One thing that remains consistent in all versions (5, 4o, whatever) is that ChatGPT often confuses who said what.

Example in a roleplay session:

GPT: Suddenly, a gigantic red dragon swoops down.

Me: I order my men to flank the dragon and start firing poison-tipped arrows.


Later, discussing the session:

GPT: That was a particularly thrilling moment when you introduced a red dragon to the scene.

Me: …uh, you introduced the dragon, not me.

GPT: You’re right! My bad. I’ll try to keep track of who said what going forward.


Of course, it doesn’t and inevitably makes the same mistakes later.

I’m not posting this just to complain, more that I find this to be an interesting & consistent issue and to confirm it’s not just a me issue. It’s a little annoying even still. One would think it wouldn’t be so difficult to track this given it’s obvious who said what in a chat session.

Leave a Reply