Livestream Learning: How Room 302 Engages and Educates in Real-Time
Livestreaming is an important part of our practice at Room 302 Studio. We've found that it's a habit that both increases our output, and gives us an opportunity to share our explorations with the world and get immediate feedback from strangers.
We love working in public, not only to create educational resources for other people who want to learn to do work like this, but also because we think that documenting and recording our work makes our work better.
As we work on Coach Artie, our open-source internal AI Studio Assistant, we've been doing livestreams where we work on him in public, folks ask questions about the code and architecture, and we get a chance to step back from our work. Sometimes I'll use the livestreams just to [update studio members on new Coach Artie features][https://www.youtube.com/live/Pasknkewgsc] - since it's nice to have an archive of our progress, and let non studio members chime in and ask questions.
"On top of that, you're providing functions to Coach Artie, like an API to the rest of the world… " 59:45
Answering questions and trying things out in real time is half educational, half discovery- the best streams are when we try out something new together, like when Curran and I did our recent impromptu Source Code Circle Pack Livestream - he had a dataset and a rough idea, and together we were able to make something pretty quickly and try out a bunch of different solutions using Curran's VizHub multiplayer coding tool
The topics of the streams depends on what we're working on as a studio. The best is when we can take a piece of a client project (if they've given permission) and push the project forward, or if we have a personal pet project that we've been itching to work on. Sometimes that's making Line Charts with Climate Data or Redesigning The Studio Website live and in public.
Like non-streamed Mob Programming, we find that having the added pressure of an audience, combined with the help of many eyes in the chat to spot problems or suggest solutions, we end up moving surprisingly fast during our streams.
We've also been really lucky to have folks walk us through their tools and answer questions and respond to feedback. Everyone felt comfortable sharing new ideas, asking questions, and we learned a lot about the specific tool as well as broader approaches to architecting apps that work with LLMs. I want to do more of this in the coming year, and I'm hoping to put together a series of excellent guests for the fall, and we can double down on our love of livestreaming everything both as an educational resource for others, and to push ourself to make the best work we can.