By skyforbes For a long time I did the usual thing, telling ChatGPT to act like a senior expert, consultant, strategist, whatever fit the task. Sometimes it worked, sometimes the answers felt stiff, overconfident or just kinda fake smart. Recently I tried something different almost by accident. Instead of asking it to be an expert, I asked it to just be a neutral conversational partner and help me think stuff through.
The difference was honestly more noticeable than I expected. The replies became simpler, less preachy, more like someone reacting to my thoughts instead of lecturing me. It started pointing out obvious gaps in my logic without trying to sound impressive, and asking clarifying questions that actually helped. I also noticed I was typing more naturally, like I was talking to a person, not trying to engineer the “perfect” prompt every time.
Now I mostly use it this way when I’m stuck or unsure. Not for final answers, just to untangle my own thinking first. It feels less like using a tool and more like borrowing a second brain for a bit. Kinda funny how lowering expectations made the output feel more human and, weirdly, more useful.