Thoughts On: Understanding is the new Bottleneck
Understanding is the new Bottleneck
As AI helps us create new software and update existing software it can be tempting to let it just go and do it's thing. There are constant videos and articles about how it does some amazing thing like rewriting SqlLite from the ground up to updating 30000 lines of code overnight. However, if you don't keep on top of it you'll eventually get stuck.
Last week I was explaining to a fellow developer how to use AI to look things up in git history. There is good information in git commits, but other than using blame I almost never read through them. However, it's all different now. Wen coming across a feature in the software I don't know or am unsure about, I can ask AI to go and read the docs, read the code, read the git history, and present me with a timeline of ideas/decisions/changes that demonstrates how the feature came to be.
Doing so gives me an idea of the original intent of the feature, improvements made over time, how old it is, how often it is used (if analytics are available) and how to move forward. Should it be merged with something similar? Is it redundant and can be dropped to simplify the software? Is it never used except in some rare edge case?
As AI is great at writing large amounts of code, many times it will duplicate features. For example, a dashboard that shows the same thing 3 different ways is fairly common. Repeating itself in a report by saying the same thing in the introduction, the data, and a summary at the bottom.
Using AI to walk you through something is a great way to familiarize yourself in a way you would have never been able to do before. Identifying features you can remove to simplify software is also great. And hey, if you made a mistake, AI can look into the git changelog and resurrect a feature in no time at all!