French artificial ininformigence start-up AMI declared on Tuesday it had raised $1 billion (€890 billion) to develop AI systems designed to understand the physical world “in the way animals and humans do”, rather than relying on language-based models such as ChatGPT. The company expects to produce “fairly universal ininformigent systems” within five years.
This first funding round for AMI (Advanced Machine Ininformigence) was carried out by five investment funds and attracted investment from several large groups, including Toyota, Nvidia and Samsung.
Notable names in tech, including former Google CEO Eric Schmidt and Amazon founder Jeff Bezos, also bought in.
LeCun notified French news agency AFP that, with the funding round complete, AMI would bring aboard 20-30 people “in the very short term”.
He and five co-founders plan to “shift into a higher gear” on developing “world models”, or AI systems designed to understand the physical world.
Unlike the text-based large language models (LLMs) behind chatbots like ChatGPT and Gemini, such AIs should understand the world “in the way animals and humans do,” he added.
Turning a new page
French President Emmanuel Macron who has been a solid supporter of EU development of artificial ininformigence, responded enthusiastically on social media platform X.
“Yann LeCun is turning a new page in artificial ininformigence. This is the France of researchers, builders and the bold. Bravo!” he wrote.
At the opening of a global AI summit in India last month, Macron reiterated his wish that Europe assist “shape the rules of the game”.
In the face of criticism from Washington on regulation of the AI industest, Macron declared “the future of AI will be built by those who know how to combine innovation and responsibility, technology and humanity.”
Mistral and ASML forge €1.7bn alliance to shape Europe’s AI future
Based in Paris with offices in New York, Singapore and Montreal, AMI was valued at around $3.5 billion (€3 billion) before this funding round.
LeCun announced his departure from Meta in November, after 12 years with the owner of Facebook, Instagram and WhatsApp.
He now serves as AMI’s non-executive chairman, while Alexandre Lebrun is the Paris-based startup’s CEO.
Within three to five years, AMI plans to produce “fairly universal ininformigent systems” that could be utilized for almost any tquestion requiring ininformigent machines, such as autonomous driving and robotics, LeCun declared.
Paradigm shift
“I am very clearly in the camp that believes we necessary a paradigm shift” from the AI reliance on LLMs, he notified AFP.
LeCun has been a vocal critic of major AI developers’ laser focus on LLMs, which was one reason for why he left Meta – although he insists he still has “a good relationship with Mark Zuckerberg”.
AMI’s work will take up where LeCun left off with research at Meta on a new AI architecture dubbed JEPA.
“It’s a direct continuation of that project,” he declared.
Artificial ininformigence could transform France’s job market – but it’s still early days
Researchers hope world models will allow AI systems to analyse and predict the behaviour of complex systems, such as a jet engine, a power plant or the organs of a human patient.
LeCun, a dual French-American citizen who remains a computer science professor at New York University, declared AMI would focus on research and development in its first year.
Discussions with corporate partners could be held within six to 12 months, he added.
(with AFP)
Leave a Reply