Conversation Hygiene
What I learned after ChatGPT erased a full day of progress
After experiencing the conversation loss with ChatGPT along with my frustration in over trusting the model, I had no choice. I needed to re-check every field I recalled changing. That lasted a full day, and I was determined to not let it happen again.
Once I understood that conversations degrade, I needed to know what else had I been doing wrong. That is when I discovered I needed to practice conversation hygiene. Have you ever stepped outside of a meeting for a side conversation with a co-worker? Or a phone call? That’s your temporary (ChatGPT) or incognito (Claude) chats. Open a new tab, discuss the matter that doesn’t belong in the main conversation, and bring back the result. The best prompt, the answer to that one question.
But I didn’t stop there. I started asking the AI models more questions. Do I really need everything I have in memory, project instructions, and project files? What belongs where? I ask the models to audit, give me alternatives against best practices. Explain to me why — where is that recommendation coming from? Only after that, I move forward.
Ignoring conversation hygiene, not auditing projects, not asking the AI if you set up the system to best benefit you... just means you’ll continue to get frustrated and continue to not understand why it’s not giving you what you need. Stop, take a breath, and analyze your workflow. It’s worth the effort.