Elon Musk not impressed by DeepSeek’s claims, doubts startup’s claims about {hardware} sources – Firstpost

After greater than every week of silence, Elon Musk has lastly spoken out about DeepSeek, the Chinese language AI startup that has made headlines with its low-cost, high-performance AI mannequin. Relatively than becoming a member of the refrain of reward, Musk has dunked chilly water on the corporate’s success, casting doubt on its speedy rise and suggesting that DeepSeek isn’t absolutely clear about its {hardware} sources. His scepticism stands in distinction to the constructive reactions from different tech leaders, elevating questions concerning the true nature of DeepSeek’s achievements.

Whereas many within the trade have celebrated DeepSeek’s skill to ship spectacular AI efficiency at a fraction of the price of US giants like OpenAI, Musk stays unconvinced. His feedback, notably across the firm’s {hardware} claims, have added a layer of controversy to the continued debate about the way forward for AI improvement.

Musk’s scepticism about DeepSeek’s claims

Musk’s scepticism has been clear because the begin. When Salesforce CEO Marc Benioff praised DeepSeek’s AI breakthrough, claiming it proved that AI’s true worth lies in knowledge, not costly {hardware}, Musk responded with a blunt “Lmao, no.” Moreover, when a consumer joked that DeepSeek’s AI mannequin, R1, had been “leaked from a lab in China,” Musk responded with a laughing emoji, an obvious reference to previous controversies surrounding China’s position in international occasions.

What number of AI Chips does DeepSeek actually have?

One in all Musk’s key criticisms revolves round DeepSeek’s reported GPU utilization. The corporate has claimed it used round 10,000 Nvidia A100 GPUs to coach its mannequin, a quantity that appears comparatively modest in comparison with the huge {hardware} necessities of different AI leaders. Nevertheless, Musk and Scale AI CEO Alexandr Wang consider DeepSeek may very well be utilizing way more highly effective sources. Wang urged that DeepSeek could possibly be working with round 50,000 Nvidia H100 GPUs, however attributable to US export restrictions on superior chips, the corporate could also be reluctant to reveal the total scale of its {hardware}.

Musk echoed this principle, merely responding with “Clearly,” implying that DeepSeek is hiding the true extent of its GPU utilization to keep away from scrutiny from US authorities. This raises questions concerning the firm’s transparency and whether or not it’s stretching the reality concerning the sources behind its AI mannequin.

Combined reactions to DeepSeek’s success

Musk’s dismissive stance is in sharp distinction to the reactions from different trade figures. Whereas Benioff celebrated DeepSeek’s achievement as a breakthrough second, Microsoft’s Satya Nadella urged that cheaper AI would speed up international adoption, and OpenAI’s Sam Altman admitted that DeepSeek had created a powerful mannequin for its value. Musk, nonetheless, appears unwilling to simply accept DeepSeek’s success at face worth, particularly when it challenges the established narrative of pricey, high-performance AI.

With Musk’s personal AI enterprise, xAI, now within the race with its Grok AI mannequin, it’s clear that his scepticism in the direction of DeepSeek is as a lot about aggressive rivalry as it’s about questioning the legitimacy of the claims being made by the Chinese language startup. Because the battle for AI supremacy intensifies, Musk’s doubts solely add extra gas to the hearth. Whether or not his scepticism is justified or simply one other occasion of trade competitors stays to be seen.