The People Who Made the Machine

The Book

Karen Hao is a journalist who covered OpenAI for years. Her book is called Empire of AI, and her argument is simple. The company takes labor and resources from poor countries and keeps the money. Then it calls the whole thing a mission to help humanity.

One story from the book tells you most of what you need. A man named Mophat Okinyi lives in Nairobi, Kenya. A staffing company called Sama hired him to work for OpenAI. His job was reading the worst text the model could produce, hour after hour, so ChatGPT would learn to refuse it. The pay was under two dollars an hour. Billy Perrigo, a reporter at TIME magazine, published that number first. Hao reports what the work did to the man, and it is hard to read.

The book covers the physical side too. Data centers use huge amounts of water for cooling. New ones are going up in central Chile while the country sits fifteen years into a drought. The towns hand over land and tax breaks. The water goes to the servers, and the profit goes to another country.

The Part Nobody Says Out Loud

Here is how the product got made. People posted answers on the internet for free, millions of them, for years. They were helping strangers and asking nothing for it. OpenAI copied all of it. Then workers in Kenya and Venezuela cleaned up the raw output for a dollar or two an hour. Then the company sold the result and called it machine intelligence. It is worth almost a trillion dollars now.

The computer did not create that knowledge. People did.

The companies have started admitting how much of their story was sales talk. In spring 25 the CEO of Anthropic said AI would write ninety percent of code within six months. IT Pro, a technology news site, checked that claim in Sep 25, and the real number came in near thirty percent. This past May, Fortune reported that Sam Altman said he was "delighted to be wrong" about the job losses he had spent years predicting. Then both companies filed to sell shares to the public in Jun 26, and both are hiring programmers as fast as they can. The scary talk ended right when the stock paperwork began.

What a Small Shop Should Take From This

The lesson is about who runs the tool. A small model trained on one trade, run by somebody who knows that trade, beats a giant rented one run by somebody who doesn't. The big companies hide the human operator because a hidden operator is what sells the product. Your move is the opposite. Put the operator, meaning you, in charge.

That is not a slogan, it is a working setup. The tools for running a model on your own computer are free and public. The hardware costs less than a service van. What the big companies rent you by the month, a shop can own outright, sized to the one trade it actually does.

Here is the concrete version. A fire alarm company could train a small model on its own service records and its own panel documentation. That model would answer questions about the exact systems the company maintains, and nothing else. The data never leaves the shop, and the shop owns the result. None of that requires a trillion dollar company. It requires a used graphics card and somebody who knows the work.

Your knowledge of your own work is the asset. The people in Hao's book got paid two dollars an hour to hand that kind of knowledge over. You don't have to. Nobody can rent you what you already own.

Next: the full review of Empire of AI