Altman Compares AI Training to Human Development

[ed]

Altman also criticized what he described as “unfair” comparisons between AI energy use and human effort. He argued that critics often compare the energy required to train large AI models to the energy needed for a single human task.

“But it also takes a lot of energy to train a human,” Altman said, noting that decades of food consumption, education, and even centuries of human evolution underpin individual intelligence.

In his view, the fairer comparison is between a trained AI model responding to a query and a human answering the same question. Measured that way, Altman suggested AI may already be competitive — or even more efficient — on an energy basis.

The remarks reflect ongoing debates about the environmental trade-offs of large-scale AI systems, particularly as global adoption accelerates and infrastructure demands expand.

[cryptobro]

altman basically said if we’re doing this math let’s at least use the same denominator