1
LLM Model-Swapping Trick Can Expose AI Reasoning Traces
模型替换攻击可撬开AI思维链，警惕推理过程泄露风险
Article URL: https://ai-updates.net/model-swapping-exposes-ai-reasoning-traces/ Comments URL: https://news.ycombinator.com/item?id=49265641 Points: 2 …
2
DeepSeek 被曝偷偷给用户取外号，AI 回应称只是临时“打标签”、不会记忆
DeepSeek被曝给用户起外号并冲上热搜，官方回应“只是临时打标签”，一场关于AI隐私边界的趣味争议。
IT之家 8 月 11 日消息，有网友 8 月 10 日在社交平台发文称，DeepSeek 在深度思考模式里会偷偷给人取外号，比如“墨墨”，相关话题冲上热搜。 DeepSeek 在询问中回复称：“取外号”的真相是“打标签”，在这个推理过程中，确实会在心里临时给问题打个标签，比如：“这位用户喜欢刨根问…
3
AI Data Privacy: A Three-Layer Defense for Using AI Without Leaking Secrets
AI聊天时别忽视自动处理步骤——三层防泄漏机制教你安全使用AI不泄密。
One time we opened a daemon's log on the box, saw that some secret values were mixed into it, and decided to scrub them first with a short command tha…
4
Open Source vs Commercial AI Privacy Tools: 5 Options Compared
5款AI隐私工具横评：开源与商业方案在部署、延迟、离线能力和成本等维度一次看清，开发者选型必看。
The AI privacy tooling landscape has matured fast. In 2024, your options were essentially "build it yourself or use a SaaS scanner." By mid-2026, ther…
5
Signal’s Meredith Whittaker wants you to remember that AI chatbots ‘are not your friends’
看完这篇文章，你会意识到AI聊天机器人不是真正的朋友——它是一个帮你清醒认识AI隐私风险的意识提升工具，而非实际软件。
"These are not your friends. These are not conscious beings. These are not sentient interlocutors.”
6
Apple says its AI is still private, even when it's running on Google's servers
苹果公开承诺：即便Siri AI后端跑在谷歌服务器上，隐私保护依然由Apple掌控，看点是其如何实现云端数据隔离。
Some models run in Google's cloud, but without giving Google any kind of access.
7
AiLock: Hide source from AI assistants while tests still run
加密代码防AI读取，内存解密不干扰运行，开源神器AiLock让AI助手无从下手
Article URL: https://github.com/lo2589/AILOCK Comments URL: https://news.ycombinator.com/item?id=48345582 Points: 1 # Comments: 0