Hi everyone,
I’m running into a consistent and critical issue with a Pickaxe I’ve built, and I’m wondering if anyone else has experienced the same or found a workaround.
The issue:
After the third or fourth message in a conversation, the AI starts replying to the previous user input instead of the most recent one. This completely breaks the conversational logic and makes it nearly impossible to hold a coherent, flowing interaction.
Example:
- U: “Hi!”
AI: “Hi! What’s your name?” - U: “How can I change my organization?”
AI: “Hi! What’s your name?” - U: “I’m Gerónimo. Inspired by Semler.”
AI: “Tell me more about your frozen organization…” - U: “Hello?”
AI: “Sure, I know Ricardo Semler!”
It feels like the AI is stuck one turn behind, creating a frustrating and confusing experience for the user.
What I’ve tried:
- Toggling memory and knowledge settings
- Adjusting prompt design
- Testing in new sessions, devices, and browsers
- Rebuilding the Pickaxe from scratch
No change.
Why this matters:
This bug makes it impossible to deliver a smooth, reflective user experience — which is essential for the type of conversational AI I’m building (coaching/facilitation style). Users get lost, responses come out of sync, and the whole thing loses impact and clarity.
Anyone else experiencing this?
Would love to know if this has been reported already or if anyone has found a workaround. Any help or insight would be massively appreciated!
Thanks a lot,
Gerónimo