Nvidia Proclaims AGI with OpenAI's Astra

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 praises Astra, the new OpenAI model trained on Nvidia chips, and claims that AGI is here. While OpenAI plans a deployment this week and boasts professional capabilities, voices like Gary Marcus challenge the AGI qualification, and Sam Altman considers the term poorly defined. Meanwhile, OpenAI emphasizes its heavy reliance on Nvidia GPUs, which reports record revenues and mentions "400K GPUs coming soon."
OpenAI Relies on Nvidia and Nvidia Reports $96.2 Billion in August
OpenAI states that its training fleet and the majority of its inference stack operate on Nvidia GPUs, having described Nvidia as the foundation of its infrastructure back in March. This dependency is part of a broader trend where OpenAI, Meta, Anthropic, and Google increasingly turn to Nvidia's advanced chips. Nvidia reported a quarterly revenue of $96.2 billion in August, more than double that of the same period last year. Revenue from its data center segment, which includes its AI chips, reached $89 billion. In a message posted on Sunday, Jensen Huang also announced "400K GPUs coming soon."
Defining AGI is Divisive: Marcus's Critique and Altman's Reservations
Gary Marcus believes that Jensen Huang has jumped to conclusions without providing evidence or a definition, arguing that announcing victory without defining it muddles the debate. He proposes a ten-point definition of AGI and asserts that Astra meets only one or two of those criteria, adding that the model falls short of conventional definitions. Sam Altman, for his part, describes AGI as a very poorly defined term, akin to a marketing term lacking relevance. OpenAI defines AGI as highly autonomous systems that surpass humans in most economically valuable tasks.
Huang Credits Astra, OpenAI Launches and Promises Performance
Jensen Huang claims that AGI has arrived and credits Astra for this, specifying that the model was trained on Nvidia chips. He posted on X, "From ChatGPT to o1 to Astra in 4 years," and congratulated the OpenAI team. OpenAI unveiled Astra on Thursday, calling this model a significant breakthrough in research and an evolution in what AI can handle. The company asserts that this model is the most advanced and aligned to date, capable of performing the most complex professional tasks with unprecedented speed, accuracy, and discernment, and plans to make it available to customers within the week. Greg Brockman stated, "Welcome to the era of AGI," and believes that AGI could be considered to have been created around this time, possibly regarding this model.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.