Now it's about the questions that really matter The consciousness debate around LLMs has faded. Agents exposed the questions that actually matter: context, tools, control, and responsibility.
Turning an inaccessible dictation scrapyard into searchable text. Locally on a Mac, without the cloud, at usable speed.
The more thinking I put into the prompt—and the more compact the output stays—the more likely I get substance instead of generic AI slop. The prompt-to-output ratio is my practical warning signal for that.

About me

My name is Raphael Kirchner. I studied philosophy and psychology and got into programming in the mid-1990s, while still at university. One reason was my interest in how brains and computers process information, what they have in common, and where comparisons between them can be misleading. Since 2001 I have been an independent developer: https://customwebcode.com .

In 2023 I began—still skeptically—to experiment with ChatGPT, and since 2024 I have been using LLMs and agents with still growing enthusiasm as a tool for programming and as a resonance space for thoughts, ideas, and decisions.

On Aporeo, I collect

notes and ideas from my own experience with AI, shared in case they're useful to others too.