RC Half-batch Retrospect

When I plan my days I have a lot of tasks that I would like to get finished, alas, at the end of the day, most of these tasks remain unfinished. I need to be more realistic. Additionally, I should do some meta-planning where I use certain days for certain related tasks instead of having many unrelated tasks (e.g. mixing admin with writing code and reading papers). This is something that I continue to work on.

I had a chat with some of the RC admin about how it has gone so far. It was super chill but very insightful. While some things have not gone as I wanted (does anything ever go according to plan? Don't think so!) it was reassuring to get some outside point of views which also highlighted what I did accomplish, like setting up this website and the more meta-skills of building volitional muscles etc.

The first sentence of my Week 6, Day 5 reads

Removed a lot of coffee chat slots, limiting them to just Fridays from now on. Think I need to go more into deep work mode.

I think this sets the tone for the later part of the batch for me. The social aspect of RC has been great fun and I've made loads of friends and connections, but it takes energy and janks you out of your flow doing all of these scheduled coffee chats. Once per week will be enough from here on.

I will focus less on the social aspects of RC and more on getting better at coding. I want to finish my project before the end of batch, and if I do it before that I want to maybe try to create a Jax package. Here's to the next year and the next RC batch!

I managed to get my workstation set up properly. Now I have a beefy gpu which I can use on-demand from my laptop. I also went a bit crazy and set up a VPN so that I can use jupyter together with emacs. For this I relied a lot on this blog post on emacs and jupyter which saved me many hours. I found the code-cells emacs package to be very well-engineered and striking a good balance between light-weight and feature complete.