Recently, Meta celebrated the 10th anniversary of its Fundamental Artificial Intelligence Research (FAIR) team. The event highlighted Yann LeCun’s views on the current state and future potential of artificial intelligence. LeCun’s remarks, a leading scholar in the field of artificial intelligence, particularly focused on the ever-changing AI technology environment and the path to human-level AI.
Nvidia’s superiority in the artificial intelligence hardware market was an important aspect of LeCun’s criticism. As a result of its cutting-edge graphics processing units (GPUs), which are essential for training huge language models like ChatGPT, Nvidia has emerged as the world’s most valuable chipmaker. Metaphorically, LeCun described the current situation as an ‘AI war’, while Nvidia described it as ‘weapons supply’. It’s important to note that this statement reflects Nvidia’s significant position in the artificial intelligence business, as well as the importance of its hardware to advancing AI research and development.
LeCun expressed doubts about the imminent advent of artificial intelligence equivalent to humans. Contrary to the positive views of certain industry executives, he emphasized the need to reach “dog” and “cat” levels of artificial intelligence as an intermediate step on the road to human-level intelligence. Based on this cautious approach, it appears that major milestones in the advancement of artificial intelligence have yet to be reached.
LeCun also expressed concerns about the current technological capabilities of quantum computing. Even though competitors such as Google and Microsoft have put a lot of effort into this technology, he still believes that classical computing is more effective than quantum solutions for solving many challenges. With this perspective in mind, Meta has made choices that differentiate it from other digital titans by directing its resources in a different direction.
LeCun also described Meta’s methodology for creating artificial intelligence, namely its investigation into multimodal AI systems. The development of these systems, which can analyze various types of data, including audio, photos, and video, has led to advancements such as the augmented reality glasses developed by Project Aria. Despite the company’s heavy reliance on GPUs manufactured by Nvidia for training artificial intelligence software, LeCun predicts a shift toward processors specifically for deep learning brain acceleration. This shows that Meta is actively responding to and changing the artificial intelligence hardware environment.
Image source: Shutterstock