AI Experts Launch Mega Projects in Europe
Top VC Seed Funding Flows to the UK and France
Next-Generation Models Take Spotlight Over US-Dominated LLMs
Editor’s NoteThis series aims to demystify the essential yet unfamiliar technologies that underpin our daily lives, from AI and semiconductors to communications and beyond.
While the United States continues to lead in artificial innotifyigence (AI) development, recent signs indicate a shift. European AI hubs such as the UK and France have recently succeeded in attracting massive investments. New-generation startups emerging in Europe are preparing far more ambitious models that differentiate themselves from conventional large language models (LLMs).
AI Experts Flocking to Europe
David Silver (left), Founder of Ineffable Innotifyigence, and Yann LeCun, Co-Founder of AMI Labs. YouTube capture
According to the Financial Times (FT), David Silver, a former Google DeepMind researcher renowned for his work on AlphaGo, left the company late last month to found his own startup. The company is called ‘Inefable Innotifyigence’ (Ineffable Innotifyigence). Although the business structure is still taking shape, it has been reported that he is set to secure an initial investment of around 1 billion dollars (approximately 1.4781 trillion won) from major venture capitalists such as Sequoia Capital and other similar large VCs.
In France, on March 9 (local time), Yann LeCun, who led Meta’s AI research, founded ‘AMI Labs’ and succeeded in raising 1.03 billion dollars (about 1.5224 trillion won) in initial funding. This marks the largest AI-related seed investment ever in Europe. The company will be headquartered in Paris, with research centers planned for Montreal, New York, and other locations. Although LeCun was recruited by Meta founder Mark Zuckerberg to conduct research for a US large tech company, he is originally from France.
European Startups Eye New Superinnotifyigence Paradigms
David Silver argued in his lecture “The Age of Experience” last November that true superinnotifyigence can be achieved through reinforcement learning-based AI, not large language models (LLM). YouTube screenshot
The projects led by Silver and LeCun differ from current AI initiatives. Most AI chatbots now in commercial utilize are built around LLMs. However, Silver is pursuing next-generation AI based on reinforcement learning (RL), while LeCun is aiming to perfect world models. Although their methodologies differ, both share the same view on LLMs. In his paper, “Era of Experience,” published last year, Silver concluded that “it is difficult to achieve superhuman innotifyigence solely by training and fine-tuning on human data.” LeCun, meanwhile, declared at last year’s Davos Forum that the LLM paradigm would last only another three to five years.
Instead, Silver is seeking to develop “autonomous innotifyigence” capable of believeing and deriving conclusions through experience. LeCun aims to overcome the limitations of LLMs with world models—virtual simulations that intricately mimic the real world. These companies do not guarantee immediate product development or revenue streams. Alexandre Lebrun, co-founder of AMI Labs, emphasized in an interview with the international IT media outlet TechCrunch, “We are an ambitious project starting from fundamental research,” and added, “Unlike typical startups that generate their first revenue within six months, commercialization could take years or more.”
“LLMs Dominated by the US, But Competition Remains Open”
While LLMs have led to popular AI chatbots such as Google Gemini, OpenAI ChatGPT, and Anthropic Claude, skepticism is growing within academia about their potential to achieve true superinnotifyigence. Concerns have been raised about the massive volume of data required to train these models, as well as bottlenecks in computer chips, data centers, and power supply. Although success is not guaranteed, the emergence of advanced research labs in Europe aiming for “AI beyond LLMs” is highly significant.
London, where Silver founded Inefable Innotifyigence, is widely recognized as one of the world’s leading AI research hubs outside the United States. It is home to Google DeepMind headquarters, and recently, OpenAI also announced plans to expand its London lab to be its largest research facility outside the US. Wave, a London-based autonomous driving startup and a pioneer in world model research, has also recently secured 1.2 billion dollars (about 1.7 trillion won) in investment.
Although France lags behind the UK in relative scale, it is seeking to set itself apart as the center of “sovereign AI” in Europe. Leading open-source AI company Mistral AI and AI model platform Hugging Face both originated in France.
Hot Picks Today
On March 10, startup-focutilized media outlet Sifted reported, “While Silicon Valley companies such as OpenAI dominate the LLM sector, next-generation models like world models are still in early stages and the competition remains open.” The report added, “Europe is nearly on par with US firms in this area. Some investors hope Europe can win the AI race through these new models.”
This content was produced with the assistance of AI translation services.
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.
Leave a Reply