By skyforbes I’ve started noticing something interesting: it feels like ChatGPT’s responses are influenced by what I’ve discussed with it in the past. For example, when I ask for ideas for customer projects, the model tends to focus on a specific product area that I’ve worked on before in my previous chats.
A colleague of mine — who uses the exact same prompt — gets completely different ideas that fit his area of focus instead. It really seems like ChatGPT “learns” from our previous interactions and then keeps steering future outputs in that same direction.
Has anyone else experienced this?
And more importantly — is there a way to make ChatGPT ignore past conversations and respond completely independently, as if it’s a fresh model with no context or bias from previous chats?
Would love to hear how others deal with this.