Google DeepMind CEO Demis Hassabis says AI scaling 'must be pushed to the maximum'

A significant discussion is unfolding in Silicon Valley regarding the potential of scaling laws in artificial intelligence. Demis Hassabis, the CEO of Google DeepMind, recently emphasized the necessity of maximizing the scaling of current AI systems. Speaking at the Axios AI+ Summit in San Francisco, he stated, "The scaling of the current systems, we must push that to the maximum, because at the minimum, it will be a key component of the final AGI system. It could be the entirety of the AGI system." Artificial General Intelligence (AGI) remains an elusive goal for many leading AI firms. This theoretical form of AI is designed to reason at a level comparable to human intelligence. Companies are heavily investing in infrastructure and talent to reach this milestone. According to Hassabis, scaling is crucial for achieving AGI, although he acknowledges that there may be a need for "one or two" additional breakthroughs in the field. Challenges persist with relying solely on scaling, particularly due to the finite amount of publicly available data. Increasing computational power necessitates the construction of new data centers, which can be both costly and environmentally impactful. Furthermore, some experts are expressing concerns over diminishing returns from massive investments in scaling. Yann LeCun, Meta's chief AI scientist, who is transitioning to lead his own startup, argues that the industry must explore alternative approaches. He pointed out, "Most interesting problems scale extremely badly," during a recent talk at the National University of Singapore. LeCun's new venture aims to develop world models that focus on spatial data rather than relying solely on language processing. His objective is to usher in a new era of AI capable of comprehending the physical world, maintaining persistent memory, reasoning, and executing complex action sequences, as he outlined in a LinkedIn post last November.