“The Worlds I See,” by Fei-Fei Li
Macmillan Publishers/TNS
It has been an overwhelming 12 months in the technology world. Rapid developments in artificial intelligence, which went into overdrive in late 2022 with the release of ChatGPT, have generated countless column inches, panel discussions and political debate.
For those who want to contribute to the discourse, understanding the core issues at play has never been more important. Yet the deluge of hype, news coverage, fear-mongering and misinformation can leave us feeling lost. What are the innovations that truly matter? What are the consequences worth thinking about?
It helps, then, to take a good step back and to take the time to slowly ingest information and consider what may lie ahead. As we enter 2024, our technology columnists wanted to share the books they read (or read again) in the past year that have helped frame their thinking. Dave Lee is Bloomberg’s U.S. technology columnist, based in New York City. Parmy Olson covers AI and the tech industry from London.
The Worlds I See — Fei-Fei Li, 2023
Fei-Fei Li was a young woman from China when she arrived in the U.S. in the early 1990s. Back then, with barely any grasp of English, let alone computer vision, she could have no idea how profoundly her work would set us on course for the current AI boom. Her project ImageNet, which to date has annotated more than 14 million images for use to train AI algorithms, laid the groundwork that made the work of pioneering AI companies like DeepMind and OpenAI possible. In “The Worlds I See,” Li weaves the story of her work tightly with her own personal journey: breaking through as an immigrant and building her way up to being one of the most respected figures in her field. Today she is co-director at the Stanford Institute for Human-Centered Artificial Intelligence and thus one of the most powerful voices lobbying on behalf of the human race as Silicon Valley speeds ahead. I found her tales not only fascinating, and humbling, but also a path to sleeping easier at night, glad she’s in our corner. At this delicate period of wondering whom we should entrust with our cutting-edge AI future, Li’s résumé is most encouraging. Dave Lee
Your Face Belongs to Us — Kashmir Hill, 2023
The human face’s romantic role as the “window” to our soul has a colder analogy for computer algorithms: just another kind of fingerprint. Kashmir Hill’s gripping book explains how a startup called Clearview AI brazenly scraped the internet to amass a database of billions of faces, then worked with dozens of police departments around the U.S. to create a hidden surveillance tool. Hill was the first reporter to identify Clearview’s sprawling network and lays out how the explosion of video doorbells and security cameras that can identify faces are turning our world into a panopticon, and a racist one at that. Facial recognition software often isn’t trained to recognize the faces of Black people properly, and Hill shows the consequences in a terrifying chapter when Robert Julian-Borchak Williams is arrested in front of his family and jailed, all because of a flawed facial “match.” Even when the algorithms are near-perfect, as with Clearview’s, they aren’t designed with civil liberties in mind. Hill also lays out how Clearview founder Hoan Ton-That spent years aligning himself with far-right figures, many of whom would love to track the faces of undocumented immigrants. What it means to go “outside” is fundamentally changing, she writes. And not in a comforting direction. Parmy Olson
I, Robot — Isaac Asimov, 1950
It’s an embarrassing admission, from a technology columnist in his mid-30s, to share that until this year I had not read any of Isaac Asimov’s work. But, scolded by a peer, and mindful Asimov’s Three Laws of Robotics might fast become actual government policy, it seemed wise to quickly dive in. The short stories in “I, Robot,” a compendium released in 1950, dreamed up some of the questions that are very real today as we edge closer to creating so-called artificial general intelligence. Asimov’s laws — the first being that “a robot may not injure a human being or, through inaction, allow a human being to come to harm,” and the others stemming from that — seem straightforward but form the basis of plot lines that remind us today of the unpredictable ways in which smart robotics or AI could go wrong. Some 80 years after being written, the foresight holds up. Case in point: If you weren’t told otherwise, you’d think Asimov himself had written the recent discussions around ChatGPT becoming “lazier” in December because it learned humans did less work around Christmas. Dave Lee
Going Infinite — Michael Lewis, 2023
The big takeaway from this book isn’t so much the story of Sam Bankman-Fried’s downfall but how a man with obvious signs of narcissism and deception could charm one of the smartest journalists on the business beat. Michael Lewis is best known for poking holes in the stories of smarmy traders, but he paints a strangely sympathetic portrait of Bankman-Fried in Going Infinite, flicking at rude acts like playing video games in the middle Zoom meetings or mistreating his girlfriend as a feature and not a bug. Lewis’ neglect of Bankman-Fried’s flaws gets egregious when he describes the young savant’s attitude toward investors in his fund. Relegated to a footnote (a footnote!) is the startling admission from Bankman-Fried that he would have answered a different question or “made a word salad” if any of his investors asked about the risk engine at the core of his trading fund and where all their money was. To get some insight into how “tech bros” like Mark Zuckerberg and Elon Musk will continue to maintain a network of powerful supporters despite their bad behavior, read this book. Parmy Olson
