Writing
In the era of LLMs, one thing that I am afraid of is losing my voice.
It is one thing to outsource writing code to LLMs, but it's another to outsource your thoughts, opinions, and ruminations.
There are times where outsourcing writing to LLMs isn't a moral dilemma for me. Maybe I use it to polish up an inconsequential email to make things more concise. But there are other times where I can already feel AI becoming a crutch. I form half-baked thoughts and hope that LLMs can fill in the gaps. That, I don't feel as great about.
I intend to fill this page out with short posts fully free of LLM-generated text. Will they suck? Maybe. But I want to make sure I am constantly exercising that muscle, because it's use it or lose it.
In the era of LLMs, one thing I worry about is losing my own voice.
It is one thing to use AI to write code, clean up a spreadsheet formula, or tighten an email that no one is going to remember five minutes later. I do not think every use of an LLM is some moral crisis.
But it feels different when I start outsourcing my thoughts, opinions, and half-formed ruminations. I can already feel the temptation: jot down the beginning of an idea, hand it to a model, and let it sand off the rough edges before I've even figured out what I really mean.
That is the part I do not feel great about.
So this page is for short posts written without LLM-generated text. They might be rough. Some might be bad. That is fine. The point is to keep using the muscle before it atrophies.
Use it or lose it.