Major AI breakthroughs are largely concentrated in a handful of leading research labs, according to Demis Hassabis, Co-Founder and CEO of Google DeepMind. He believes algorithmic innovation will be crucial for these labs to maintain their competitive edge.
Hassabis defines Artificial General Intelligence (AGI) as a system capable of mimicking all human cognitive capabilities. He suggests that achieving AGI within the next five years is a realistic possibility, marking a significant milestone in AI technology.
Computational resources are vital for both scaling AI systems and validating new ideas, Hassabis noted. He expects DeepMind to continue leading in AI breakthroughs, despite current systems lacking continuous learning capabilities post-training and struggling with long-term planning and consistency.
Open source models typically lag behind frontier models by about six months, underscoring the rapid pace of innovation within leading labs.