Anthropic tells AI builders to delete most subagents: one agent loop wins on cost and speed
suraj_sharma14 · x · 2026-11-17
Anthropic's Applied AI team found that a single agent loop with skills and tools often beats complex multi-agent setups on quality, cost, and speed — one company cut runtime roughly in half by collapsing its multi-agent harness. Multi-agent fails because teams copy org charts into code, subagents play a lossy 'telephone game' with summaries, and system prompts become conflicting Frankenprompts. What works: skills for playbooks loaded on demand, tools/MCP for live data and permissions, subagents only for parallel work, isolated context, or real context bloat.
Related event: Anthropic Advises Developers to Fold Subagents into Single-Agent Loops(4 posts)→
More from coding & agent
- Karpo Decision 31B open weights: a decision model scoring all options in one forward pass — _jaydeepkarale · 2026-11-17
- Riley Brown: After 50+ DMs, Nothing Beats Claude Code Projects Right Now — angrypenguinPNG · 2026-11-17
- MCP server turns your saved reels and posts into searchable notes for Claude — PaRtHuS · 2026-11-17
- Stardock CEO shows agents building games across PCs and Macs, with hard-won engineering lessons — draginol · 2026-11-17
- sandbox-mcp: Hanselman's MCP server gives AI agents disposable real Windows 11 machines — unixterminal · 2026-11-17
- ProceduralTerrains: open-source Three.js GPU terrain generator hits 2.0 — anselm · 2026-11-17