1
Trust Architecture in AI Products
AI产品成败在信任：从开发者视角拆解可解释性架构，让非确定性行为变得可理解。
When I first started exploring AI products, I focused on prompts, models, and response quality. Recently, I came across another concept that feels jus…
2
Ethical Hyper-Velocity (EHV): A Hardware-Rooted Zero-Trust Runtime Enforcement Architecture for Agentic AI Systems
一种基于硬件根信任的零信任运行时强制架构，专为高频策略更新的代理型AI系统提供安全治理。
arXiv:2605.17909v2 Announce Type: replace Abstract: As autonomous agentic systems scale across regulated critical infrastructures, the lack of mechani…