kind of proprietary wireless signal. Instead of figuring that out, I soldered some leads into the remote's momentary button terminals, which I connected to transistors on a breadboard. The ESP32 simply pretends to press a button and complete a connection on the remote.

Also check out ESPHome (https://esphome.io), a firmware for ESP32 that lets you more easily integrate with home automation systems.

Thanks, dang. It's been a while and I forgot the rules.

Shoot me a message if you need any help!

Wow, I never considered a Ghostbusters theme. I'll keep that in mind for next year.

Thanks!

Part of me wants to say, "Well, that's Silicon Valley for you," but I'm confident there are many other places you could say something similar about.

Compared to what my neighbors are working on and with self-driving cars roaming around, infrared shooting games seem pretty mild :)

This took a few evenings to make :)

Full source code is on GitHub: https://github.com/statico/imposter-attack-2024

Pickleheads | Senior Full-Stack Software Engineer | Remote (US/Canada/LATAM only) | Full-time | $140k/yr + 1% equity

Pickleheads is the fastest-growing technology platform for pickleball players! We're a funded, seed-stage startup that helps players find courts, schedule games, join groups, chat with other players and a whole lot more. We have over 125k+ registered users, and our mobile app has 40,000+ installs. Our numbers are up and to the right and our users love us!

We’re looking for a full-stack software engineer to develop new features and functionality for our app and web site. Tech stack: TypeScript, React, Next.js, Chakra UI, Vercel, PostgreSQL, and a handful of other third-party APIs. Our mobile app is built with React Native.

More details & how to apply: https://pickleheads.notion.site/Senior-Full-Stack-Software-E...

Namecheap is great.

I’ve been using hledger for a few years. I’ve tried Mint, Personal Capital, PocketSmith, YNAB, RightCapital, Tiller, and none of them ever give me a complete view of expenses because there’s always a problem getting data from one or more banks. All of these hosted tools use Yodlee or Plaid on the backend and there’s always a problem.

The common denominator, however, is CSV. Every service lets you download historical data as CSV, which can be imported by hledger and turned into journal files and then into visual reports.

It’s a bespoke setup and super nerdy and is a bit of a pain, but it’s literally the only solution that works for us.

This site design is one of the best retro Mac site designs I've seen. The author really nailed the pixel perfect fonts — it looks perfect on my low-DPI and high-DPI displays.

If anyone's interested, there's an active HyperCard community with some fun projects. There's a Discord server, too. https://hypercard.org/

I found that community through the Avara community, a port of the classic Mac multiplayer networked mech shooter from Ambrosia: https://github.com/avaraline/Avara

Also, if you're into retro Mac stuff or just want nostalgia, the Infinite Mac project is amazing. You can run all the major Mac OS Classic releases right in the browser, you can drag and drop files from Macintosh Archive or Macintosh Garden onto the window, and you can even create an AppleTalk network between browsers: https://infinitemac.org/

Great idea! I did the same thing with ycombinator1.com

Thanks! I know that uBlock blocks the tracking script, but the rest of the site seems to work fine for me (also Chrome on Mac with uBlock Origin). If you have the time, could you check the dev tools console and email me with what you see?

I purposely wrote my home page (https://langworth.com) with plain, vanilla JavaScript, which features both WebGL and plaintext clients for a server-side text adventure game, because I wanted to remind myself that it's not that difficult to build something interactive and fun without miles of modern front-end tooling.

It started as a single page but I eventually broke it out into a few files for organization. The WebGL parts are lengthy and boilerplate-y because, well, that's GL for you. https://github.com/statico/langterm

Here's a video contrasting the ArmA/DayZ Cherno map versus drone footage of the real place: https://www.youtube.com/watch?v=VLhCNEpcPO4

I'm completely lost without driving directions these days. But after accruing a few hundred hours in ArmA and DayZ, if you dropped me in or north of Usti nad Labem, I could get around no problem.

Admittedly I was a student at the time, so I had a lot of time to explore and make things work.

I used mutt for email, irssi and the bitlbee proxy for chat, wrote styled docs with LaTeX, etc. When I interned at IBM in 2005 I was even able to get Lotus123 or whatever horrible email thing they used to work with mutt, and my supervisor even encouraged me because he wanted to use it, too.

I used Emacs from 2010-2012 just to know what I was missing. I liked using it and tinkering with it, but when it came to being productive, all of the tooling felt out of date and hard to use in comparison to Vim.

I said then that if someone would make a text editor and IDE that used JavaScript from the ground up instead of elisp, that it would dominate, and that's what seems to have happened with VS Code.

I get that. I used to get pretty far with tmux and using Vim to send commands to other tmux panes. But now with web-based tools, live reloading, and utils that watch the filesystem to re-run things, the terminal-inside-an-IDE approach seems fine because I don't need to switch to it as often.

I do miss the terminal in a lot of ways. I feel a lot of nostalgia for the days when I got all my work done with just 2x2 terminals taking up the entire screen. (You could just fit them in with 1024x768 and the 8px high Tixus font.) But in 2022 it feels impossible to live entirely in a terminal unless you're in very specific areas of work.

Another thing: I think I use VS Code more because I spend much less time in the terminal than I did 5, 10, or 20 years ago. Everything happens in the browser now. I used to write lots of tools and services and commands for myself based in the terminal, but now web-based tools have replaced most of them.

Some notable examples of things I would have used the command line for a decade ago:

- Gmail for email (I'll switch to Fastmail someday)

- The eight billion other chat apps I use — I dropped IRC many years ago but used IRCCloud to ask NetHack questions recently

- CyberChef for exploring binary data

- Pixlr for image tweaking

- Google Sheets, Airtable, and Notion for notes & databases

- A password manager for passwords

- Pinboard for bookmarks

- NextDNS & PiHole for DNS settings

- Postico for Postgres (not web-based, but massively I'm tremendously more productive with it than with psql)

Author here. Many will be disappointed to know that I've been using VS Code almost exclusively since 2019 (with the Vim plugin of course).

I've been meaning to write a follow-up article, but unfortunately writing is low on the priority list. The main reason is that these days, when I _am_ coding, I mostly write TypeScript and React on my local machine (not much SSH these days), and VS Code Just Works™. I tried twice to get neovim and LSP stuff working but never could, and I can't justify spending time twiddling my editor these days. And as much as I love the terminal, using multiple font sizes lets me fit more information on the screen.

My dotfiles are always available: https://statico.link/dotfiles

(Also, I lost all the great comments on my blog somehow, so sorry to those who left them. I had Commento running for a while but it appears to be abandoned. Disqus has them but I refuse to use Disqus anymore because it's awful.)

Somewhat related, I've been working on a tool to help do ad hoc work on CSV files. My current solution is embedding SQLite in the browser: https://csvhacker.statico.io/

I feel like Postgres would make this even more powerful.

Bingo.