Sam Altman Compares Training AI To Raising Kids

2026-03-01 · Source: Matt Wolfe · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Fundamental Awareness, quick

Summary

Sam Altman recently compared the process of training an AI model to raising a human child, sparking discussion. He argues that while critics often highlight the energy consumption of AI training versus a single human inference query, a fairer comparison considers the total energy expenditure to "train" a human. Altman points out that human development involves approximately 20 years of life, continuous food consumption, and the cumulative learning of billions of ancestors. He suggests that when comparing the energy cost of a single query from a trained AI model like ChatGPT to a human, AI is likely already more energy-efficient, as it bypasses the extensive developmental and evolutionary overhead required for human intelligence.

Key takeaway

For AI Product Managers evaluating the long-term sustainability and cost-effectiveness of AI systems, consider Altman's perspective on energy efficiency. Your analysis should encompass the full "training" lifecycle of both AI and human intelligence, including developmental and evolutionary costs, to accurately assess AI's comparative efficiency rather than focusing solely on operational inference costs.

Key insights

AI training costs should be compared to the full developmental and evolutionary costs of human intelligence.

Principles

• Fair comparisons require equivalent scope.
• Human intelligence has immense hidden costs.

Topics

• AI Training
• Energy Efficiency
• Human Development
• Sam Altman
• AI Comparisons

Best for: AI Product Manager, CTO, AI Ethicist

Related on AIssential

• Neil Thompson: Is there a 'Secret Sauce' driving AL Model development? — AI model development combines systematic efficiency with significant, art-like variation in outcomes.
• Why comparisons between AI and human intelligence miss the point — Human intelligence is fundamentally collective, embodied, and social, a dimension AI systems currently lack.
• Distinguish between inference scaling and "larger tasks use more compute" — Distinguishing between task size and efficiency is crucial for understanding LLM inference scaling and its economic implications.
• "How to measure intelligence?" | Six researchers debate — Current AI lacks human-like developmental learning, struggling with common sense despite advanced problem-solving.
• AI IQ is here: a new site scores frontier AI models on the human IQ scale. The results are already dividing tech. — AI IQ provides a unified framework for benchmarking LLMs across IQ, EQ, and cost, despite methodological debates.
• LLM Vs AI: A Practical Guide to Differences, Use Cases, and Tools — Generative AI and LLMs are transforming enterprise operations through diverse applications, requiring careful selection and deployment.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Matt Wolfe.