Alibaba's Tsai Sounds Alarm On Signs Of AI Bubble In US: 'People Are Building Data Centers On Speculation'

BenzingaMar 25 04:53 ET

After the low-key release of DeepSeek-V3, the industry was shocked and once again questioned the Silicon Valley model.

① DeepSeek released the V3-0324 model on Monday evening, with preliminary tests showing that it can run on Consumer-grade Hardware, breaking the traditional notion that large models require a Datacenter; ② DeepSeek's model continues to reduce the energy consumption and computing costs of large models while maintaining an open-source approach to continuously drive technological innovation, leading to the rapid development of the domestic AI Industry and raising questions about Silicon Valley's closed and paid model.

cls.cnMar 25 00:13 ET

Insurance capital is about to "operate" Gold! Four major insurance capital giants have been approved to become members of the Shanghai Gold Exchange.

Insurance funds broaden the channels for risk avoidance.

wallstreetcnMar 24 21:23 ET

Express News | Jack Ma-backed, Ant Group Co., utilized semiconductors manufactured in China to create methods for training AI models, resulting in a 20% reduction in costs, according to sources (Bloomberg)

Moomoo 24/7Mar 24 02:45 ET

Kai-Fu Lee: In the future, there will likely only be three major models left in China.

They are DeepSeek, Alibaba, and ByteDance. Li Kaifu stated that the competition in the large model industry in China is still very fierce, with DeepSeek currently having the strongest momentum, while ByteDance has an advantage due to its user base and monetization channels.

wallstreetcnMar 21 01:35 ET

The "gap" in the implementation of the DeepSeek large model integrated machine | Depth

① The customer demand for the DeepSeek all-in-one machine has significantly increased, more than doubling compared to the customer density brought by ChatGPT in the previous two years. ② Currently, there is a severe shortage of relevant technical talent in government and enterprises, and there is a considerable gap between large models and their Business implementation.

cls.cnMar 20 23:08 ET