Barry Zhang

Anthropic Says Simple Workflows Are Safer Than Agent Complexity

Anthropic finds stable enterprise AI usually combines retrieval, tools, and memory—not complex frameworks or fully autonomous agents.