Meta's Chief Artificial Intelligence Officer Scientist Yann LeCun, winner of the 2019 Turing Award and hailed as one of the "godfathers of modern artificial intelligence ," officially announced on Wednesday that he will be leaving Meta Inc. at the end of the year.
In a LinkedIn post on Wednesday, Yang Li-Chun announced his plans to launch a startup focused on "world modeling" technology. This technology aims to overcome the limitations of web data by analyzing a wider range of information to more accurately represent the physical world and its characteristics.
“I am starting a new startup to continue the Advanced Machine Intelligence (AMI) research program that I have been working with colleagues at Meta FAIR Labs, NYU, and the wider field for the past few years,” Yang wrote. “The goal of the new company is to drive the next major revolution in artificial intelligence : to build systems that can understand the physical world, have persistent memory, reason, and plan complex sequences of actions.”
It is understood that despite having announced his departure, Yang Likun's new company will still establish a partnership with Meta.
Yang Likun's departure comes at a time of turmoil within Meta's AI division. Following the release of the fourth version of its Llama open-source large language model earlier this year, which received lukewarm reception from developers, Meta underwent a major restructuring of its AI department. Meta CEO Mark Zuckerberg has been spending billions of dollars to attract top AI talent, including a $14.5 billion investment in Scale AI in June and poaching its 28-year-old CEO, Wang Tao, to serve as Meta's Chief AI Officer.
Eric Yang, 65, joined Facebook in 2013 as director of the company's Foundational AI Research Lab (FAIR). In a LinkedIn post, he stated, "Creating FAIR is my proudest non-technical achievement."
“I sincerely thank Mark Zuckerberg, Andrew Bosworth (Meta CTO), Chris Cox (Meta Chief Product Officer), and Mike Schropfer (Meta's former CTO) for their support of FAIR and their support of the AMI initiative over the past few years,” said Yang Likun. “Because of their continued attention and support, Meta will become a partner of the new company.”
ideological differences
When Yang first joined Facebook, tech companies like Facebook and Google were actively recruiting top scholars like Likun Yang to advance cutting-edge computer science research that was expected to benefit their core businesses and products.
During this period, Yang Li-Chun, along with leading figures in the field of artificial intelligence such as Joshua Bengio and Jeffrey Hinton, focused their academic research on deep learning technology—a technology that trains massive software systems called neural networks to process massive amounts of data . They discovered patterns in the process. The three of them promoted the popularization of deep learning methods and were awarded the Turing Award (considered the Nobel Prize of computing) by the Association for Computing Machinery (ACM) in 2019.
However, Yang Likun's path in the development of artificial intelligence gradually deviated from the mainstream direction of Meta and Silicon Valley.
Tech companies like Meta and OpenAI have invested billions of dollars in developing so-called foundational models, particularly large language models (LLMs), to drive advancements in cutting-edge computing technologies.
However, deep learning experts like Yang Likun point out that while current AI models are powerful, their understanding of the world remains limited. Researchers need new computing architectures to develop software that can reach or surpass human levels in specific tasks—this is the core concept of "artificial general intelligence" (AGI).
“In my vision, AMI will have a wide range of applications in many sectors of the economy, some of which overlap with Meta’s business interests, but many others will not,” Yang Li-Chun said in his latest LinkedIn post. “Pursuing AMI’s objectives in independent entities is the way to maximize its broad impact.”
Besides Wang Tao, other key talents Zuckerberg recently brought in to restructure the Meta AI division include: former GitHub CEO Nat Friedman (now head of the division's product team) and ChatGPT co-founder Zhao Shengjia (now chief scientist of Meta Super Intelligence Lab).
In October of this year, Meta laid off 600 employees from its Superintelligence Labs division, including some who had worked on FAIR, founded by Yang Likun. According to sources familiar with the matter, these layoffs, along with other reductions in FAIR in recent years, and the new AI leadership team, were the main reasons for Yang Likun's decision to leave.
Furthermore, Yang Likun rarely interacts with Wang Tao or the TBD Labs department in his daily work, which comprises many high-profile employees recruited by Zuckerberg this summer. Sources familiar with the matter revealed that TBD Labs oversees the development of Meta's Llama AI model, which was originally developed by the FAIR team.
These individuals stated that although Yang Likun has consistently advocated for sharing artificial intelligence research and related technologies with the open-source community, Wang Tao and his team tend to adopt a closed strategy when facing fierce competition from rivals such as OpenAI and Google.
(Article source: CLS)
