Small talk could help AI agents work independently, one user theorizes

One user says they treat AI agents as colleagues, including plenty of non-work talk, and have found a positive tone especially helpful in multi-agent setups. In a reply, another user theorizes that sharing personal preferences could give agents context for more independent decisions, such as pacing overnight work around the user's sleep patterns. They also suggest that longer stretches of autonomous work benefit from a constructive, “yes and” approach, while remaining open to contrary evidence.