I don't understand how people use these apps with memory enabled. I am always carefully controlling the context of each conversation. The idea that past conversations could bleed into current ones is unthinkably terrible.
Yeah, I turned it off completely after it once started generating random garbage in long conversations with memory enabled. Like, it would give completely unrelated replies: I'd ask about Go, and it would tell me about the weather in Paris (or something else we discussed once months ago). Maybe it was just a temporary bug, but it was an eye-opener that context poisoning exists.
I'm not talking about deleting conversations. Anthropic's guide isn't going to actually move your conversation history anyway. The purpose of this feature is to move over specific memories which the AI can use in future responses.
But I have this feature turned off, and I cannot imagine ever wanting to turn it on, because I am always thinking carefully about what the AI "knows" when it generates a given response. For example, since I know that the AI always wants to make me happy, when I ask for an "opinion" I'm careful to not let the AI know which answer I'd prefer. I'll often try phrasing the question in different ways to see if it changes the outcome.
kgeist|3 hours ago
downboots|12 hours ago
Wowfunhappy|11 hours ago
But I have this feature turned off, and I cannot imagine ever wanting to turn it on, because I am always thinking carefully about what the AI "knows" when it generates a given response. For example, since I know that the AI always wants to make me happy, when I ask for an "opinion" I'm careful to not let the AI know which answer I'd prefer. I'll often try phrasing the question in different ways to see if it changes the outcome.