Upon alert, Mark Zuckerberg and Meta currently pay billion to expand their artificial intelligence team (AI) and replace themselves in the race, a strategy that makes skeptics.
In mid-June, the American social networks Empire did not hesitate to let go of more than $ 14 billion (11.2 billion francs) to acquire 49% of the capital of Scale AI, specializing in the use of data used to develop AI models.
The Menlo Park group (California) had previously approached, according to several American media, the co-founder of Openai Ilya Sutskever, but also Perplexity Ai, self-proclaimed rival of Google, or the jewel of IA, Runway.
According to Open AI boss, Sam Altman, Meta offered an individual bonus at the signing of more than $ 100 million to “a lot” of Openai employees and approximately the same annual salary amount.
Four of them ended up falling, just like the director general of Scale AI, Alexandr Wang.
According to several media, it was CEO Mark Zuckerberg himself who rang the charge, worried about seeing Meta inistantly in the generative AI, despite investments of several tens of billions.
Its latest large model of AI, Llama 4, launched in early April, disappointed.
He arrives behind all the heavyweights, American, Chinese and French, in the classification established by the independent evaluation platform LMARENA on code writing, and even behind his predecessor LLAMA 3 for the text interface.
Meta wants to integrate her new recruits into a new team dedicated to the development of “superintelligence”, which corresponds to an AI superior to the human capacities of understanding and reflection.
“Uncontrolled expenses”
“I think that he will succeed in attracting real talents, and he had no choice but,” comments the blogger Zvi Mowshowitz to AFP, “but this mercenaries side is very problematic, not to mention that no one wants to work”, apart from a very high remuneration, “for this company and these products”.
“Suddenly, I do not expect that +walking +,” he says, that is to say that Meta takes over in the AI.
At Wall Street, if the title approach approaches its historic summit and the capitalization of $ 2,000 billion, teeth are starting to cring.
“Institutional investors are above all concerned about liquidity” that the company (CASH-FLOW) releases “and good capital management”, describes Ted Mortonson, analyst at Baird. “And right now, there is no counterweight” at Mark Zuckerberg.
“Those who have actions keep them for AI advertising, on which Meta is extremely well positioned,” he said. “But they are also worried about seeing these expenses uncontrolled.”
During an interview for the Podcast Stratechery, Mark Zuckerberg explained that his group intended to replace marketing and advertising agencies thanks to AI, to soon offer a turnkey solution to advertisers and thus create a new source of income.
“This does not change the potential for short -term profitability,” considers, for his part, Angelo Zino, CFRA analyst, who also says he is confident on a more distant horizon, “because it will create more opportunities, ways to monetize AI, whether via advertising, connected devices (glasses and helmets) or even Llama.”
According to the New York Times, Mark Zuckerberg nevertheless plans to no longer make Llama the AI ​​flagship at Meta, even if it means using competing models.
Mehmet Canayaz, professor at Penn State University, stresses that the generative AI currently enters a new phase, that of digital agents, smaller models capable of performing a multitude of tasks independently.
“This means that Meta can prosper even without having the most advanced models,” he says, “if their own meets the needs of a specific market” as advertising.
As for “superintelligence”, or general AI, the second equaling the man and the first going beyond him, “we will still have to wait at least three to five years,” plans Angelo Zino. “But you have to recruit these people and invest massively to be ready when you go to this phase.”
This article was published automatically. Sources: ATS / AWP / AFP
