Atlas of AI: Power, Politics, and the Planetary Costs of Artificial Intelligence

Reading Atlas of AI in 2026 is a necessary reality check. When you work with these systems every day, it’s easy to get caught up in the software and the "cloud" and forget what actually powers it.

Crawford’s best point is reminding us that AI isn't magic, It’s physical. It’s not just algorithms. It’s lithium mined from the earth and real people working long hours to label data. I really appreciated that shift in perspective. It makes you realize that "computing" is actually a heavy industrial process involving energy and labor, not just invisible math.

However, I think the book is a bit too eager to paint the whole industry as the villain. The author often blames the scientists who built these tools for how they are used. I disagree with that. The researchers gave us incredible technology that has helped humanity in huge ways. The real issue isn't the invention itself, but when governments or companies blindly trust imperfect AI without testing it enough, and when corporations oversell technology that is actually far from their promises.

Some parts of the book also feel a little outdated now, given how much LLMs have advanced by 2026. But the main lesson still holds up: we need to stop looking at AI as just code and start seeing the physical and human chain behind it. It’s a good read if you want to understand the "hidden costs" of the industry, even if it is a bit one-sided.

— Oleksandr ★★★★4, прочитано у 2026

Тем ще немає. Обговорити в Клубі

Опис: Claude, коротко про книжку