SEOUL – Sam Altman, CEO of OpenAI, has weighed in on the recent debate surrounding AI overinvestment and the pursuit of low-cost, high-performance AI, sparked by the emergence of Chinese AI company DeepSeek. In a blog post titled "Three Observations," Altman reiterated his belief that the cost of using AI is decreasing tenfold annually and emphasized the imminent arrival of artificial general intelligence (AGI), an AI system that matches or surpasses human capabilities.
Altman acknowledged the rapid pace of AI development, stating, "We continue to witness the rapid progress of AI development." He outlined his vision for AGI and highlighted three key trends related to AI's economic viability. Firstly, he argued that the intelligence of AI models is directly proportional to the resources used for training and execution. This counters the notion that massive investments in AI development do not guarantee proportional improvements in model performance. Altman cited the "Scaling Law," which suggests that AI development follows a pattern where increased parameters and data lead to better results.
Secondly, Altman emphasized the dramatic decrease in the cost of using AI, stating, "The cost of using AI has decreased approximately tenfold each year, and this decrease in cost is driving a surge in AI usage." He pointed to the significant price reduction per token (data unit) between GPT-4, released in 2023, and GPT-4o, released last year, as evidence of this trend. Altman highlighted that the current rate of AI development surpasses Moore's Law, which predicts a doubling of semiconductor integrated circuit performance every 18 months.
Thirdly, Altman stressed the immense socioeconomic value that AI development will bring, arguing that there is no reason to halt investment in the field. He suggested that the ripple effects of AI will far outweigh its performance advancements, indicating that the "DeepSeek shock" should not lead to decreased AI investment but rather encourage further investment.
Looking ahead to the era of AGI, Altman shared his thoughts on the potential changes and concerns. He predicted the rise of AI agents that function as virtual assistants and colleagues, transforming knowledge work. He also acknowledged the potential for AGI to exacerbate existing inequalities and warned against the use of AI for mass surveillance and control by authoritarian governments.
Despite the potential challenges, Altman expressed optimism about the future of AI, stating, "Our goal is for AGI to benefit all of humanity." He envisioned a world where AGI could lead to economic growth, disease eradication, increased family time, and the full realization of creative potential. Altman concluded, "Within the next 10 years, we may enter an era where everyone on Earth can achieve more than the most influential people today."
[Copyright (c) Global Economic Times. All Rights Reserved.]
