Anthropic: fold most subagents back into your main agent loop
_AustinCalvert_ · x · 2026-10-25
Anthropic's new post argues most multi-agent systems are just Conway's Law for AI—specialist subagents plus an orchestrator yield lossy handoffs, subagent sprawl, and 2x latency. Across enterprise deployments, a single main agent with progressively disclosed skills and tool search often beat both one-big-prompt and subagent designs on quality, cost, and latency (e.g., a customer-facing agent with a dozen specialist subagents that hurt consistency and speed). Subagents still make sense for true parallelism, context bloat, and permission isolation; Anthropic ships a migration kit to identify and fold the rest. As IntuitMachine puts it: most of it is organizational debt wearing an agent costume.
More from coding & agent
- Graphify, a 135K-star repo, maps codebases into queryable knowledge graphs for AI agents — techNmak · 2026-10-25
- Matt Pocock: deterministic checks could safely automerge refactor PRs under strict rules — mattpocockuk · 2026-10-25
- Streamrunner ships: AI-powered native macOS reimagining of OBS Studio for $39.99 — jasonkneen · 2026-10-25
- Dev shares crude cross-interface agent setup: Linux box with limited apps plus CUA monitoring — tokenbender · 2026-10-25
- Dev uses agents to monitor group chats and send greetings on his behalf, eyeing voice cloning next — tokenbender · 2026-10-25
- agentic-gate: open-source layer blocks LLM agents' bad tool calls, 150→0 in benchmark — Nice_Current_8581 · 2026-10-25