The persistent memory is spotty many of the times; despite proper governance rules in place, continuous doc uploading, etc., it still forgets. Also, it feels like you need to be a lot more detailed and specific when it comes to medium-complexity coding versus Claude AI in the chat or Claude code, where you can be quite general to a certain degree, and it can output with quite good accuracy.
Deeply attuned' sounds like a high bar for emotional intelligence. Is the AI's personality static, or does it evolve based on the long-term history of your interactions?
Hi @luke_cartright, you mentioned parallel agents running tasks together. If two agents try to modify the same file or function at the same time, how do you handle the merge? Is it last‑write wins, or do you have a conflict resolution system? I am curious about it.