• Copy link Facebook X Whatsapp Reddit Pinterest Flipboard Email Get Tom’s Hardware’s best news and in-depth reviews, straight to your inbox. • You are now subscribed Your newsletter sign-up was successful OpenAI CEOSam Altmantook part in a wide ranging Q&A on Friday, answering dozens of rapid-fire questions duringa 60 minute sessionhosted by The Indian Express. • Not for the first time, Altman stoked controversy. • This time, he bemoaned “unfair” comparisons between the efficiency of AI inference queries and human thought. • In Altman’s view the comparison is skewed as humans have millennia of evolutionary smarts and technology teachings behind them, yet individuals require “like 20 years of life and all of the food you eat during that time before you get smart.” 🚨 SAM ALTMAN: “People talk about how much energy it takes to train an AI model … But it also takes a lot of energy to train a human. • It takes like 20 years of life and all of the food you eat during that time before you get smart.” pic.twitter.com/vRuVnnmzjBFebruary 21, 2026 Photonics and high-speed data movement is the next big AI bottleneck The data center cooling state of play Massive AI data center buildouts are squeezing energy supplies Ultra Ethernet: The data center interconnection of tomorrow Chief Nerd clipped the eyebrow-raising Q&A segment for convenient sharing.
Article Summaries:
- OpenAI CEO Sam Altman criticized “unfair” comparisons that pit AI inference energy use against human cognition during a 60‑minute Q&A hosted by The Indian Express. Altman argued that training a human also consumes significant energy-“20 years of life and all the food you eat before you get smart”-and that, when accounting for the evolutionary and cultural energy invested in human intelligence, AI may already match or exceed human energy efficiency. He urged greater use of sustainable power for AI workloads. The session followed a high‑profile meeting between AI leaders and Indian Prime Minister Narendra Modi, underscoring India’s role in AI development.
Sources: