Nvidia and AI: Jensen Huang Defends the Crucial Role of Software
Le brief IA que les pros lisent chaque soir
Les 7 actus IA du jour, décryptées en 5 min. Gratuit.
Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.
Choisis ton rythme
Gratuit · Pas de spam · Désabonnement en 1 clic
Jensen Huang: AI and Software, a Necessary Coexistence
In a recent appearance on Lex Fridman's podcast, Jensen Huang, the CEO of Nvidia, shared his vision for the future of software in light of the rise of artificial intelligence. Contrary to some alarmist opinions, Huang believes that AI will not replace existing software but will use it to enhance its own capabilities. He illustrates this idea with an analogy: even a sophisticated humanoid robot would use a traditional microwave rather than attempt to create microwaves on its own. By connecting to the Internet, such a robot could read the user manual and become immediately proficient. This perspective underscores the importance of software as essential tools for AI agents.
OpenClaw: A Turning Point for Agentic AI
Huang compares the potential impact of OpenClaw to that of ChatGPT, emphasizing its significance for agentic systems. According to him, OpenClaw could transform the AI landscape in a manner similar to how ChatGPT revolutionized generative systems. He describes OpenClaw as "the iPhone of tokens," referring to its accessibility and potential to go viral among consumers. This popularity is attributed to the ease of access for users, making it a rapidly growing application.
From an economic standpoint, Huang anticipates a shift towards the commoditization of tokens, with prices ranging from free to premium. He predicts that paying $1,000 per million tokens will soon become a reality, highlighting a transformation of data centers into true token production factories, where revenues are directly tied to production.
A New Rack Architecture for a New Era
Nvidia has taken concrete steps to adapt to this new era of AI. The company has rethought its rack architecture, moving from systems optimized for language model inference to the new Vera Rubin platform, which consists of five types of specialized racks. These racks include dedicated units for sandboxing agents, storage systems for massive contextual caching, and solutions for ultra-low latency inference. Huang emphasizes that this new architecture is radically different from the old one, designed specifically to execute AI agents rather than just simple language models.
This strategic realignment by Nvidia demonstrates the growing importance of AI agents in today's technological landscape. By adapting its infrastructure, Nvidia positions itself to meet the future needs of the industry while asserting that software will remain at the heart of this evolution.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.