This is the newest affirmation that the “nice man” principle of historical past continues to thrive in Silicon Valley. As such, it joins a style that features Walter Isaacson’s twin tomes on Steve Jobs and Elon Musk, Brad Stone’s ebook on Jeff Bezos, Michael Becraft’s on Invoice Gates, Max Chafkin’s on Peter Thiel and Michael Lewis’s on Sam Bankman-Fried. Notable traits of the style embody a bent in direction of founder worship, discreet hagiography and a Whiggish interpretation of the life underneath examination.
The nice man underneath Witt’s microscope is the co-founder and chief govt of Nvidia, a chip design firm that went from being a small however plucky purveyor of graphics processing models (GPUs) for pc gaming to its present place because the third most useful firm on this planet.
Two issues drove this astonishing transition. One was Jensen Huang’s intuitive appreciation that Moore’s regulation – the commentary that computing energy doubles each two years – was not going to use for ever, and {that a} radically totally different sort of computing structure can be wanted. The opposite was his resolution to wager the way forward for Nvidia on that proposition and switch the corporate on a dime, a lot as Invoice Gates had achieved with Microsoft within the Nineteen Nineties when he had realised the importance of the web.
Huang was born in Taiwan however introduced up in Thailand and the US, of which he’s now a citizen. He was a intelligent, diminutive, pushed little one who survived bullying at a boarding faculty in Kentucky however thrived academically. After graduating from Oregon State College he first labored at Superior Micro Gadgets (AMD), a Silicon Valley chip firm that competed with semiconductor large Intel, earlier than shifting to LSI Logic, an revolutionary agency that developed the primary design instruments for chip architects. So he was in on the semiconductor business from the very starting of his profession.
When Chris Malachowsky and Curtis Priem, two chip designers who labored at Solar Microsystems, the fabled Silicon Valley workstation producer, determined to arrange a startup, they requested Huang to affix them, on the grounds that he was already an achieved and profitable product supervisor at LSI. And so Nvidia was born, and Witt’s narrative actually will get going.
Like most startups, Nvidia had preliminary ups and downs, however quickly turned recognized for offering the sort of processors that gamers of video games comparable to Doom and Quake valued. As a result of correct rendering of images in video video games required computing energy that typical serial-processing processors struggled to offer, Nvidia moved to experimenting with chips that might do a number of calculations concurrently, often called parallel processing, and thus to the event of GPUs.
Parallel processing supplied an astonishing advance in computing energy and velocity. However writing software program to harness the know-how proved difficult: consider it as switching from considering in three dimensions to considering in 5,000. With the intention to make parallelism extra extensively usable, Nvidia launched into a secret challenge, codenamed Cuda, to create a software program platform that may make GPUs – hitherto used just for gaming – into instruments that could possibly be helpful for any activity requiring brute computing energy.
Cuda turned out to be a masterstroke in two methods: first, it instantly elevated the vary of purposes wherein GPUs could possibly be deployed – and due to this fact the scale of the potential market; and second, though the software program was free, it solely labored with Nvidia {hardware}!
The one remaining drawback was to discover a market with an insatiable demand for high-end processors. And, as luck would have it, one such market hove into view, crusing underneath the banner of AI (AKA massive language fashions), the coaching of which required unconscionable quantities of computing energy. Immediately the tech giants developed an insatiable starvation for Nvidia’s little supercomputer modules. The AI gold rush was on, and Huang was the person promoting the picks and shovels – and on his method to a web value of $100bn.
It’s an awesome story and Witt tells it properly. He paints a rounded image of a exceptional entrepreneur – half visionary, half maniacal workaholic, half inspiring company chief. Oh, and half ranting screamer: “yelling at folks is a part of his motivational technique”, says one worker; he has “embraced candour to the intense” says one other, tongue firmly in cheek. So it’s a “warts and all” portrait, as Oliver Cromwell would have put it.
And simply to underline that time: when, in his closing interview for the ebook, Witt asks him what new jobs may be created by AI, Jensen loses it and begins shouting at him. “His anger appeared uncontained, omnidirectional, and wildly inappropriate,” he writes. “I used to be not Jensen’s worker, and he had nothing to achieve from raging at me. He simply appeared uninterested in being requested concerning the adverse features of the instruments he was constructing. He thought the query was silly, and he had been requested it one too many instances.”
-
The Pondering Machine: Jensen Huang, Nvidia and the World’s Most Coveted Microchip by Stephen Witt is revealed by Bodley Head (£25). To assist the Guardian and Observer order your copy at guardianbookshop.com. Supply costs might apply
Leave a Reply