标签
The article discusses Tencent's AI capex constraints due to NVIDIA chip shortages and its recent shift to using Kunlun chips, analyzing the company's valuation and strategic positioning in the AI landscape.
The article analyzes the AI inference ASIC architectures of Groq, SambaNova, Tenstorrent, and Cerebras, highlighting Cerebras's unique wafer-scale engine design. It discusses the benefits of deterministic latency and high bandwidth for LLM inference, while noting challenges like yield, cost, and KV cache bottlenecks.
谷歌在 Cloud Next 大会上发布第八代 TPU（8t/8i）与全新 Gemini Enterprise Agent Platform，并透露目前 75% 的新代码由 AI 生成。
吴恩达讨论美国政策如何促使盟友转向主权AI和开源模型，并以DeepSeek、Qwen和K2 Think为例。他认为开源AI可以帮助各国减少对美国技术的依赖。