Meta’s Chief Scientist Yann Likon made a keynote at the Meta launch event in Paris, announcing that Generative AI, the technology powering ChatGPT, has come to an end. Instead, he promised the development of new models of artificial intelligence that would demonstrate human-like rationality.
LeCun expressed his dissatisfaction with the current state of AI and machine learning, asserting that humans have intelligence, machines do not. He presented Meta’s latest AI project called Image-Based Joint Embedding Predictive Architecture (JEPA). The ambitious effort aims to surpass the capabilities of creative AI, enabling machines to imagine abstract ideas rather than simply regurgitating existing online information.
According to LeCun, generative models are now a thing of the past, as Meta plans to shift its focus to joint embedding predictive architectures. He said confidently, “My prediction is that in a few years, the big generative language models will no longer be used; we’ll have something better to replace them.” LeCun, known for his expertise in AI, has been highly critical of generative AI models such as ChatGPT and image-based Dall-E since their launch last year.
LeCun believes that the fear and excitement surrounding creative AI has created an exaggerated perception of its capabilities. Meta CEO Mark Zuckerberg announced on Facebook that the JEPA tool will be open source, allowing researchers to experiment with it. The goal is to develop AI that closely aligns with human understanding of the world. “We need models that can see the world and make predictions. This research is another step in that direction,” Zuckerberg said.
Unlike its competitors, Meta has taken a more sensible approach to ChatGPT-style AI for its social media platforms, Facebook and Instagram. While Meta has incorporated generative AI into its products, it has done so with less publicity than companies like Microsoft or Google. Also, Meta has released open-source AI models that require less computing power than the technology behind ChatGPT.
LeCun’s announcement marks a significant shift in the AI landscape, with Meta leading the development of new AI models that attempt to mimic human rationality. The JEPA project represents an innovative step toward extending machine intelligence beyond the boundaries of generative AI.
