People who don’t closely monitor developments in artificial intelligence ( AI ) wouldn’t have known about DeepSeek on January 1. But, by late January, it became a widely celebrated experience. The obvious advances made by this largely mysterious Chinese start-up shocked US tech companies like Meta and OpenAI.
Advertisement
The Chinese” David” revealed in a complete and widely accessible post that it had been able to create an R1 model with functionality similar to OpenAI’s o1 at a fraction of the cost that National” Goliaths” have invested in AI over the years. However, the achievements of DeepSeek should not be taken as a blow to China or as a loss to the US. Otherwise, the milestone vindicates the importance of Sino-American cooperation over AI creation and management.
The China-located walk succeeded by fusing the best of Silicon Valley-style innovation, the grit and determination of home-grown Chinese expertise and a stroke of genius rooted in its academic flexibility. The start-up has a truly vertical and non-hierarchical corporate culture, just like many other major Chinese AI and advanced technological companies.
Its employees are fueled by a shared desire to succeed, but they also have a ton of freedom at work. However, creator – an engineering-trained endeavor capitalist-turned-AI whizz – once declared, “experience is not that crucial”. DeepSeek had acquired 10, 000 A100 graphics processing units ( GPUs ) before export controls imposed by Washington.
In embracing the open-source attitude, however, the start-up eschewed an iron hand over its technology and data, as embraced by many of its American peers.
Flat Earth is ideal for the academic Earth’s growth. When users can download and modify designs anywhere in the world, everyone benefits. However, for those who criticise DeepSeek for its obvious repression, they can download the software, run it on their own GPUs or cloud servers, as they wish.