 | A WRITER'S WIT |  Y. Albats |
Hao, Karen. Empire of AI: Dreams and Nightmares in Sam Altman’s OpenAI. New York: Penguin, 2025
I occasionally read out of my fields of study (literature, writing, and music) or out of my depth, to give myself a challenge. In the case of this title, I have done both, but once having begun the task, I was determined to finish the book. I perhaps liken the experience to being a second-year French student attempting to read Camus or Molière; you might comprehend only a fraction of what you’re reading. Problems for the general citizen hoping to understand Artificial Intelligence (AI) can be considerable: apprehending modern technology in general; absorbing specialized vocabulary (AI engenders a very slick and plentiful argot); and fully grasping the culture of people in this field who may believe that AI is capable of solving all human problems, and that there is very little chance that AI could ever advance far enough to think for itself and thus rule our lives without human intervention.
As the subtitle would seem to indicate, the book is about one forty-year-old-man named Sam Altman, the CEO of his own company, OpenAI. It begins as a nonprofit, using its lofty goals of employing AI technology to benefit all humanity. However, the organization gets lost in the weeds along the way. Repeatedly, throughout the book, Hao demonstrates a certain duplicity on the part of Altman. He assuages one party in the company by taking their side and then does the very same thing with a different party. Then he acts all What? Me? when confronted. It is a behavior that cannot be sustained over time, especially in light of the fact that the company, and Altman particularly, begin to make money by the fistful.
There are two chapters out of the twenty or so that regular people will more readily understand. One is when Hao takes readers to South America (Venezuela) to view how OpenAI exploits its employees there to perform tasks for cheap that no one in the States would do: really tedious work for which workers are paid literally pennies for each task. In other words, AI uses—let’s be honest—slave labor. Oh, and along with its other insipient problems, OpenAI consumes a great deal of electricity which, in turn, creates a great deal of heat. Such heat must be reduced by using cooling units like water, exhausting a resource already in short supply in Venezuela (later the company does develop air-cooled methods).
Another chapter is about Annie, Altman’s younger sister, who, through a series of misadventures winds up nearly destitute. Her family members believe that she is largely responsible for her own misfortunes, and they do not help her beyond minimal financial gestures. Sam Altman, the family member who, by now is the most able to help his sister, does perhaps the least. The situation becomes stickier when she accuses him, quite convincingly to readers (although Hao makes it clear we cannot know for sure), that he has sexually molested her throughout her childhood. The remaining chapters of the book, at least to this reader, are largely written for people in the biz.
Readers must understand the vocabulary of AI, that “compute” is no longer a verb but a noun, according to Hao, “a term of art for computation resources, that Open AI would need to achieve major breakthroughs in AI capabilities” (59). As in “Does the program have enough compute to do the job?” Although Hao does define certain words from their first use, it would also be nice for the publisher to provide readers with a glossary should there be a second edition in the offing (that is if the publisher is really attempting to inform the hoi polloi about AI). Despite this shortcoming, Hao is a well-respected and well-informed journalist who has done a tremendous multi-year job of researching the heck out of this book, yet, for the most part, it is not easy for the non-IT person to understand. Seems as if there might be two books squeezed between these covers: one about AI and its many wonders and downfalls, its unexplored routes of endeavor. The other is about the miscreant personality of one man who gains power too early in his life and does not know how to manage it. If one wants to read about how these two subjects merge, then this is that book. Happy reading!
I occasionally read out of my fields of study (literature, writing, and music) or out of my depth, to give myself a challenge. In the case of this title, I have done both, but once having begun the task, I was determined to finish the book. I perhaps liken the experience to being a second-year French student attempting to read Camus or Molière; you might comprehend only a fraction of what you’re reading. Problems for the general citizen hoping to understand Artificial Intelligence (AI) can be considerable: apprehending modern technology in general; absorbing specialized vocabulary (AI engenders a very slick and plentiful argot); and fully grasping the culture of people in this field who may believe that AI is capable of solving all human problems, and that there is very little chance that AI could ever advance far enough to think for itself and thus rule our lives without human intervention.
As the subtitle would seem to indicate, the book is about one forty-year-old-man named Sam Altman, the CEO of his own company, OpenAI. It begins as a nonprofit, using its lofty goals of employing AI technology to benefit all humanity. However, the organization gets lost in the weeds along the way. Repeatedly, throughout the book, Hao demonstrates a certain duplicity on the part of Altman. He assuages one party in the company by taking their side and then does the very same thing with a different party. Then he acts all What? Me? when confronted. It is a behavior that cannot be sustained over time, especially in light of the fact that the company, and Altman particularly, begin to make money by the fistful.
There are two chapters out of the twenty or so that regular people will more readily understand. One is when Hao takes readers to South America (Venezuela) to view how OpenAI exploits its employees there to perform tasks for cheap that no one in the States would do: really tedious work for which workers are paid literally pennies for each task. In other words, AI uses—let’s be honest—slave labor. Oh, and along with its other insipient problems, OpenAI consumes a great deal of electricity which, in turn, creates a great deal of heat. Such heat must be reduced by using cooling units like water, exhausting a resource already in short supply in Venezuela (later the company does develop air-cooled methods).
Another chapter is about Annie, Altman’s younger sister, who, through a series of misadventures winds up nearly destitute. Her family members believe that she is largely responsible for her own misfortunes, and they do not help her beyond minimal financial gestures. Sam Altman, the family member who, by now is the most able to help his sister, does perhaps the least. The situation becomes stickier when she accuses him, quite convincingly to readers (although Hao makes it clear we cannot know for sure), that he has sexually molested her throughout her childhood. The remaining chapters of the book, at least to this reader, are largely written for people in the biz.
Readers must understand the vocabulary of AI, that “compute” is no longer a verb but a noun, according to Hao, “a term of art for computation resources, that Open AI would need to achieve major breakthroughs in AI capabilities” (59). As in “Does the program have enough compute to do the job?” Although Hao does define certain words from their first use, it would also be nice for the publisher to provide readers with a glossary should there be a second edition in the offing (that is if the publisher is really attempting to inform the hoi polloi about AI). Despite this shortcoming, Hao is a well-respected and well-informed journalist who has done a tremendous multi-year job of researching the heck out of this book, yet, for the most part, it is not easy for the non-IT person to understand. Seems as if there might be two books squeezed between these covers: one about AI and its many wonders and downfalls, its unexplored routes of endeavor. The other is about the miscreant personality of one man who gains power too early in his life and does not know how to manage it. If one wants to read about how these two subjects merge, then this is that book. Happy reading!
Up Next:
TUES 9/9: A Writer's Wit | Kimberly Willis Holt
WEDS 9/10: A Writer's Wit | Faith Hunter
THURS 9/11: A Writer's Wit | Andre Dubus III
FRI 9/12: A Writer's Wit | Han Suyin
My Book World | Federico García Lorca, The Selected Poems of Federico García Lorca
TUES 9/9: A Writer's Wit | Kimberly Willis Holt
WEDS 9/10: A Writer's Wit | Faith Hunter
THURS 9/11: A Writer's Wit | Andre Dubus III
FRI 9/12: A Writer's Wit | Han Suyin
My Book World | Federico García Lorca, The Selected Poems of Federico García Lorca
RSS Feed