AMI Labs, co-founded by Turing Prize winner Yann LeCun after his time at Meta, has made waves by raising $1.03 billion, giving it a pre-money valuation of $3.5 billion. This innovative venture aims to create “world models,” AI systems that learn from real-world experiences rather than just text.
In a chat with TechCrunch, CEO Alexandre LeBrun shared his optimism for this field. He believes “world models” will soon become the next hot topic in AI. “In six months, every company will claim to be working on world models to attract funding,” he said with a grin, highlighting how competitive the landscape could become.
LeBrun explained the ambitious nature of AMI Labs. Unlike many startups that rush to launch products, AMI is dedicated to foundational research. This long-term focus means their commercial applications might take years to develop. The goal? To create AI that truly understands the complexities of the world around us, especially in fields like healthcare. AMI’s first partnership is with Nabla, a digital health startup aimed at creating safer AI applications.
LeBrun understands the risks tied to current AI systems, especially large language models (LLMs) that can produce unreliable results. A misstep in healthcare could lead to serious consequences. That’s why he and LeCun are focused on refining their approach with the Joint Embedding Predictive Architecture (JEPA), a concept introduced by LeCun in 2022.
Investment in this sector is booming. For instance, SpAItial, another company in this space, raised $13 million, while Fei-Fei Li’s World Labs secured $1 billion in funding just last month. With AMI Labs joining this trend, the amount raised far exceeded initial expectations, thanks in large part to its talented team.
Last December, AMI Labs aimed for €500 million but ended up gathering around €890 million. This success was fueled by a strong roster of backers, including LeCun and experienced individuals from Meta and beyond.
LeBrun noted that the excitement around this venture allowed them to pick and choose investors, ensuring alignment in values and vision. Notable firms like Cathay Innovation, Greycroft, and Bezos Expeditions led the funding round. The support also comes from influential figures including Mark Cuban and Eric Schmidt.
With this significant funding, AMI Labs can focus on crucial operational costs like computing power and recruiting top talent. Their team will span key locations: Paris, New York, Montreal, and Singapore.
Though it won’t generate revenue right away, AMI Labs prioritizes real-world engagement. “We need to test our models with real data,” said LeBrun. This approach will involve active collaborations, starting with Nabla as a partner.
In addition to corporate backers, recognition from major players like NVIDIA and Samsung strengthens their position. Investments might take time to yield products, but AMI Labs remains dedicated to transparency in research. LeBrun emphasized their commitment to publishing findings and making a significant portion of their work open-source.
The aim is to build a collaborative community around their research, speeding up progress in this rapidly evolving field. As the AI world shifts, AMI Labs is paving the way for a new understanding of the technology’s potential. For further insights on the influence of such AI developments, check out articles from reputable sources like TechCrunch or Bloomberg.
Source link
Yann LeCun,world models,AMI Labs
