I’m coming up on day 30 of 90
I had an original burst of productivity, and then fell into old ways.
I can totally see why ‘sprints’ are a thing. It’s so easy to get lost down rabbit holes and lose the forest for the trees that very short-term, focused bursts have been the only ways I can be reliably productive.
This blog post about pushing through while making games was a really good read, and seems equally applicable to tons of things beyond making games.
I struggle with focus (see: the mess that is this blog), and so I’m really trying to focus on focus. One weird method? I’ve got an elastic band around my left wrist, and I snap it everytime I start to drift off course. I don’t know how long that’ll work for, but the physical sensation does seem to be an effective re-focuser right now.
The goal is to build something simple and ship
The problem with software is it’s so powerful that while playing with buliding it, your imagination can really run wild with possibilities. Of course, it turns out even building super simple things that work well is hard, and so the minute I start getting even a tiny bit ambitious, it’s essentially game over.
So, I’m trying now to pick very short-term sprint goals towards building a very simple thing - an interface to take a legal query and search terms, retrieve the Canlii results and stick them in a vector database, and then query reliably with sources. As correctly pointed out in the blog post above, there’s no reason to reinvent the tools, so I’m using basic langchain and trying hard not to prematurely optimize with changing prompts, identifying the best vector store, etc. I get distracted a lot, and my wrist is feeling it, but we seem to be getting somewhere productivity-wise. It’s about snapping my band, simplifying at every stage, and shipping the next chunk.
Snap, simplify, ship.