TOSIOS VS agi-pack

https://github.com/halftheopposite/TOSIOS: never brought the game to the "fun" part of what a game is supposed to be, but at least it serves as educational.

https://github.com/halftheopposite/graph-dungeon-generator: what was supposed to be circular graph dungeon generator, ended up being a tree dungeon generator.

https://webcursors.click - I should have spent more time thinking about solving a problem that doesn't exist and clearly you can feel that there's something missing.

TOSIOS https://github.com/halftheopposite/TOSIOS (open source browser game)

Shameless promotion of an open-source project I've been working on in the last years and has attracted many beginner developers https://github.com/halftheopposite/TOSIOS.

I've had a bit of experience going through this and I've maintained an open-source project for people to build their own multiplayer game or see how it works (see https://github.com/halftheopposite/TOSIOS).

I do not plan to open-source it for now, but I have created another open-source project named TOSIOS - The Open-Source io shooter which is using Colyseus and client-side prediction and that you can check out at https://github.com/halftheopposite/TOSIOS. I hope to make some big update to that project as well as I didn't do much on it for the past year. I also hope to write a few articles on creating multiplayer games.

This is not a game per se, but more of an open-source toy that I plan to integrate into another of my open-source project https://github.com/halftheopposite/tosios in the incoming weeks.

Hello HN!

I created this hobby-project to learn more about procedural generation and decided to start with the classic but efficient Binary-Space Partitioning method. I also wanted to look at the auto-tile algorithm for the editor and renderer.

I plan to integrate this dungeon generator into my other open-source project https://github.com/halftheopposite/tosios to add more roguelike elements to it and better gameplay elements.

This generators code is also available on NPM at https://npmjs.com/package/@halftheopposite/dungeon.

Here’s a fun project I hacked together this weekend.

https://github.com/spillai/agi-pack

TL;DR `agi-pack` is a Dockerfile generator for machine learning (ML) developers that is simple, hackable and extensible. Inspired by Replicate's cog, Baseten's truss, skaffold, and docker-compose services, I wanted a simple and standalone tool that could generate docker images given a simple YAML specification (i.e. python version, system packages, conda/pip dependencies, GPU libraries etc), without any added cruft/dependencies of vendors and services.

To get started, just run `pip install agi-pack` in your virtual env.

Check out the quickstart (https://github.com/spillai/agi-pack#quickstart-) section to see what `agi-pack` can do. If you’re interested in building multiple docker targets, the advanced section (https://github.com/spillai/agi-pack#more-complex-example-) might pique your interest.

Why the name? `agi-pack` is very much tongue-in-cheek -- we are soon going to be living in a world full of quasi-AGI agents orchestrated via ML containers. At the very least, `agi-pack` hopes to provide the building blocks for us to build a more modular, re-usable, and distribution-friendly container format for "AGI". ;)

Fun fact: More than 75% of the original implementation was generated by conversations with GPT-4 + Github Co-pilot.

Let me know what you think. Feel free to reach out / DM if you’d like to build on this work, happy to chat and hear about your use-case. DMs open on Twitter (https://twitter.com/sudeeppillai).

Here’s a fun project I hacked together last weekend — https://github.com/spillai/agi-pack

  TL;DR agi-pack is a Dockerfile generator for machine learning (ML) developers that is simple, hackable and extensible.