There is a new “era” in the artificial intelligence (AI) field. Recently, Jensen Huang, the CEO of Nvidia, hinted that “artificial general intelligence” (AGI) might become a reality within the next five years. However, many experts argue that this innovative technology does not present an immediate threat to human jobs or lifestyles. Huang, the leader of the renowned global company specializing in AI chip manufacturing, discussed the timeline for achieving Silicon Valley’s longstanding objective of creating computers capable of human-like thinking at an economic forum held at Stanford University. He stressed that the progress of this timeline heavily relies on how the goal is defined. If the benchmark includes excelling in tests designed for humans, he suggested that we could soon witness the emergence of artificial general intelligence (AGI).
AGI, short for artificial general intelligence, is often misconstrued by the general public in comparison to AI. While AI functions based on specific instructions, AGI aims for autonomous intelligence, as elucidated by Duncan Curtis, the Senior Vice President of AI Product and Technology at the AI data firm Sama. Models like ChatGPT may encounter difficulties in generating entirely new content or integrating diverse information to solve unfamiliar problems, occasionally displaying excessive confidence even when not entirely accurate. In contrast, AGI is designed to effortlessly address new challenges, continually learn, and improve its capabilities over time.
Despite Huang’s recent prediction, numerous experts believe that AGI is not imminent. The term “AGI” remains vague as there is no exact definition of intelligence, according to tech advisor Vaclav Vincalek. Yann LeCun, Meta’s chief scientist and a pioneer in deep learning, has previously questioned Huang’s claims regarding the imminent arrival of AGI. LeCun argues that current AI systems are still years away from achieving a level of consciousness that encompasses common sense. He proposes that society is more likely to develop AI comparable to the intelligence levels of animals like cats or dogs before reaching human-level AI. LeCun also emphasizes that solely focusing on language models and textual data in the tech industry is insufficient to realize the advanced AI systems envisioned by experts over the years.
