Anthrope, a American heavyweight artificial intelligence (AI) and rival of the OpenAi giant, intends to participate in the dynamism of European “technology” strengthening its presence in the continent, its product director, its product director, told AFP.
The company that causes the AI Claude model wants to be “the engine behind some of the new largest companies in tomorrow”, which will use its tools. “Many of them can and should come from Europe,” Krieger said at the Vivatech fair, the largest event dedicated to technology in the former continent held this week in Paris.
Therefore, it establishes the recruitment of one hundred people in Europe, which will strengthen their research and development capabilities in their offices in Dublin and London.
“French and European intellectual capital”
If technological bosses and political leaders have often regretted the inability of the continent to take advantage of their researchers to create important companies, particularly due to the frequent exits of young businessmen throughout the Atlantic, Mike Krieger wants to be more optimistic, as well as the CEO of the US giant of the electronic fleas of Nvidia.
Jensen Huang, who announced this week an association with the French champion of the Mistral AI and an investment of several billions of dollars in Europe, insisted on the French and European “intellectual capital”.
If the French company, founded in 2023, is younger and smaller than its US rivals, Mistral was “certainly part” of the important actors, said Mr. Krieger. The tricolor startup recently followed the traces of its competitors by publishing a “reasoning” model capable of addressing more complex tasks.
“I am talking about all the time with customers that can use Claude for certain tasks of long -term agents, but they have also optimized Mistral for one of their data processing tasks, and I think (models) can coexist in this way,” said Mike Krieger.
“Agent” models
The so -called “agent” models function as independent or semi -autonomous agents capable of working on longer horizons with less human supervision, including interaction with tools such as web browsers and emails. But the capabilities that the latest news of these agents show the fears in some researchers.
Yoshua Bengio, a professor at the University of Montreal and considered one of the founding parents of the AI, considered that an independent AI could soon represent a risk of humanity.
Anthrope still works in measures aimed at limiting any negative impact of their AI models, such as the implementation of security levels, he added.
After activating level 3 to supervise the capabilities of the version of its Claude Opus 4 model in May, the company works “actively” in the development of level 4, it detailed to the manager, without giving a launch date. According to him, future anthropic models should achieve by 2026 or 2027 in certain aspects the prediction of the CEO Dario Amodei, according to which the company could offer access to a “genius country in a data center.”
The latest anthropic models of AI are “the level of genius for certain very specific things,” he said. “Next year, they will continue to surpass in certain aspects, while requireing a lot of human intervention,” he predicted.
Source: BFM TV
