Most production agents fail from missing systems engineering, not bad models
bgoncalves · x · 2026-10-14
A thread argues most AI agents fail in production not because the model is bad, but because nobody engineered the system around it — enumerating what actually breaks and how to fix it.
A resharing data scientist agrees "the model is the easy part" and is running an Oct 31 workshop covering the full engineering layer: permissions, memory, and supervision.
More from coding & agent
- Nous Research ships Hermes Agent v0.22.0, the 'Threshold' release — Polymarket · 2026-10-14
- ResearchTwin MCP server lets agents discover researchers, papers and datasets — modelcontextprotocol · 2026-10-14
- Hermes Agent v0.22.0 ships: API-key-free onboarding, 600 plugins, one-click local models — NousResearch · 2026-10-14
- Nous Research Ships Hermes Agent v0.22.0, the 'Threshold Release' — NousResearch · 2026-10-14
- RustGIS vs QGIS: an 18-30 MB Rust app with browser WASM build and built-in AI assistant — giswqs · 2026-10-14
- 0sql fixes text-to-SQL accuracy by never letting the model write SQL — nothinginausername · 2026-10-14