The world becomes more amazing by the day. The surest evidence of the previous assertion is that wealth inequality continues to soar.
Yes, you read that right. Inequality is a wonderful thing opposite the apologetic tone about it taken by left and right. Members of the left plainly disdain it, while members of the right claim it doesn’t exist in the way the left imagines, that thanks to transfer payments (yes, wealth redistribution by government) inequality isn’t that “bad.”
Except that it’s not bad. It’s great. How could remarkable achievements that power wealth creation be bad? If someone comes up with a cure for paralysis, will readers demand that the cure never reach the marketplace lest the creator grow rich?
The beautiful, freedom-infused meaning of wealth inequality came to mind while reading Stephen Witt’s new book, The Thinking Machine: Jensen Huang, Nvidia, and the World’s Most Coveted Microchip. Nvidia co-founder Huang was already billionaire bracket rich when Nvidia was largely a company popular with gamers, but it’s been Nvidia’s central role in the rise of AI that will increasingly do and think for us in ways that will propel billions into amazing work they can’t not do, that has resulted in Huang achieving centi-billionaire status.
Thank goodness for inequality. It signals stupendous, life-enhancing improvement. Huang was already a relentless worker, but it was in 2014 when employee Bryan Catanzaro truly opened Huang’s eyes to the possibilities of AI, and Nvidia’s potential to lead it. After that, and in the words of Witt, “there was only work” for Huang. “There was only AI.”
As Huang saw it, AI was “O.I.A.L.O.,” a once in a lifetime opportunity. Huang had the best talent, and he would use it to usher in a future where machines yet again think in addition to doing. Per Witt, the new software “can speak like a human, write a college essay, solve a tricky math problem, provide an expert medical diagnosis, and cohost a podcast.” And that’s just the early stuff.
To read Witt is to find out that the known with AI has pedestrian qualities relative to what remains unknown, and even better, what’s not yet been discovered. Markets are a look into the future, equity prices in particular, and Witt quotes Nvidia employee Bas Aarts as saying about what’s known only to those in the proverbial arena, that it’s of the “Wow, I cannot believe this is possible in this age” variety, that “people are oblivious about what is already going on. People have no clue” about the amazing leaps on the way.
There’s your inequality. It’s once again born of making the world so much better. The bet here is that AI will unearth human achievement that will render the poorest parts of Africa prosperous, all the while unearthing human flourishing globally that will make the abundant, incredibly prosperous, and – yes – beautifully unequal present appear Bangladesh poor and relatively wealth equal by comparison.
Which requires a pause. Just as everyone reads a different book, everyone reviews the same book differently. The review you’re about to read is being written by someone who hasn’t a faint clue about how to “back up” a computer, “upload” a medical form, or even connect a laptop to a printer. This is a way of saying in advance that while Witt’s book is heavy on technical jargon of the CPU, GPU, and CUDA variety, this review will largely focus on the presumed meaning of Nvidia’s ascent to one of the world’s most valuable and most important corporations.
Some will ask what’s the point of reading a review of a book about the man and corporation behind a remarkable technological leap since the writer is a technical dolt. It’s a fair point, but one that arguably misses the point. That’s because it doesn’t take someone technologically proficient to grasp the brilliance of work divided across as many hands, machines, and mechanical minds as possible. Even better, it’s not unreasonable to suggest that the mechanized technological genius that’s here and that’s on the way will render the technological illiterates of today the technologists of tomorrow exactly because so much formerly done by humans will be handled by machines. Time will tell.
For now, it’s best to start with gratitude for the opposite thinkers. Huang is clearly one of them. Witt writes in the book’s introduction that he’s setting out to tell “the story of a stubborn entrepreneur who pushed his radical vision for computing for thirty years.” As is frequently stated here, there’s no “majoring” in entrepreneurialism simply because no one chooses to be an entrepreneur. Entrepreneurialism is a state of mind, it’s a path certain individuals can’t not take.
It’s not planned because it’s not obvious. Furthermore, if it were obvious it wouldn’t be entrepreneurial, and it wouldn’t be entrepreneurial simply because what’s obvious is already being done. To show how outlandish Nvidia’s path was, just look at its valuation along with Huang’s net worth. Oh wow did they ever discover something amazing. Wealth inequality is the measure of this happy truth.
Huang as most know is of Taiwanese descent, though his parents (his father was a chemical engineer) sent him and a brother to a boarding school in Kentucky until they arrived. Not a fancy school (though it’s fancier now thanks to Huang), it’s believed Huang’s parents didn’t really know. They perhaps just guessed it was a school full of somebodys instead of one full of misfits. Which on its own is a story. What’s important is that where Huang went to school wasn’t going to matter, nor did it.
To his parents, and ultimately him, the U.S. was the goal. Good athletes in the capable sense just need to get here. That's the hard part, while the rest is elevation.
When he went off to college, it was Oregon State, not Stanford. In Huang’s words, “I just followed my best friend.” Really, what could school teach him? Genius can’t be taught, which is no insight. Genius was Huang’s parents getting his talents to the United States. The rest was his own invention. How to major in AI since AI wasn’t a thing, or at the very least wasn’t an operable thing when Huang was making his way up. Gaming was primitive. No need. Entrepreneurs create. They invent. Huang and the remarkable talent he cultivated would invent the future, not be told how to do it by some academic.
The individual in Bryan Catanzaro whose passion helped spark Huang’s own passion for AI was a humanities major. Hopefully conservatives convinced colleges and universities should pay the federal government back when their graduates don’t thrive read Witt’s book. One’s major is immaterial, particularly in an economy as dynamic as ours. Work that can be taught is, by definition, old news.
Out of Oregon State in 1984 Huang secured a job with Advanced Micro Devices. The pay was $28,700/year.
Following AMD, Huang went to work for LSI Logic. He rose quickly, and was running a $250 million division when he, along with Sun Microsystems employees Curtis Priem and Chris Malachowsky, decided to pursue what became Nvidia.
Even though they were told that the market for PC video gaming hardware was “crowded,” entrepreneurs have evangelical qualities about what they’re doing. Huang clearly did. The bullet-riddled Denny’s where they started making plans for Nvidia is open to this day, and is located at 2484 Berryessa Road in San Jose. Apparently Huang still goes there, and tips in the thousands after ordering all manner of menu items (including chicken fried steak – good taste!) that he eats part of.
Huang had even once worked at a Denny’s. Anecdote sucks, but Huang plainly had the immigrant desire to assimilate himself. Of course he did. Conservatives who should know better claim “we” can’t properly assimilate immigrants, but there would never be a need even if there were a real “we” beyond the rhetorical one. Implied in immigration is a desire to assimilate simply because immigration is an expression of a rich-person desire within some of the world’s poorest to increase the value of their labor.
Sequoia Ventures, arguably the world’s greatest VC, was along with Sutter Hill the early source of Nvidia’s funds. No less than Don Valentine (Cisco Systems, among others) told Huang after his subpar pitch that “Wilf Corrigan says I have to fund you, so you’re in business.” Corrigan was the head of LSI. He saw up close that Huang was going places.
Notable about Nvidia is that it started strong. It’s NV1 chip for video gamers sold in the 100,000 range, and as sales grew so did Nvidia. The problem was that the chips soon enough stopped selling, and their prices were lowered. In Huang’s words about Nvidia’s beginnings, “Every single decision we made was wrong.” Which is an economics lesson on its own.
Nvidia’s “recession” was the recovery, or would-be recovery from an NV1 error seen as “catastrophic.” When governments fight recessions they dampen the recovery. That’s the truth about the Great Depression that still eludes 99.99999% of economists.
So bad were things from 1993-96 at Nvidia that Valley graphics expert David Kirk made his contract employment there contingent on receiving a paper check every week. Stock options were not an option. Amid all this, including another failed chip, the sign went up at headquarters about how “our company is thirty days from going out of business.” Recessions signal recovery since recessions are when errors are fixed. Huang and Nvidia’s troubled times of old are useful as a way of exposing some of the many fallacies that inform “economics.”
For one, there’s a view that won’t die among economics types that the Fed funds rate is the cost of credit. Oh please. There wasn’t a number high enough in the mid-1990s to properly compensate a lender to Nvidia. Except that it wasn’t just that the loan market was shut to Nvidia. Equity finance was a non-starter too as evidenced by Kirk’s demand for actual paper checks over stock options.
It all asks readers to contemplate again why Huang is so rich today. It’s so easy to say that he’s that way because he’s a co-founder of Nvidia, but that’s the point. It’s too easy. The better explanation is that few initially wanted Nvidia’s shares. There’s quite simply no such thing as “easy money,” but since there isn’t, Huang almost certainly has more shares than he otherwise would. But that only tells part of the story.
In January of 1999 Nvidia went public, but consider the year: 1999. When did you, the reader, first hear of Nvidia? If you own shares, when did you buy them?
It’s worth asking mainly because of those who owned Nvidia shares in 1999, it’s no wild speculation that most don’t own those same shares today. It would have been too gut-wrenching for the average bear. Witt reports between the summer of 2001 and the fall of 2002, Nvidia shares lost 90 percent of their value. While Nvidia became part of the S&P 500 after replacing Enron in 2001, and while Huang was briefly a billionaire that same year, Witt notes that Nvidia’s shares didn’t just decline 90%, but that “It would fourteen years before he [Huang] saw that much money again.”
To have kept faith in Nvidia on the way to enjoying enormous upside, the typical shareholder would have held on through some rough times, including 2006 when Nvidia’s shares dropped 90 percent once again. Witt quotes one tech pundit as writing amid the 2006 collapse that “For a long time, we have wondered when Nvidia’s abject stupidity would have a price.” Stop and think about that. Or re-read it while thinking more about why Huang is so rich. He believed deeply when few did.
It's a reminder that with extraordinary wealth that spurs so much inequality handwringing, it’s perhaps better to think about the unseen. This is particularly useful with Nvidia since Huang, though known to have a “terrifying” temper, was also loathe to fire employees. He yelled them into line with his “cathartic” rants that he was very public about. Ok, but how many left after a tirade, or another dip in Nvidia shares, or because a “better” company like Intel offered them a job. How many individuals walked away from billion-dollar fortunes?
One long-time Nvidia employee, Dwight Dierks, sold some shares right after the float in 1999 to buy his father a car. When he describes it as a “billion-dollar-Cadillac” in joking fashion, there’s serious reality underlying the joke. Since going public in 1999, Nvidia shares are up 300,000%. Which is the point, but perhaps not the expected point.
It’s so easy to say how foolish it was for early Nvidia employees to leave, or for co-founder Curtis Priem to sell all of his founders shares for hundreds of millions from 2004-2006 when they would be worth $100 billion+ today, but as evidenced by how much departed employees and founders like Priem could have made, Huang pulled off the miracle of miracles. There’s no other way to explain the would have, could have been fortunes. Naturally individuals sold, quit, left, came back, left again, and all sorts of actions unrelated to amassing shares that, by virtue of their miraculous climb, were never supposed to climb this high.
In Huang’s case, and amid the occasional costly stumble rooted in his relentless investment “in speculative technologies that would either revolutionize computing or flop,” the future was far from certain. Exactly because any successful chip was immediately a target for imitation, Huang had to keep seeing ahead. In his own words, “If we don’t reinvent computer graphics, if we don’t reinvent ourselves, and we don’t open the canvas for the things we can do,” we “will be commoditized out of business.” Imagine the pressure of having to constantly see around the proverbial corner, while knowing full well the brutal market reaction that awaits after the inevitable mistake.
It all speaks to the entrepreneurial mind. And this includes people like Elizabeth Holmes who tragically sit in jail for believing so deeply in her vision. Never asked by Holmes’s myriad critics is why, assuming she knew she was committing fraud, she never sold any of her shares. There was a huge and very liquid market for private shares, yet Holmes held on to hers. It’s that state of mind thing again. It’s why the unequal are unequal. Huang held on, and did so through incredibly difficult stretches that included him fighting to keep his job. The unequal believe in themselves, and they don’t lose faith as others are expressing their lack of faith through share sales. Again, much more interesting than the billionaires minted by Nvidia would be those that weren’t because they sold, exited, or both too soon.
Which was once again logical. Witt is clear that earnings represented to Huang the chance to try something new. Rather than pivot toward the obvious, Huang was routinely searching for the “zero-billion-dollar market,” as in the market that’s not yet there. Is it any wonder that truly entrepreneurial visions are rejected by nearly everyone, and most notably the existing commercial successes? Writing about Huang, Witt indicates that he seemed to find comfort in markets “that only he would participate in – one that only he would even see.”
Huang indicates to this day that Intel isn’t nor was it ever a competitor. Nvidia was yet again in search of the "zero-billion-dollar” markets that Intel wasn’t interested in, but that Huang was. And is. He saw things and sees things.
Very interesting in consideration of Nvidia’s wild ride was the talent that Huang lured into the fold. Keep in mind that this is Silicon Valley, which means the most ferocious battle of all is the ongoing one for talent. Yet despite Nvidia’s scary ups and downs on the way to arguably the world’s most important company, “Nvidia employees were unabashedly elitist.” It was a known that Nvidia was full of talent, which says so much about Huang. With sharks all around him trying to poach the most crucial capital of all, Huang managed to maintain a very deep bench. Terrifying as he could be, Witt notes that almost all of existing and former Nvidia employees interviewed by him “had a tender story about Huang to relate.” What fun it would be to follow him around for a day, or days. And not just Huang.
A company this dynamic is logically dense with characters. There’s Bill Dally, who dropped out of high school given an aversion to sitting through history class, became an auto mechanic, but ultimately got a BA at Virginia Tech, a Masters from Stanford, and a PhD from MIT. Dally had options, including Intel, and was told he was “crazy” for not taking an Intel job that surely more than a few Nvidians (?) did.
There was John Nickolls who always yelled instead of talking. Witt notes that Nickolls “had no interest in video games at all,” but was passionate about proving Moore’s Law wrong. He did. Nvidia did. Cancer ultimately took Nickolls way too early, but to read about him was to want to ask him (or others in the Nvidia orbit) if he felt the mechanization of knowledge and thought might eventually save him from the cancer that existing medicine couldn’t.
Alex Krizhevsky didn’t work for Nvidia, but he trained neural nets to see and think. By his actions, this soft-spoken Ukrainian immigrant conveyed to all manner of academics that (per Witt) “they had so far wasted their careers” on beliefs that Krizhevsky discredited with a week’s worth of “training,” and enormous amounts of electricity used in the training. Krizshevsky oversaw a giant leap, one indicating that “if you could teach computers to see, you could teach them everything.”
Of course, in writing what was just written it’s almost certain that all sorts of things were mangled in the telling. But that’s surely the point. And it’s why the AI advances elicit so much optimism in your reviewer.
AI isn’t a threat to humanity, rather it will elevate humankind in ways too grand to describe. Is there a way to say how it will reveal its genius? Certainly not. If there was a way to know what’s ahead, then those who know would be billionaires for knowing.
All we know, or should know, is that it will be brilliant. Think of the pin factory Adam Smith entered in the 18th century. One man working alone could maybe produce one pin per day, but several men working together could produce tens of thousands. Multiply the latter hundreds of thousands of times over with AI as technology thinks and does as though thousands and millions of brilliantly skilled individuals are at work. Oh, to be born today. Or fifty years from now.
Surprisingly, Witt expresses fear throughout the book that AI could have evil qualities, but in the words of Huang, “In order to be a creature, you have to be conscious.” AI will lift humans to the best versions of themselves precisely because it will enable specialization that will have more and more of us working all the time not because we have to, but because we want to.
In a very real sense, the AI that has been authored by the remarkable chips created by Nvidia will turn the world’s workers into Nvidia’s workers. Which will be a beautiful sight to see. Impossibly rich as Nvidia employees are, they’re still coming to work. They can’t not. They work all the time because they love what they’re doing, and they’re in love with discovery.
It cannot be stressed enough that progress in life is defined not by what we’re doing and learning, but by what we no longer need to do and learn. That’s why Huang is so right that AI “can only lead to good.” It’s so true.
About halfway through The Thinking Machine, Witt quotes AI visionary (and oddly enough, skeptic) Geoffrey Hinton talking about “the uncontainable thrill of sneak previewing embargoed AI technology that would shock the world once unveiled.” Yes! Once again, to be young. There’s no limit to human advance, particularly as production is spread across a soaring number of hands, machines, and thought that’s mechanized. Read Stephen Witt’s book to get a sense of what’s ahead as wealth inequality soars thanks to human flourishing going skyward, but almost as fun, sit back and watch. We haven’t seen this movie before, but it will make all that we’ve seen before seem incredibly bland by comparison.
