How I use AI

LLMs I use

For general inquiries, I use Gemini, ChatGPT, Claude, Z.ai, Grok, etc. It’s basically just my browser now. AI just gives me a summary and I just click on the reference links in the footnotes of the output.

When coding or building apps, I use CLI tools like Gemini CLI, Claude Code, Codex, OpenCode, and OpenClaw.

How I use them

I use AI to proofread my posts, alongside Grammarly, and to keep my code documentation up to date. When writing code, I use them to build and explain errors for debugging.

They’re also handy in summarizing videos, articles, and forms so I can get the gist of the content first and return to it more thoroughly when I have more time. Beyond that, I rely on it for breaking down complex topics into simpler terms, aggregating references on subjects I’m interested in, and roleplaying — particularly for stepping into the perspective of someone with a specific condition or understanding why a person might hold a certain opinion or test.

What I don’t use them for

I don’t use AI for image and video generation, nor do I use roleplaying platforms like Character.ai to chat with “someone” (I’m not that lonely). I also don’t use it to answer inquiries from patients, interns, and students on the spot.

For a time, I used Freewrite. But I realized I’d rather talk to my mentor, my wife, or my friends and listen to their opinions (in case they have one).

Note on the em dashes (—)

When I started writing with Apple devices, a double dash would automatically become an em-dash or ⌥+SHIFT+-. I loved this feature and never let go of it. Later, I realized that LLMs use this often in their outputs. Some bloggers stopped using the em dash entirely.

Because of this I’m aware that I can be accused of using AI to write articles for me. But no. Anything with the pronoun “I” is written by me.

Just because I put em dashes doesn’t mean it’s written by AI.