In the age of AI, I started writing a diary in my own words

Introduction

Hi, I’m kittchy, an ML engineer. Thanks for visiting my blog.

I built this blog and then left it alone for about a year, so I’ve decided to put it to use by keeping a diary here.

Why a diary now?

You might wonder why I’d start a diary when AI is everywhere. Today you write a prompt and AI produces the text right away, even polishes it for you, so putting articles out into the world is easier than ever.

I’m starting my own diary anyway because I’ve come to feel these three things.

• My vocabulary and writing skills are declining
• I’m tired of talking to AI
• I want to give the AI/ML community living writing as open data

My vocabulary and writing skills are declining

At work I’ve recently learned to write text at top speed with Notion and ChatGPT. But the text I end up with isn’t something I thought through myself. It’s AI fan fiction. I just keep copying and pasting what AI recommends, so when I reread it later, I don’t really understand it. That kept happening.

If this goes on, my ability to write on my own will keep slipping, and I feel my ability to get things across to people is slipping along with it.

The best way to build vocabulary is probably to think with my own head and put things out. So, to build that habit, I decided to write a diary using AI as little as possible. (In the end I did hand this article to Claude for polishing.)

I’m tired of talking to AI

These days I often work through throwaway code and implementations by talking to Claude Code. But don’t you find it tiring to give orders to AI in natural language?

I had gotten tired of loosely instructing AI in natural language at work too. I work remotely, so the line between time spent talking to humans on Slack and time spent talking to AI has blurred. I’ve started thinking about how to word things for AI without meaning to, and it has made work quite awkward.

To cut off that conversation with AI for a while, I decided to write freely in a place that only I, or a few people, will read, and shake off my everyday AI fatigue.

Giving the AI/ML community living writing as open data

In November 2025, the Wikimedia Foundation asked AI companies to “stop scraping and use the paid Wikimedia Enterprise instead.” AI companies’ bots were crawling Wikipedia heavily and loading its servers, and some even disguised themselves as human visitors to avoid detection.

Wikipedia urges AI companies to use its paid API, and stop scraping | TechCrunchWikipedia, the online nonprofit encyclopedia, laid out a simple plan to ensure its website continues to be supported in the AI era, despite its declining traffic. TechCrunch

I think the Wikimedia Foundation is right. If we want sustainable AI/ML development, we should give respect and money to the people who provide data to the web for free.

Still, it honestly frustrates me if that slows down development and progress in AI/ML. So, as someone who has worked on ML, I decided to keep a diary, thinking that even my clumsy Japanese could add a little to the diversity of the data.

Closing

Honestly, I don’t know how many days this will last. It might end in three days, or it might turn out to be more fun than I expect and keep going. So far it’s surprisingly fun.

I’d be happy if you dropped by now and then to read what kittchy has been thinking and going through.

See you.