So I was using ChatGPT in the browser (that’s the only way I use it), and I glanced up and saw a little icon that said “>_ Codex”. Ok, I have a bit of time. Let’s check this out.

Within minutes, I had given OpenAI access to my Github, installed the Codex CLI app, and was able to create agents locally. Scary easy.

I have been curious (and honestly, a bit horrified as well) about the whole “agentic” thing and I have wanted to try things out, but I hadn’t taken the time because I thought it would require quite a bit of setup and I was also afraid of the agents doing something stupid to my machine (due primarily to mistakes my own permissions or instructions).

I don’t know the first thing about what to put in AGENTS.md, which seems to be how you configure an agent for Codex, but I think I will be exploring this further.

Are we flying too close to the sun? Only time will tell!