Continual improvement

Continual improvement has been something on my mind recently. In large part, I think it's a reaction to the increasingly widespread usage of Large Language Models which I have the misfortune of hearing every day. Over time, it's solidified into something of a mindset I hold and believe is important in the (theoretically) coming age of automation.

The crux of my mindset of continual improvement is that almost everything I find myself needing to do can form some kind of practice or learning experience. Writing documentation will help me practice conveying information, writing unit tests helps me better understand the code I have, and coming up with TTRPG character names on the fly is an exercise in improvisation. Each of these tasks are things I see LLMs gradually expanding into, and for some of them I completely understand why. I don't particularly enjoy writing documentation and would love to get around it if I could. But I think being forced to do it allows me to practice my skills at it. By writing down the details of an API or a process, it gives me time to reflect on how I built it and whether or not the instructions sound insane for someone else to follow. Similarly with a unit test.

What it comes down to, I have found, is deciding what processes I can do without. I have no issue with the washing machine automating away my laundry, why if I could also have it hang my laundry out to dry I would, because I have no desire to become better at washing clothes. What I do care about is being a good programmer, game master, communicator, and so on and so forth until the cows come home. And so I try and understand every small process as a step in practicing and polishing the skillsets I care about.

I've tried to expand this mindset throughout my life. I recently started reading non-fiction books on subject matters that I want to be knowledgeable on, having felt like my existing media diet has left me rather stagnant with my understanding of most things. I've always been one to read the Stack Overflow conversation threads, rather than just taking a quick answer and dipping, because I value the insight it gives me into the problem I'm trying to debug. Lots of small things which can probably be quickly automated over are actually opportunities for growth, and I feel ignoring them for the sake of efficiency is only likely to stifle my growth in the long run.

With LLMs I feel things have become the most egregious. A chatbot today can solve all kinds of problems (to more or less of a degree) but using one is really no better than offloading your chances to grow in return for an instant payoff. I think it's a continuation of the trend for instant gratification, results right now at the cost of anything else. Regardless, you can count me out of offloading my life to an amalgamation of my own words and thoughts sold back to me.