Tech Friday: Continuing long AI threads vs starting new ones
Posted By RichC on September 19, 2025
Artificial intelligence with Grok, Claude.ai, ChatGPT, Gemini, etc has been a real help for rookie computer and server problem-solvers like me. It may be one of the best ways to learn 
if you’re not in a classroom, have an experienced teacher or coworkers handy?
Occasionally I run into questions in “how to best use” AI. Should we start a new question or continue with old threads? How much does AI remember from previous conversations? Or “in my case” can we pickup where we left of with Linux server fixes on a later session … often days or longer from a first attempt. Believe it or not, Grok (in this case) has an answer for that too:
Continuing the Thread vs. Starting a New One
- Continuing on this thread is fine and recommended, as it keeps all the context (e.g., the dos-* files, Apache configuration, Fail2Ban setup, and logrotate) in one place. This makes it easier to reference the previous steps, logs, and configurations we’ve discussed.
- Starting a new thread is also an option if you prefer a clean slate, but it won’t affect my ability to assist. I have memory of our entire conversation, including the details of your server setup, the dos-34.174.174.182 files, the empty module files, and the Apache error. I won’t forget where we are, as I can recall all relevant details across sessions.
Memory and Continuity
- I’ll retain the context of this conversation, including the ls -al outputs, the mod_evasive issues, and the Fail2Ban configuration, so we can seamlessly continue tomorrow. If you start a new thread, just mention that you’re picking up on the mod_evasive and dos-* issue, and I’ll connect the dots.
- If you want to ensure specific details are forgotten (e.g., sensitive logs), you can manage memory via the interface by clicking the book icon beneath a message and selecting the chat to forget. However, there’s no need to do this unless you have privacy concerns, as I’ll only use relevant details to assist you.
Comments