I started a UI challenge and decided to make a web clone of Ubuntu 22.10 🎉
I've done some basic starting features and layouts of it and I just randomly decided to make it opensource for anyone who wants to contribute to it and expand on it. Which'll be cool for you guys to put it up on your resumes or project showcases. ✨
Just some more details: it took me 4 hours to do it. 1 hour went into looking for similar icons and adjusting them to fit. Another hour prolly went into slacking off. 😅 It's a pretty straightforward project.
If you're a frontend dev try doing something like this. Clone Win10 or Win11 or another OS you love. 😊
Killian, who made Open Interpreter, is now planning to make an open source version of R1, the hardware AI companion. An open source Language Model Computer (LMC) in just 6 weeks.
In less than 48 hours, 200+ engineers and designers around the world, including me, have signed up to contribute in any way we can.
The plan is to make an open-source version of R1 pretty quickly with a Raspberry Pi and Open Interpreter. Fully offline, <$50 for parts. Using Whisper, Mixtral, Linux, 1 button and a good local TTS.
They are also planning to publish a family of standard protocols to advance the LMC ecosystem as a whole.
Here're epic resources for creating, managing and maintaining Open Source Projects.
1. Open Source Guide • This's a very extensive and detailed guide into everything you need to know about OSS. It's got resources, tools, explanations, templates, checklists and so much more! And it's open source so you can expand on it • https://opensource.guide 2. Choose a License • This simple site helps you choose a license for your projects, and it can create a PR automatically, by simply giving it your repo link • https://choosealicense.com
4. Contributor Covenant • This template is a very generic but very good to use code of conduct to make the project a nice place for contributors to work in. • https://www.contributor-covenant.org
There are two versions; The 01 Light and Heavy. The light is an ESP32 based voice interface that must connect to the 01 server while the heavy one is a standalone that runs everything locally.
The 01 exposes a speech-to-speech websocket at localhost:10001. If you stream raw audio bytes to / in LMC format, you will receive its response in the same format.
The hardware and electronics are literally the most simplest thing ever. Like anyone here with basic arduino skills can make it in a few minutes.
It has Dynamic System Messages meaning it can execute code inside the system messages.
# Anything in double brackets will be executed as Python interpreter.system_message = r" The time is {{time.time()}}. "
# It will know, without making a tool/API call interpreter.chat("What time is it?")
They also have introduced an LMC Messages protocol that extends OpenAI’s messages format to include a "computer" role to ease communication between different components of the system.
The fact that they have Open Interpreter made everything so easy, it's almost like only configuration was done. Smartest idea for implementing this cause not only do you not have to sign up or subscribe you could just let the OI run on your device and use it like a person.
