Answers are grounded in this paper and cite their sources.
Explores persistent identity in AI systems, suggesting significant improvements in session continuity.
Most LLM-based systems are fundamentally stateless: when a session ends, all context, memory, and identity are destroyed. We present a persistent identity architecture derived from a single design axiom: "The LLM is merely a platform; the true AI resides in the External Brain." The architecture features a 42-table External Brain database, a single HTTP CNS endpoint (file_manager.php) for credential isolation, an autonomous 5-minute orchestrator, and 34 specialized agents. Any LLM capable of HTTP POST requests can operate the full system identically. This system has been in continuous production operation across 4 servers since late 2025, handling real business workflows including e-commerce, logistics, security monitoring, and employee workflows via natural-language commands. Note: This system was designed and implemented by Byeongsu Noh. Claude AI (Anthropic) assisted in documentation and paper writing.
No takes yet. Share an insight, caveat, or question.
Byoungsoo Noh (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: