Subagents make coding slower and dumber: dev's first-hand vibe check
_AustinCalvert_ · x · 2026-09-15
A dev's hands-on observation: solving similar problems with subagents is noticeably slower than without, subagents perform worse (likely from having to pass summaries of the problem around), and it burns more tokens.
More from coding & agent
- ZGCM-1: A fully open 7B foundation model for math and agentic search — zgcagi · 2026-09-15
- Google: Structured Intermediate Specs Unlock Diverse UI Exploration for Vibe Design Agents — google · 2026-09-15
- Sprint planning is a colossal waste of time now that AI can draft the work — arpit_bhayani · 2026-09-15
- SailPoint president: 'We standardized on Copilot and nobody is using it' amid shadow AI — SuB8u · 2026-09-15
- Solo dev argues LLMs shouldn't be the default center of an agent runtime — HmmmThisIsOdd · 2026-09-15
- How Much Code Is Throwaway Now? Engineers Rethink Project Lifespans in the AI Era — rakyll · 2026-09-15