At the VivaTech conference in Paris on May 22, 2024, Yann LeCun, the former chief AI scientist at Meta Platforms, expressed concerns regarding the company's current leadership and potential staff departures. LeCun, a prominent figure in the AI field, characterized Alexander Wang, Meta's chief AI officer since 2025, as "young" and "inexperienced." Despite acknowledging Wang's ability to learn quickly, LeCun criticized his lack of research experience, suggesting this could affect Meta’s innovation capabilities.
LeCun’s remarks came in the context of a broader talent competition in the AI sector, where Meta has invested heavily to attract top talent, reportedly offering substantial signing bonuses to entice professionals from competitors like OpenAI. He noted that Meta’s CEO, Mark Zuckerberg, appears to have lost confidence in the AI team following controversies related to the company's Llama 4 model, which were perceived as attempts to manipulate performance benchmarks. This loss of confidence has reportedly led to significant turnover, with LeCun predicting further exits.
Additionally, LeCun commented on the limitations of large language models (LLMs), describing them as a "dead end" for achieving superintelligence. He noted that while LLMs have their applications, they also face challenges such as hallucinations and difficulties with multimodal data processing. In contrast, his new venture, Advanced Machine Intelligence Labs, is focusing on developing "world models," which aim to enhance AI's learning capabilities.