How Video Games Ushered In the A.I. Revolution

New York Times

05-04-2025

A challenge of writing a book on the tech industry, especially something as rapidly evolving as artificial intelligence, is that the story will be slightly out of date in the few months between the final draft being turned in and the hardback hitting the shelves. You wish you could tell the future, but you're stuck with what you have when you send the book to the printer.
For example, Stephen Witt's 'The Thinking Machine,' a lively biography of the C.E.O. Jensen Huang, whose company Nvidia makes microchips that power A.I. systems like ChatGPT, came out this spring, but it recounts events only up to a mid-2024 climactic showdown between the author and his subject over the possibility of A.I. destroying humanity, which means that a line that appears in an earlier chapter — about how Elon Musk differs from Huang in temperament — mentions that the Tesla C.E.O. has 'at least' 11 children. That count is now woefully behind. By most estimates, he's up to 14.
It also means Witt's account doesn't include the recent drama that arose after the release of a new A.I. chatbot from the Chinese company DeepSeek. A rival to ChatGPT, the Chinese chatbot was allegedly built for a fraction of the cost, with fewer fancy chips. This bucked the accepted wisdom that the only way to improve A.I. was to shovel vast sums of money at Nvidia to buy more and more of its hardware. When the markets absorbed this fact in January, Nvidia's stock price tumbled.
Before that fall, however, there was an astonishing rise. The story of how Nvidia became the hottest investment on Wall Street and a household name is fascinating because its trajectory differs significantly from that of its Big Tech peers. For most of the time that companies like Apple, Meta and Amazon have been around, regular people used their products and services every day. But, unless you were a hard-core gamer, you probably hadn't heard of Nvidia until recently.
Huang doesn't offer Witt much on how his upbringing may have led to his current status as a technology apex predator ('I try not to analyze myself in that way,' he tells Witt), but there are early glimpses of his incredible drive and focus. In 1973, at 10 years old, he immigrated to the United States from Thailand and eventually landed in Oregon, where, between homework and shifts at Denny's, he played competitive Ping-Pong at the national level.
By the early '90s, popular video games like Myst and Doom were coming out and the industry around personal computing was ramping up. Like many ambitious people on the West Coast, Huang had been looking for a way to rise up through the hardware market. In 1993, instead of trying to compete against giants like Intel and Sun in the general computer chip space, Huang co-founded Nvidia, a company focused specifically on P.C. video games; their chips were robust enough to process the immersive visuals that the new games were creating.
This push for extra processing power would come in handy down the line, but for much of Nvidia's history, success was far from assured. Over 30 years, the company had ups and downs, nearly facing bankruptcy and fighting off activist investors.
Huang's tolerance for risk pulled his company through again and again. Huang was also notorious for his management style; his trademark technique is rage and yelling. In 2008, one of the company's new graphics chips had a design flaw that caused mass customer returns and a plunge in stock price. In front of a large group in the company cafeteria, including more than a hundred executives, Huang reamed out the chip architect responsible for the error. 'I still vividly remember Jensen just nonstop berating him for a good hour and a half,' an employee who was there tells Witt. 'Honestly, maybe it was two hours — he was just livid.'
The screaming and rage sessions seem at odds with other, kinder aspects of Huang's personality — his friends from his personal life said they didn't ever witness any blowups. And somehow he's retained many longtime loyal employees, even people from the early days when it wasn't obvious that working at a gaming chip company — not the most glamorous part of the tech industry — was a golden ticket to enormous wealth. He's also a loyal boss: The guy from the cafeteria wasn't fired after the showdown.
Over and over, Huang made decisions that worked out. He hired well, he saw opportunities around corners. The smartest risk Huang took was listening to a midlevel researcher who, in 2013, pitched him on a technology called 'neural nets,' then a fringe area being explored by a handful of academic researchers. Huang saw the potential, and set Nvidia on a path that would make his chips the premier tool for today's A.I. revolution.
I should offer a warning here: 'The Thinking Machine' does contain a fair amount of somewhat technical passages about A.I. and the making of the hardware it runs on. Witt does his best to render all this legible to a lay audience, but you also won't really lose much if you just sort of glaze over those parts.
Witt's previous book, 'How Music Got Free,' a history of digital music piracy, had some technical baggage too, describing in detail how the MP3 file format was first developed, but the telling was balanced by the compelling cat-and-mouse drama between record label executives and a CD factory employee who was pirating their music.
His task here is trickier. The arc of Nvidia's rise to dominance can't really be built out of explosive interpersonal moments at the office or reflections on its founder's stranger personality traits, as entertaining as those things are. Instead, the drama most naturally arises from a series of technical achievements where Huang and his gang pulled ahead of the competition by sheer feats of computer engineering. Thankfully, Witt does a decent job at drawing the reader into those moments.
It's hard not to compare Huang with Zuckerberg, Bezos and Musk and see him as a kinder, less evil version of his tech overlord peers. For all his verbal abuse, he hasn't attempted to reshape global society or exploit low-wage workers.
Witt's book suggests that Huang built a highly profitable company by dint of his own brilliance, hard work and a bit of luck. Still, as compelling as Huang and his decades-long journey to get to the present are, the big questions about what's ahead for A.I. are left unanswered: How will it change humanity? Will A.I. lift us up or destroy us? When Witt tries to ask Huang about these things, the C.E.O. brushes him off, saying, 'I feel like you're interviewing Elon right now, and I'm just not that guy,' before getting irate and yelling.
That Huang isn't worried about a dystopian runaway A.I. should be comforting, but, on the other hand, he's obviously got a financial interest in A.I. continuing to spread into every aspect of our lives, and his company, although weakened, is still the clear market leader.
In the end, 'The Thinking Machine' leaves us unsure of its subject's vision of the future, and Huang will probably never give a satisfying answer. He is someone who dreamed of beating Intel in Q2 sales numbers, not of completely changing the world and ushering in a new technological age. It just so happens that by achieving the first goal, he also ended up doing the latter.