Experience Layer for AI

CortexDB V1는 AI 에이전트의 경험 레이어를 구축하는 데이터베이스로, 기존 LLM과 지식 레이어를 넘어 사용자가 겪은 사건과 신념을 체계적으로 관리한다. 다섯 개의 메모리 계층(이벤트, 에피소드, 사실, 신념, 이해)을 통해 AI가 왜 그렇게 판단했는지 근거를 추적할 수 있으며, LongMemEval-S와 LoCoMo 벤치마크에서 최첨단 성능을 기록했다. 8주 만에 개발된 이 시스템은 PASETO 인증, 계층적 네임스페이스, GDPR 준수 삭제 기능 등 실무 적용에 필요한 기능도 포함한다. 향후 V2에서는 경험을 종합하는 통합 단계가 추가되어 더욱 지능적인 메모리 기능을 제공할 예정이다.

https://cortexdb.ai/blog/v1

#aiaagents #memorymanagement #cortexdb #llm #distributedinfrastructure

CortexDB V1: The Experience Layer for AI Agents

93.8% on LongMemEval-S. 86.9% on LoCoMo. Five memory layers. Hierarchical scopes. PASETO auth. The third layer every AI agent has been missing — shipped.

"From Censys’s vantage point, this rapid adoption is reflected by its exposure footprint on the public Internet. Although OpenClaw is designed to run locally on TCP/18789 or be accessed through protective mechanisms like SSH or Cloudflare Tunnel, Censys has identified more than 21,000 publicly exposed instances as of 31 January 2026. Observed deployments span major hosting providers and regions, with the largest concentration in the United States, followed by China and Singapore. As these assistants increasingly operate with access to highly sensitive user data, the scale and speed of their Internet-facing deployment underscores the importance of careful configuration, monitoring, and security review early in their lifecycle."

https://censys.com/blog/openclaw-in-the-wild-mapping-the-public-exposure-of-a-viral-ai-assistant

#OpenClaw #AI #GenerativeAI #AIAAgents #AgenticAI #Moltbot #Clawdbot

OpenClaw’s Rapid Rise: 21,000+ AI Instances Exposed

Censys tracked OpenClaw’s explosive growth, finding 21,000+ exposed AI assistant deployments online and highlighting urgent security risks.

Censys