
Training Humans vs. Training Machines
There is a particular species of genius that flourishes in Silicon Valley, the kind that confuses audacity for wisdom, disruption for progress, and a keynote address for something resembling a moral philosophy. One of these high priests Silicon Valley is Sam Altman, who, speaking at an event hosted by The Indian Express, addressed concerns about artificial intelligence and answered several inconvenient questions regarding the water and energy consumption of the AI industry.
Let’s first take water. Worries about AI’s water usage are, in Altman’s view, “totally fake” — or at least, they were legitimate once, back in the dark ages of evaporative cooling. Anyone suggesting today that ChatGPT consumes alarming quantities of water is spreading something with “no connection to reality.” Reassuring, given that there is currently no legal requirement for tech companies to disclose how much energy or water they actually use. We’ll simply have to take his word for it.
Energy consumption, he conceded, presents a “fair” concern, though his defense soon took a more philosophical turn. Critics of AI’s appetite for electricity, he suggested, might reflect on the fact that human beings themselves are hardly paragons of efficiency. Do you know how much energy takes to train a human mind?
Training a human, he explains, takes twenty years of life, mountains of food, and the entire evolutionary history of one hundred billion people who somehow survived predators, disease, famine, and war long enough to reproduce.
If you really think about it, you will acknowledge that this logic reveals a kind of elegant circularity: your children — those inefficient, slow little mammals demand two decades of caloric and emotional investment before contributing meaningfully to society. Never mind that this comparison conflates two entirely different categories of necessity. We’ are talking about a computational scandal of the first order.
What makes this genuinely worth pondering, beyond the spectacle of a man debunking inconvenient studies from a stage he presumably reached aboard a private aircraft, is the broader context. The individuals shaping artificial intelligence are not merely developing another consumer product. They are designing what may be the most consequential technology in human history, one already altering economies, labour markets, political structures, and the very future of our species And their defence of social and environmental responsibility reduces to this: evolution was pricey too.
Well, I suppose every civilisation gets the oracles it deserves.
Thanks for reading Chronicles of a Liquid Society! This post is public so feel free to share it.
✍️ Written by
Newsletter
Stay updated with new articles.

No comments yet. Be the first to comment!