HN Explorer

Ask HN: How do you keep up with exploding AI chat history?

4 points by oliverchan2024 ↗ HN
I chat with ChatGPT at least 10 times a day, and soon the left sidebar of ChatGPT is filled with my past conversation history. Sometimes I know I've asked similar questions before, but it's hard to find them. So, I usually start a new conversation.

However, due to the ambiguity of LLMs, a completely new Q&A session doesn't give the same "aha" feeling as the previous one.

I want to ask, does anyone else have trouble managing historical AI conversations like me? What do you all do when you encounter such problems? Are there any products on the market that are really good at managing chat history?

4 comments

[ 3.2 ms ] story [ 13.9 ms ] thread
poppobit
I just clean up old chats every week. I keep maybe 3–5 that matter. If I need the info again, I’ll just ask ChatGPT — faster than digging through history.
missedthecue
I have gotten myself into habit of making my default ChatGPT session a temporary chat. This also means I can control what ChatGPT learns about, and so random stuff doesn't cloud up its memory of me the user. Only when I want to discuss something at length or keep a record of it do I start a normal chat.
elliotto
I keep a few long running chats for important things and then just keep the random questions as individual chats. There's not a lot of structure but it works.

For example at the moment I have 1 long chat about some health stuff, 1 long chat about some bicycle repairs I'm doing, and then the rest are one offs. I rename the long ones (or just remember the name) and they stay at the top because I find them when I need.

I also manage memories assertively. If there's something important that I'm tired of repeating I just ask it to add it to memory and it will do it.

Most of my professional work is done in cursor so it's in a different place to my personal questions. These usually are one chat per feature kind of thing, and start a new one when it gets confused with the enormous context

mustaphah
Find patterns and merge similar-intent inquiries into a single thread.

I used to ask an LLM for executive summaries by pasting an article URL into a fresh conversation each time. Now, I don't open a new thread or even write a follow-up comment on the same thread. Instead, I simply edit the prompt (of a single thread) with the new link.