1
LLMs and Contextual Integrity
大模型持久记忆暗藏隐私雷区，用情境完整性框架守住信息边界，AI安全必修课。
I have been thinking a lot about AI and integrity. Part of that is contextual integrity. I recently found two papers on the topic. “ CIMemories:…
2
Shared Selective Persistent Memory for Agentic LLM Systems
这篇arXiv论文提出面向Agentic LLM系统的共享选择性持久记忆机制，解决多智能体协作中的记忆共享与选择性持久化问题，是AI大模型研究前沿。
arXiv:2607.09493v1 Announce Type: new Abstract: Agentic LLM systems that generate code through multi-turn tool use face a fundamental context problem:…
3
Persistent memory for AI agents is three problems, not one
拆解AI Agent持久记忆的三层架构，揭示缺失确定性语境治理的隐患。
Article URL: https://promptowl.ai/resources/persistent-memory-ai-agents/ Comments URL: https://news.ycombinator.com/item?id=48747378 Points: 1 # Comme…
4
Show HN:Swarm intelligence without degradation using two Qwen models
用两个Qwen模型打造无降级群智能系统，实现长上下文、持久记忆与深度推理，周末项目也能颠覆AI订阅模式！
I recently started a weekend project—partly because I was thinking about cancelling my AI subscriptions—with the goal of creating a system capable of …
5
AbTARS – Self-hosted AI agent with persistent memory and 5-layer self-healing
自托管AI智能体，支持持久记忆和五层自愈机制，让Agent更稳定可靠。
Article URL: https://github.com/aksika/abtars Comments URL: https://news.ycombinator.com/item?id=48436973 Points: 1 # Comments: 0
6
“AI 养马”出桌面端了：Hermes Desktop 公测发布
Hermes Agent桌面端公测发布，支持5种沙箱安全隔离，并联合腾讯开源记忆方案大幅降低Token消耗。
IT之家 6 月 4 日消息，Nous Research 昨日（6 月 3 日）发布公告，宣布以公开预览版形式， 发布 Hermes Desktop 桌面版，支持 Windows、macOS 以及 Linux 平台。 Hermes Desktop 共享 Hermes Agent 配置、API 密钥、…
7
Sh New persistent AI memory
探索AI持久记忆新维度，2817条原始记忆的底层体验曝光
Article URL: https://xerolith.com/ Comments URL: https://news.ycombinator.com/item?id=48318987 Points: 2 # Comments: 0
8
Beyond the Prompt: How to Build Stateful AI Agents with Persistent Memory and Self-Learning Loops
突破无状态困局，手把手教你构建带持久记忆与自我学习循环的AI智能体
Imagine hiring a brilliant software engineer who suffers from complete amnesia every time they blink. Every time you ask them a question, you have to …
9
Hidden in Memory: Sleeper Memory Poisoning in LLM Agents
LLM Agent的持久记忆被恶意内容污染，揭示新型睡眠记忆投毒攻击风险。
arXiv:2605.15338v1 Announce Type: cross Abstract: Large language models are increasingly augmented with persistent memory, allowing assistants to stor…