A recent experiment using artificial intelligence (AI) to create a novel dish has unexpectedly gone viral, sparking widespread online discussion. Popular blogger "Los Angeles Yingzheng W" shared a video on Douyin (TikTok) showcasing the process of using the DeepSeek model to generate and cook an "unprecedented AI dish." The video quickly climbed the Douyin trending charts, attracting significant attention and debate. Despite the blogger's best efforts to follow the AI model's somewhat "abstract" instructions, the final result appears to have fallen short of expectations.

Today, OPPO announced that it has successfully applied the Expert Mixture of Experts (MoE) architecture to mobile devices for the first time globally. This groundbreaking advancement significantly enhances AI processing efficiency, opening new pathways for the further development of terminal AI technologies and laying the foundation for the deep integration of AI and mobile hardware. As AI technology rapidly advances, an increasing number of AI tasks need to be completed on terminal devices. However, large AI models often require powerful computing capabilities, which poses a huge challenge for mobile devices with limited hardware resources. This breakthrough aims to overcome this limitation.

Recently, edge computing chip and software startup SiMa.ai announced the launch of its new product—MLSoC Modalix chip, further solidifying its position in the edge AI market. SiMa.ai has just completed a $70 million funding round, with investors including industry giants like Dell Technologies Capital.

Cloud TPU v5p is Google's most powerful tensor processing unit, featuring 8960 chips per pod with significant performance improvements. The AI Hypercomputer uses a collaborative system design to optimize the efficiency and productivity of AI tasks, revolutionizing supercomputer architecture. It introduces the Dynamic Workload Scheduler and traditional consumption models, supporting Cloud TPU and Nvidia GPUs, optimizing user expenditures. This innovation marks a milestone in artificial intelligence.

IBM's latest NorthPole brain-like chip runs AI image recognition algorithms at an astonishing speed, 22 times faster than commercial chips, with a 25-fold increase in energy efficiency. This chip employs brain-inspired technology to overcome the von Neumann bottleneck, accelerating AI image recognition. NorthPole integrates computation and storage, revolutionizing computer architecture by placing memory information on the chip, eliminating bottlenecks.

Xiaomi has released the new bionic quadruped robot CyberDog 2, equipped with a 21 TOPS AI processing power. CyberDog 2 has undergone comprehensive technical upgrades, providing enhanced dynamic responsiveness and a rich variety of movements. With up to 21 TOPS of computing power, CyberDog 2 supports complex intelligent interactions and precise responses.