降低AI代理调用延迟，英特尔研究CPU与存储堆叠架构

Odaily News: Chuan Mu posted on X platform, stating that in a podcast interview on "Tech Surge," Lip-Bu Tan revealed that Intel is researching a new memory architecture that stacks CPU and memory together, packaging the CPU and memory in one unit to reduce invocation latency during AI agent operations and improve AI agent efficiency.