Pat Gelsinger, former CEO of Intel, has reacted to the meteoric ascent of Deepseek AI, highlighting three key lessons of the computer history that believes that the community of AI must adopt in light of the disruptive success of the Chinese company.
Sharing his ideas about X, Mr. Gelinsger said the computer science obeyed the gas law, which means that when computer science becomes dramatically cheaper, it does not shrink the market, it expands it. He criticized the market reactions that the actions of AI saw, arguing that reducing the costs of only “will make the AI display much more widely.”
Engineering thrives with limitations: with limited access to avant-garde chips due to export restrictions, the Deepseek Chinese engineers had to innovate with less resources and deliver a world class model at a lower 10-50x cost than The competitors.
Open Vences: Mr. Gelsinger expressed concern for the increasingly closed ecosystem of AI, supporting Deepseek’s decision to obtain his model. He requested more transparency in the development of AI, aligning with the defense of Elon Musk of Grok for the open AI instead of the closed approach of the Operai CEO Sam Altman.
“Thanks to the Deepseek team.” He concluded his publication.
Wisdom is learning the lessons we thought we already knew. Depseek reminds us of three important learning history:
1) Computing obeys the gas law. Doing it dramatically cheaper will expand the market for it. The markets are wrong, this will do Ai …– PAT GELSINGER (@Pegelsinger) January 27, 2025
In response to a commentator about X, Mr. Gelsinger, who retired from Intel last year after receiving the option to resign or be retired from the Board, said that if the company would still lead, he would be open to work with Deepseek.
Yeah
– PAT GELSINGER (@Pegelsinger) January 27, 2025
The Deepseek innovative approach is an important industry disruptor. Instead of depending on expensive and high -end hardware such as its competitors, Depseek trained R1 using 2,000 GPU H800 Nvidia in just two months at a cost of approximately $ 5.5 million, a fraction of billions of US companies that typically spend In AI infrastructure.
The immediate response of the market has been dramatic, since Nvidia’s shares received a hard blow, losing almost $ 600 billion, and the Deepseek application increased in popularity.
But while the market panicked, Pat Gelsinger doubled. In A detailed publication of LinkedInHe revealed that he was still buying Nvidia shares, convinced that the cheapest would expand the potential of the industry, he does not shrink it. “Today I am a buyer of Nvidia and AI shares and happy to benefit from the lowest prices,” he wrote.
In his publication, he also reflected on a conversation with the legendary computer scientist Donald Knuth, who once said that “did his best job when the resources were more severely limited and the most demanding programs.” Gelsinger connected this philosophy with Depseek’s achievements, praising his highly limited Chinese team for delivering an innovative AI model.
In addition, making comparisons with Linux, USB and Wi-Fi, said that the story has repeatedly shown “open victories every time it is given an adequate take.”
Finally, Pat Gelsinger called Depseek “an incredible engineering piece” that will mark the beginning of a greater adoption of AI and restore the industry approach for open innovation.
By
