Harrison Chase outlines async subagent notification options for deepagents

1. 3:23 PM
   Harrison Chase@hwchase17X

   Harrison Chase says agent harness work is moving to infrastructure layer

   AIHarrison Chase says the most interesting agent harness work is at the intersection of harness and infrastructure, citing dreaming, async subagents, heartbeats, and agent-scheduled reminders. He says this is a reason he is bullish on managed agent platforms, and that all four belong in managed deep agents.

1. 5:23 AM
   Harrison Chase@hwchase17X

   Harrison Chase outlines async subagent notification options for deepagents

   AIHarrison Chase says LangChain is working to make deepagents more async native, with better support for async subagents. He lays out two options for notifying a parent agent that is still running: interrupting it to apply the update, or batching the update after the run finishes. For a paused parent awaiting human approval, he proposes either batching the update with the human response or resuming the thread, which would cancel the interrupt.

The trend compares only continuously observed participants, so its scope may be smaller than the current heat total. Hover or tap for hourly heat; use arrow keys with a keyboard.