Killfile (4 page)

Eli Preston is not on the Forbes 500 list, but everyone in the financial press says it's only a matter of time. He's the founder of OmniVore Technology, a small, privately held company, which is expected to make him a billionaire once it goes public.

Preston is supposed to be the next Zuckerberg, or at least one of the candidates: smart, ambitious, and insanely rich; lives on bulletproof coffee, energy drinks, and high-protein sushi prepared by a personal chef; lectures about the future at TED one week, then shows up courtside at a Knicks game the next.

He just turned twenty-six. I know this because he rented out the entire Bellagio in Las Vegas for the party. I read about it on Gawker.

I'm not, however, completely clear on what exactly OmniVore does, or why it's worth so much money. So I ask.

For a brief instant, I get a slight weariness from Sloan. His IQ scrapes the limits of the tests made to measure it. He's constantly explaining things, stopping his own personal train of thought and waiting for everyone else to catch up. To him, the rest of us move in slow motion, taking forever to understand what is clear to him in an instant. For a moment, I sense how tiring it must be to be so much smarter than everyone else. To know the answers so long before everyone else does. To see clearly while they're still blindly groping around in the dark.

But he explains anyway.

“OmniVore is what we call a data-miner,” he says. “It sifts through massive amounts of information for its clients. You see, most companies are drowning in facts. Thanks to the Internet and cheap digital storage, they have access to incredible levels of detail now—they can track down to the minute the last time you purchased razor blades at the store, and it goes into a little file that contains every other fact about you. Your
credit-card number. Your birth date. What car you drive. Your sexual preference. Your favorite flavor of ice cream. Everything that might be considered even remotely relevant to extracting another couple of dollars from your wallet. And they have that for everyone. They can track the number of left-handed divorced women who visit their website, or gauge the reaction on Twitter to their latest ad campaign from Denver Broncos fans or vegetarians or even rape victims, if they want to go that far. But it's all too much. They have more raw data than ever, and they actually know less. It's like drinking from a fire hose, as they say. They can't narrow it all down. That's where OmniVore comes in. It applies its software to the data, and it finds out exactly what the companies actually need to know. It identifies problems before they happen. It learns the patterns inherent in a business, and makes predictions for the future. It discovers threats and eliminates them.”

“Threats? Like what?”

“It depends. For a tech company, it could be hackers or industrial espionage, someone trying to steal next quarter's product designs. For a retailer, credit-card fraud or employees stealing merchandise. For a bank, embezzlement or tax fraud. Could be anything. The idea is that OmniVore's software is smarter than humans. It's like chaos theory in reverse: we see a hurricane, but they know where and when a butterfly flapped its wings.”

He's dumbed it down significantly for me, but it's close enough.

“And that's profitable?”

“God, yes. Eli is turning away the biggest companies in the world. He's got government agencies, automakers, studios, TV networks—everyone you can think of, really—lined up and waiting on him. Unlike Twitter or Facebook, there are people with deep pockets willing to pay a great deal for this service.”

“So what's your problem with him?”

“Not many people know this, but Eli used to work for me.”

“That doesn't seem like something he'd be shy about. I imagine it would look good on his résumé.”

“We didn't part on the best of terms. I recruited him out of Harvard as an analyst. I thought Eli would be perfect for the job. As it turned out, he got bored very quickly. He didn't care for sitting in an office and crunching numbers. He wanted to be out in front, playing at the high-stakes tables.”

Right. Where the hookers and free drinks are. It wouldn't be the first time a young guy thought he could do better than his boss. “You fired him.”

“We came to a mutual understanding. A few months later, he had raised enough capital to start OmniVore. And his success story began.”

“But you don't think he did it himself.”

“You really must be psychic,” he says, smiling. I hear that joke a lot. I smile along with him anyway. “Yes. He's built his company—everything he's done since he left—on the strength of my ideas, my intellectual property.”

Something about the timeline seems off. I interrupt Sloan. “Let me ask you a question here: He hacked your files two years ago, and you're just finding out now?”

“Do you understand what I do, Mr. Smith?”

He's right. I don't. I'm not ashamed to admit it. “Not even a little.”

“You invest in the stock market?”

“Not really.”

That surprises him. “Where do you keep your money?”

“In my wallet, mostly.” He laughs at that, but it's true. My fees are high, but so are my expenses. I also don't trust money I can't pull out and spend whenever I want. I know it's not very smart, but old habits die hard.

“Well, if this all works out, maybe we can set you up with a starter portfolio,” he says. At the same time, in his head, he prepares to give me the Fisher-Price explanation of his job, the one he uses for people he meets at parties.

“Stock picking used to be about people making educated guesses based on the companies,” he says. “Profit and loss statements, supply and demand, market conditions—but really, nothing more scientific than throwing darts at a board with names on it. Most investment firms are lucky if they do as well as an index fund—that's a collection of stocks that simply mirrors the market, goes up when it goes up, goes down when it goes down. What I do is different. Have you ever heard of algorithmic trading?”

You don't work for rich clients without picking something up. “You use computers to analyze the market and make stock trades for you.”

“Close enough,” he says. “I invented it, more or less. I created an algorithm—that's a mathematical formula that you enter into a computer—that analyzes data. Essentially, it was a way to look at any set of data and organize it, and even make predictions based on it. It could find the underlying patterns in the numbers, when a human being would see only a mass of random information. This was something I discovered when I was looking into Sienkiewicz-Moore theorems, a subset of Big Number theory, back when I was still in graduate school.”

All I get from that is the ice wall again. He sees he's lost me and tacks quickly back into simple concepts. “So I take a series of facts, translate them into numbers, plug them into the formula, and it makes a series of educated guesses about the future. What made the algorithm so interesting was that the predictions were almost always right. I could take any facts that could be reduced to mathematical inputs—migration patterns of birds, or tide tables, or annual rainfall in the Gobi Desert, for instance—and I would get a very good idea of how those same
numbers would turn out in the future. Then it occurred to me, what if I entered something a little more concrete than rainfall estimates?”

I'm starting to get it now. “Like, say, stock prices.”

He smiles again, and I sense genuine pride still lingering there. “Exactly. I quickly found the algorithm was a great deal more valuable to me in practice than as a theory published in an academic paper. That's how I began trading. I'd use the algorithm to predict the rise and fall of the market, and I'd make bets accordingly.”

“And you got very rich.”

Another small burst of pride. “Yes. Other people noticed. They hired their own computer engineers and math professors. Now there's a whole industry of traders and programmers analyzing market data using algorithms and computers. Each firm has something they call their ‘secret sauce.' That's a proprietary algorithm that's the heart of their trading. It tells their computers how to interact with the markets. I called mine Spike. To find the spikes in the markets.”

He looks at me. I smile, to show him I get it. “Your secret sauce is better.”

“Much, much better. Not to boast, Mr. Smith, but nobody else has come close to understanding what I did when I broke that problem back in graduate school. And we're constantly refining the process, feeding more data to the algorithms. Spike, like every other piece of trading software, makes millions of decisions every second. Literally billions of dollars every minute, all moved around by computers. That requires incredibly smart people to analyze market trends, to see risks and opportunities and then translate them into the kind of math that machines can understand. Everyone is trying to beat the odds, trying to get their computers to think a little faster, a little smarter. It's not easy. As I said, most firms are lucky to match the market. The best ones can offer you perhaps a ten percent return over time.”

“What's your return?” I ask.

“Eighty to ninety percent,” he says. “Even when the economy collapsed, we managed to make a profit. All with Spike. It's simply smarter than anything anyone else can come up with. Other people will promise you pennies. We double your money, or close to it.”

Bullshit. That's Ponzi scheme territory. Nobody can guarantee that kind of return. I don't get any active deception off Sloan, but people have a habit of buying into their own hype. After all, it's not a lie if you really believe it.

Some of my skepticism must show up on my face, because Sloan smiles and asks, “You don't believe me?”

I try to be diplomatic. “That's quite a return,” I say. “Warren Buffett only manages nineteen percent, and he's supposed to be the most successful stock picker in history.”

Sloan smiles again. “Warren's a friend. But he's a very public figure. People follow him. They jump into his stocks when he buys and run away when he sells. We don't allow that. We keep our trading secret. We've got a dedicated dark pool that hides our trades, and we spread them over a variety of market makers. And we control our overall investment. I could have a half a trillion dollars in assets under management if I wanted. I've got the clients lined up outside my door. But then we'd be big enough to tip the market. People would see our trades move the prices. We'd have information leakage, and we'd lose our advantage. I don't need that. That's why we've kept the heart of Spike a secret ever since I invented it.”

“And you think Preston stole it from you.”

Some anger finally slips past the ice wall.

“He
did
steal it. He's taken the knowledge I've spent fifty years building, and is now getting rich from it.”

“I thought you said he was in data mining. Not stock picking.”

“It's all numbers,” Sloan says. “The heart of Eli's business is a piece of software called Cutter. It analyzes the data for him. But the engine that drives Cutter is the same one that drives Spike. The one I built. It can be used on any set of facts, provided you can reduce those facts to computer input. Phone calls, credit-card purchases, social-media posts—my algorithm can find the patterns hidden in all of it. Everything OmniVore does, it's doing on the back of my work. That's the genius of my discovery.”

“Again, not to boast or anything.”

Sloan waves that off. He doesn't have time for false modesty.

“Still doesn't explain why you've waited two years to go after him,” I say.

“If I knew he'd stolen from me, I certainly would have done something about it sooner. I thought my security was adequate. There are no records of any breach. The software behind Spike is located inside secure computers that would have recorded any attempt to hack them. Access is strictly controlled. Every email is monitored. My analysts walk through a scanner and a strong magnetic field on their way into and out of the office, which means I would know if they carried any thumb drives or disks from the office, or they would be wiped clean if I didn't. Then, about six months ago, I received information from one of Eli's former employees. Someone upset with his own pay package, of course. He told me that he'd heard Eli bragging to a client that his computer models were better than Spike—that he'd improved them.”

“That doesn't prove he stole them from you.”

“No. But Eli's security isn't as thorough as mine. This informant showed me several blocks of the source code Eli uses for Cutter. With a few changes here and there, it's the same as Spike. No question.”

He takes a manila folder and places it on the table. He opens it and shows two different printouts. One is marked
SPIKE
and the other is marked
CUTTER
. They look identical to me, but only because I can't make sense of either of them. Sloan seems pretty convinced, however.

“So if he didn't download it from you, how was he supposed to steal it? Did he break your security system? Pay off one of your other employees?”

Sloan shakes his head. “No. Again, I would have noticed any breach. I have redundant systems and loyal, well-paid people. He might have been able to corrupt one or two, but not all of them. It would have shown up.”

“Then how did he do it?”

“He took my ideas out of here the old-fashioned way—in his brain.”

That staggers me for a moment. “Is that even possible?”

Sloan radiates ironclad certainty. There's no doubt in him. “He's smart enough, yes. He read the underlying computer code of Spike, line by line, until he found the algorithm, and then he memorized it. He walked out of here and wrote his own version, and then used it to create Cutter.”