Strachman’s and Gibson’s modest goal for 1517 was to help a new generation of technologists reshape the incumbent system of the world. Birthed in the fogs of Stanford, hosted in the clouds of Google, and aiming at superhuman artificial intelligence, that system exalts academic honors, all-purpose Turing machines, software supremacy, open-source modularity, and secret-sauce integration. But do you really trust this system to handle your money, motivate your children, shape your worldview, or maneuver your car through a crowded parking lot?
The bearded libertarian Gibson was working at Thiel’s Clarium Capital in 2010 when he learned that his boss was launching an “anti–Rhodes Scholarship.” “Sign me up,” Gibson said. Post-graduate fellowships for Americans and other colonials to get polished at Oxford, Rhodes Scholarships are the ultimate establishment honor. They tend to produce politicians like Bill Bradley, media celebrities such as Rachel Maddow, and academics rather than entrepreneurs.
Strachman, an advocate of “project-based learning,” had previously started a charter school in San Diego. She already believed that many students could learn far more starting their own companies than sitting in classrooms. When she heard about the Thiel Fellowship, originally called “20 under 20,” she joined Gibson there. They picked their first group of twenty under twenty in 2011. The next year an eighteen-year-old freshman at Canada’s University of Waterloo, Vitalik Buterin, applied.
The Thiel team found Buterin’s project, an innovation in digital education, unimpressive. “At the outset,” says Strachman, “we were seeing lots of education projects; his didn’t stand out.” They passed him over, and he didn’t reapply. But that year they did choose his high school classmate and friend Chris Olah, a nineteen-year-old graphics virtuoso.
In 2013, Olah brought Buterin, just shy of twenty, to the big “Hack the North” software jam session, held every September at Waterloo, where he met Strachman and Gibson. In his upward-inflecting, faintly lisping tones, Buterin told them that since his application, he had changed course a bit, having fallen in love with “this cryptographic construction called a blockchain.” He had devoted himself to studying its possibilities, using the bitcoins he’d earned from founding Bitcoin Magazine to travel the world from Israel to Las Vegas to Amsterdam, talking with other blockchain obsessives. Realizing that “crypto projects were taking up thirty hours per week of my time,” he dropped out of university in April 2013.2
In the Waterloo meeting with Strachman and Gibson, his last chance to get their attention before he transgressed into his over-the-hill twenties, Buterin proposed to revolutionize the Internet and the global financial system. “That’s what Peter Thiel wanted, right?” As he recited a litany of ambitiously cockeyed schemes—“Turing-complete” blockchains, new software languages, currencies, computer platforms, smart contracts—Strachman and Gibson could see that he was a genius. But his grandiosity and apparent lack of focus defied all the rules of successful enterprise.
They decided to support him anyway. In November 2013, Buterin wrote the Ethereum white paper, and on June 5, 2014, Peter Thiel announced a new group of twenty Thiel Fellows, which included Buterin. A year later, Ethereum went live, with the announcement, “What bitcoin does for payments, Ethereum does for anything that can be programmed.” It was another step in the decentralization of the Internet.
Just as Ethereum was entering the larger world, in July 2015, Strachman and Gibson were leaving the Thiel Fellowship to start a new but related project, the 1517 Fund, which would invest in Thiel Fellows and other high school and college-aged company founders.
The fund’s name alludes to another historic decentralization, launched on October 31, 1517. That was the day that Martin Luther posted his Ninety-five Theses on the church door at Wittenberg. Among the abuses Luther was protesting was the selling of indulgences. The remission of the temporal penalty due to sins, an indulgence, like other spiritual goods, must not be sold. The pardoners who perpetrated this abuse would issue a document memorializing the transaction. The 1517 Fund explains the parallel: “Likewise, universities today are selling a piece of paper at great cost and telling people that buying it is the only way they can save their souls. Universities call it a diploma, and they’re making a fortune doing it. Call us heretical if you like, but the 1517 Fund is dedicated to dispelling that paper illusion.”3
The Thiel Fellowships and the 1517 Fund are protesting the layers on layers of government grants that impose a stifling conformity on our universities through the indoctrination there of a single system of the world. Above all, they denounce the horrendous debt load of more than $1.5 trillion—roughly 7 percent of the U.S. gross domestic product—heaped on the hapless American student to pay for a bloated academic establishment, debt that has driven whole generations out of the entrepreneurial economy that enriched their forebears and endowed the universities themselves.
Traveling in early 2017 to the fund’s annual meeting in Silicon Valley, I Ubered through the brindled meadows of Portola Valley and far up the Old Spanish Trail to the locked gate of Number 495. This is the site of Jacques Littlefield’s famous five-hundred-acre Pony Tracks Ranch. When Littlefield, the heir to a construction fortune, died in 2014, his collection of 220 rare and exotic tanks—“enough mechanized weaponry to invade a small country” as the San Jose Mercury News reported at the time—was auctioned off.4 What remained at his ranch was the sensational view, overlooking all of Silicon Valley, from Intel in Santa Clara to Google in Mountain View. On clear days you can glimpse the towers of San Francisco.
In the future, the ranch may become better known as the site of the early headquarters of Luminar, the company begun in 2012 by Austin Russell, a wunderkind engineer and 2013 Thiel Fellow. At the time of the 1517 Fund’s annual meeting in 2017, Luminar had 250 employees in Silicon Valley and Orlando, Florida, and was adding a new chip design facility in Colorado. About to emerge from five years of “deep stealth” development, it was announcing unique new lidar (laser-radar) chips, orders of magnitude better than the competition from Google and others, which provide the core vision capability for self-driving cars.
1517 chose Pony Tracks for its annual meeting to exhibit Russell, who started Luminar at age seventeen and ranks with Buterin as the Thiel Fellow most capable of launching a reformation. While Buterin is unleashing a new global architecture for computing and finance, Russell is launching a revival of U.S. innovation in manufacturing hardware. Defying the frenzied hype of its software-based rivals in the self-driving sweepstakes, from Google to Tesla, Luminar’s product is comparable in its transformative reach and ambition to Henry Ford’s early-twentieth-century assembly lines at River Rouge in Detroit. Software will not end up eating everything in Silicon Valley.
The Luminar story begins in Newport Beach, California, with the precocious twelve-year-old Russell, whose parents were denying him a mobile phone. He responded by transforming his Nintendo gaming portable into a WiFi phone and followed up with inventions and patent applications in optics and holography. His chief setback seems to have been losing a national robotics competition as the fifteen-year-old leader of his high school team.
Buterin too was a child prodigy on a Mozartian scale. At four his favorite toy was an Excel spreadsheet. At seven, he instructed himself in Mandarin and today fluently debates in the language during trips to Shenzhen. Both the Californian and the Canadian benefited from early guidance from their fathers. Dmitry Buterin introduced his son to bitcoin, and Robert Russell, now CFO of Luminar, pushed Austin ahead in optics. Like Buterin, Russell is not cowed by the powerful forces arrayed against him.
Looking at Google’s Waymo, Elon Musk’s Tesla, Levandowsky’s Uber scheme, and the host of imitator “autonomy” projects, Russell sees over-hyped schemes that cannot ultimately succeed. They can sue each other all they want, but the approach of the most highly vaunted ventures is to try to enhance the existing lidar, radar, and camera systems with artificial intelligence, big data, mapping, and software. Russell, who spent
his boyhood preoccupied with optics, lasers, and machine vision, understands that no amount of big data can make up for bad data from hopelessly inadequate vision systems.
Lidar (light detection and ranging) is how vehicles see and evaluate the world around them. Russell understood from the outset that if the car cannot deliver an utterly reliable real-time image and interpretation of the road ahead for at least two hundred meters in the dark, with just 10 percent reflectivity, it is a death-trap. Two hundred meters give a car seven seconds to react, compared with the one or two seconds afforded by existing systems, if they can register the objects at all.
Such systems are worse than useless, Russell thinks, because they produce an illusion of success by meeting the low-cost targets and performance specs hyped by the industry. Like Google’s Waymo vehicles, they could perform adequately on repetitive twenty-five-mile-an-hour commutes from Silicon Valley to San Francisco. Answering Elon Musk’s complaints about regulation, Russell said prophetically in early 2017, “The real threat is premature launch of systems that end up killing people and killing the industry.” He added, “For existing technology, it may even be underregulated.”
To Russell, it is obvious that autonomy is not chiefly a software problem. Most existing software will have to be changed anyway to adapt to faster and better lidar systems. Much of the work already done by Waymo and others will have to be redone. Hardware is the gating factor. Nearly all of Russell’s competitors use existing devices—all of them inadequate—from companies such as Velodyne and Quanergy.
The hardware for self-driving cars has hardly improved in twelve years since the original DARPA competition that launched the industry. These approaches still entail a “paint-can” on the roof with sixty-four lasers and photo-detectors in rows. Velodyne’s idea of how to improve performance is to double the number of lasers to 128 and add more software. It all strikes Russell as a kludge—a system too complex and cumbersome to work.
Governing this challenge is Clayton Christensen’s model of “integration and modularity.” When a product fundamentally underperforms the market’s requirements, integration is essential. You can’t just go down to Fry’s and cobble together existing devices. Every interface must be optimized. Modularity—interconnecting standard components from a variety of suppliers—works only when the product easily fulfills the function, leaving tolerances for less than top performance.
To Russell, it is obvious that current self-driving vision systems are nowhere near the level the market requires. As bad as human drivers are, the self-driving systems of 2017 had to be turned off or augmented far more often than human drivers failed. The best record by far is Google’s Waymo, but it has to be disengaged every 5,500 miles or so, compared with nearly 500,000 miles for human drivers between accidents and 84.6 million miles per fatality. Tesla has to be turned off every three miles. Sure, these systems will get better with time. But they’re nowhere close to where they need to be. Although you would never know it by listening to the promotional literature, the existing vision systems cannot collect enough good data fast enough. Fancy software just gilds that rotating sow’s ear on the roof.
In 2012, the seventeen-year-old Russell resolved to start from scratch and build entirely new integrated systems with at least fiftyfold more resolution and tenfold more range than the prevailing standard. Russell’s technology targets were not “inevitable.” They did not spring from Darwinian trends in the Valley. It would require, in Russell’s words, relentless “top-down planning and bottom-up engineering.”
To Russell, nothing matters but performance. There is no long-run demand for a system that kills people. Aim at performance, and low cost will follow. Aim at low cost, and you will not achieve sufficient performance to have an enduring business. After a sufficient system is devised, demand will foster economies of scale and learning curves that bring the price down over time.
Doing research at Beckman Laser Institute while joining with the photonic genius Jason Eichenholz of Open Photonics to fill out his company, Russell eventually raised $36 million from 1517 and other investors. Throwing out the conventional wisdom about lidar, his team researched two thousand different ways to build a fundamentally more effective system. In 2013, they decided to make a drastic change in the prevailing practice.
A crucial decision was to reject a wavelength of light preferred by his rivals—905 nanometers—which is close to the visible spectrum and can simulate what can be seen by human eyes. Unfortunately, the 905 nanometer near-infrared also can affect human eyes. So Russell moved deeper into the infrared band to 1,550 nanometers (153 terahertz), which is employed in fiber optics. The longer wavelength can be used at sixtyfold greater power levels without endangering human beings. Russell and the team also radically changed the architecture. Rather than using scores of lasers, they raster-scan the scene thousands of times a second with a single laser, like the device in a cathode-ray tube television.
Rather than separating the vision systems from the interpretive processing, they integrated them, as human vision systems do. Throwing out silicon and adopting an exotic high-performance alloy, indium gallium arsenide, for their chips allowed them to implant the lasers directly on the same substrate as the computational element.
All the other companies link their sensor systems and their processor technology with three-thousand-dollar analog-to-digital converters. Luminar produced a microchip “eye” coupled to the processor that resolves the image, sampling the road hundreds of thousands of times a second. Now on its seventh-generation application-specific integrated circuit (ASIC) design, Luminar has a fully integrated technology stack. It is manufactured exclusively by the company and its specialist foundry partner, TowerJazz Semiconductor. Headquartered in Israel and headed by the visionary entrepreneur Russell Ellwanger, TowerJazz happens to run a world-beating analog-digital fabrication facility (a “fab”) in Austin Russell’s hometown of Newport Beach. After making a few score working systems that outperformed existing lidar roughly fiftyfold, Luminar committed to producing ten thousand units of working systems in 2018.
1517’s Strachman stresses the importance of cultivating an entrepreneurial community. After becoming a Thiel Fellow in 2013, Russell found his community. He moved into 216 Park Lane, a stucco mansion on a palmy drive in Atherton, a few miles north of Stanford. While real estate agents tried to sell the property for ten million dollars in the wake of the “Great Recession,” the eighteen-year-old Californian rented the secluded space with two other youths engaged with the Thiel program—the twenty-three-year-old Vermonter Stephen Balaban, a graduate of the University of Michigan, and the seventeen-year-old Thomas Sohmers, from Massachusetts. The trio looked an entrepreneurial version of the Hardy Boys, heroes of the classic mystery books—with Russell as the tall, ping-pong-playing, blond Joe Hardy; Balaban as his dark-haired intellectual martial-artist older brother, Frank Hardy; and Sohmers as Chet Morton, their slightly chubby friend whose tinkering and inventions often save the day.
Sohmers, like Russell, was a 2013 Thiel Fellow, but while Russell had made it to Stanford before dropping out, Sohmers had left his science-focused high school as a junior. Thanks to his silicon-trap mind, though, he was soon at Stanford anyway, not as a student but as an expert brought in to teach Russell’s ex-classmates. He had started his REX Computing company at sixteen, aiming at overthrowing prevailing chip architectures that waste most of their energy in “wait-states” and wires connecting to memory.
Disgusted with the departure of silicon from Silicon Valley, Sohmers saw many paths for reviving the industry. He focused on inventions that moved the prevailing chip designs toward the zero-power end of Tredennick’s “leading-edge-wedge.” Low power is imperative in an age of mobility serving the mobility of human minds and bodies.
Balaban, the oldest member of the trio, was still almost as precocious as Russell and Sohmers. Like Buterin, he was fluent in Mandarin and software. But he missed out on a possible Thiel Fellowship when he went to Bei
jing instead to start a “Y-Combinator” incubator fund.
Befriending Danielle and Mike, then still with the Thiel Fellowship, Balaban became the adult supervisor of the younger fellows while working on a wearable, hands-free camera embedded in a baseball hat.
The three young men often talked technology and libertarian philosophy late into the Atherton nights and worked on their companies nearly all other hours. All of them shared and re-enforced each other’s frustration with Silicon Valley’s software obsession and abandonment of manufacturing.
And all of them soon came to see this abandonment as a gigantic opportunity.
CHAPTER 11
The Heist
In January 2009, Satoshi Nakamoto announced “the first release of Bitcoin, a new electronic cash system that uses a peer-to-peer network to prevent double-spending. It’s completely decentralized with no server or central authority.” He went on to specify:
Total circulation will be 21,000,000 coins. It’ll be distributed to network nodes when they make blocks, with the amount cut in half every 4 years.
first 4 years: 10,500,000 coins
next 4 years: 5,250,000 coins
next 4 years: 2,625,000 coins
next 4 years: 1,312,500 coins
etc . . .
Hal Finney, cryptographer supreme, posted his congratulations and offered these thoughts:1
It’s interesting that the system can be configured to only allow a certain maximum number of coins ever to be generated. I guess the idea is that the amount of work needed to generate a new coin will become more difficult as time goes on. . . .
As an amusing thought experiment, imagine that Bitcoin is successful and becomes the dominant payment system in use throughout the world. Then the total value of the currency should be equal to the total value of all the wealth in the world. Current estimates of total worldwide household wealth that I have found range from $100 trillion to $300 trillion. With 20 million coins, that gives each coin a value of about $10 million.
Life After Google Page 12