rag failure modes

3 articles · 15 co-occurring · 0 contradictions · 0 briefs

Articulates specific failures of naive RAG (contradictions, noise, outdated facts) and proposes architectural response.

Related concepts

token efficiency 1 system prompt as production code 1 state management 1 retrieval augmented generation 1 reinforcement learning 1 prompt engineering commoditization 1 multi turn conversation management 1 multi agent state management 1 memory persistence across sessions 1 memory persistence 1 langchain context strategies 1 evaluation as context proxy 1 context window management 1 context window exhaustion 1 context window as ram 1

Evidence chain (3 articles, showing 3)

Prompt Engineering Best Practices 2026 | Thomas Wiegold Blog supports

Author identifies 'retrieved the wrong documents' as primary failure mode, supporting thesis that RAG quality is context engineering, not prompt engineering

@victorialslocum: Dumping entire conversations into a vector database isn't memory. extends

Articulates specific failures of naive RAG (contradictions, noise, outdated facts) and proposes architectural response.

State of Agent Engineering - LangChain supports

If agents are hitting consistency/hallucination problems at scale, retrieval-augmented context (RAG) quality becomes critical. The write-in data suggests retrieval/context freshness issues.

query this concept

$ db.articles("rag-failure-modes")

$ db.cooccurrence("rag-failure-modes")

$ db.contradictions("rag-failure-modes")