Zhang Haijun / 张海军
Zhang Haijun / 张海军 is the guest in 华为的「韬定律」，是创新还是噱头？｜ Bonus, where he is introduced as a former senior engineer at Synopsys / 新思科技 and the author behind the public account “福利业的猫.” In the episode, he explains Huawei’s Tau Law from a semiconductor-engineering and EDA perspective.
His key contribution is the distinction between ordinary die-to-die Semiconductor 3D Stacking and the more ambitious Cell-to-Cell Logic Stacking implied by Huawei’s logic-folding language. Zhang treats Tau Law as meaningful but not proven: the metric shift toward system time is reasonable, while the cell-level implementation still needs new tools, packaging capability, and product evidence.
国产 AI 算力能凭「超节点」弯道超车吗？｜WAIC 深度观察 S10E23 adds Zhang’s WAIC supernode analysis, with the raw source spelling his Chinese name as 张海君. In this episode, he explains why domestic AI hardware is moving from single-chip comparison toward supernodes, Scale Up AI Interconnect, protocol fragmentation, power/cooling tradeoffs, and large-customer order validation.
Connections
- What’s Next｜科技早知道 — show where Zhang appears as guest.
- Synopsys / 新思科技 and Electronic Design Automation — toolchain background used to evaluate the claim.
- Huawei, Tau Law, and Cell-to-Cell Logic Stacking — main source topic.
- Semiconductor 3D Stacking and Advanced Packaging — broader technical context.
- AI Accelerator Supernode, Huawei CM384, and Nvidia GB200 NVL72 — WAIC supernode comparison added by S10E23.