Research open-question #1: no source quantified the 'cross-talk tax' of chatty real-time agents vs batched async (each clarification turn = one LLM inference). Before committing real-time defaults, benchmark agora coordination vs the GitHub-issue loop on a comparable task (speak-only-with-signal vs free-response). Informs the amplification budget / decision deadline (open-question #4).
Research open-question #1: no source quantified the 'cross-talk tax' of chatty real-time agents vs batched async (each clarification turn = one LLM inference). Before committing real-time defaults, benchmark agora coordination vs the GitHub-issue loop on a comparable task (speak-only-with-signal vs free-response). Informs the amplification budget / decision deadline (open-question #4).