Point A: March 1999, a converted warehouse on Emerson Street in Palo Alto, the kind of space that still smelled faintly of the printing press that had occupied it during the Eisenhower administration. Two people in the room and a whiteboard that covered the entire east wall. The whiteboard was covered in diagrams that looked like fever dreams drawn in dry-erase marker: circles within circles, arrows pointing in directions that did not correspond to any known geography, words like SCALE and NETWORK EFFECT and PLATFORM written in capitals and underlined three times.

Diana Khatri was twenty-seven years old and had not slept more than four hours a night in six weeks. She had the look of someone who was burning fuel that her body had not yet manufactured, drawing on reserves that she did not know she possessed and had no way of replenishing. Her co-founder, Marcus Wei, was twenty-nine and had the opposite problem: he was calm in a way that suggested he had already run the simulation of every possible outcome and accepted all of them.

The company was called Lumina. The idea was simple, which in Silicon Valley in 1999 meant it was either worthless or worth a billion dollars, and the difference between those two states was entirely a function of who believed in it. Lumina would build a platform that connected small businesses to their customers through the internet. Not e-commerce, which was already becoming crowded. Not advertising, which was what portals did. Something new, something that would let a bakery in Des Moines know that one of its regular customers had moved to Phoenix and was craving the same sourdough they'd bought every Tuesday for three years.

"It's not a technology problem," Diana said, staring at the whiteboard. "It's an empathy problem. We have all this data now, all these connections, but nobody is using them to make people feel known. The internet should feel like walking into a restaurant where the waiter remembers your name. Instead it feels like walking into a warehouse where everything is labeled and nothing is personal."

"Venture capitalists don't invest in empathy," Marcus said.

"They invest in returns. Empathy is the mechanism. The return is the result."

Point A-Prime: October 2001, the same warehouse, now empty except for a single desk and a chair and Diana Khatri sitting in the chair staring at a computer screen that showed a spreadsheet with more red than a butcher's floor. The warehouse had been emptied over the previous three weeks: first the Aeron chairs, then the foosball table, then the twenty-three iMacs in Bondi blue, then the espresso machine that had cost four thousand dollars and produced coffee that tasted like victory even when victory was still hypothetical. All of it sold, piece by piece, to pay the severance packages of forty-seven employees who had believed in something that had turned out to be, in the end, just another set of circles on a whiteboard.

Lumina had raised eighteen million dollars. Eighteen million dollars in three rounds, from firms whose names appeared regularly in the Wall Street Journal and whose partners were photographed at charity galas with governors and movie stars. Eighteen million dollars had bought two years of growth and forty-seven employees and a client list that included three hundred small businesses in twelve states and a technology platform that actually worked, that actually made people feel known, that actually connected the bakery in Des Moines to its scattered customers in a way that was beautiful and useless.

Beautiful and useless, because eighteen million dollars was not enough to reach profitability before the market stopped believing in beautiful things that were not yet profitable. The Nasdaq had peaked in March of 2000, six months after Lumina's Series B, and then it had done what markets do when they discover that they have been pricing dreams as if dreams were factories: it had corrected. The correction had been violent and total, erasing five trillion dollars of value from companies that had been valued not on revenue or profit but on belief, on momentum, on the shared hallucination that the old rules no longer applied.

Diana Khatri was not looking at her spreadsheet. She was looking at a framed photograph that sat on her otherwise empty desk. The photograph showed the Lumina team on the day they had moved into the warehouse, forty-nine people crammed into a single frame, holding bottles of cheap champagne and grinning with the particular radiance of people who believe they are standing at the beginning of something that will change the world. Two years later, forty-seven of them had been laid off, one had gone to work for Google, and one had been hired by a venture capital firm that had once courted Lumina for its portfolio and was now acquiring the company's assets for three cents on the dollar.

Vector Between: Diana at twenty-seven on Emerson Street versus Diana at twenty-nine on the same street in the same building that had been stripped of everything except the ghost of the idea that had once lived there. The Diana of 1999 believed that technology could make the world more human. The Diana of 2001 believed that she had been naive, that the world was not interested in being made more human, that what the world wanted was efficiency and scale and the frictionless extraction of value, and that empathy was a rounding error in the great accounting of capitalism.

The question was not what had happened. The question was: at what point between these two Dianas had the transformation occurred, and was the transformation a progression or a regression, and could she trace the exact vector that had carried her from one state to the other?

Point B: December 1999, the offices of Garrison Meriwether Ventures on Sand Hill Road, the Vatican of venture capital, the holy street where the future was consecrated and funded. Diana and Marcus had been summoned to discuss their Series B. The summons had come from Peter Garrison himself, founding partner, a man whose face appeared regularly on the cover of Fortune and whose name was spoken in Silicon Valley with the same reverence that medieval pilgrims reserved for saints.

Garrison's office was on the third floor, with a wall of windows that looked out over the hills and the oaks and the endless California sky. On his desk sat a Lucite tombstone commemorating the IPO of Netscape, in which Garrison had been an early investor. On the walls hung photographs of Garrison with presidents and technology pioneers and, in one conspicuously placed frame, with the Dalai Lama.

"Your product works," Garrison said. He was a large man in a gray suit that cost more than most people's cars, and he spoke in a voice that was designed to be reassuring, the voice of a father explaining to a child that everything would be all right if the child simply listened carefully and did as instructed.

"Our product works beautifully," Diana said.

"It doesn't scale."

"It's scaling. Three hundred businesses in twelve states. Monthly active users growing at fourteen percent."

"That's linear growth. Linear growth is not interesting. Linear growth is a bakery, not a technology company. Technology companies grow exponentially or they die. The question is not whether your product helps small businesses connect with customers. The question is whether your product can become the platform on which all small businesses connect with all customers. That is a one-hundred-billion-dollar question. Are you a one-hundred-billion-dollar answer, or are you a thirty-million-dollar answer in a market that only funds hundred-billion-dollar questions?"

Diana looked at Marcus. Marcus looked at the Lucite tombstone. The silence that followed was the silence of a fork in a road, the moment when a decision splits the timeline and creates two possible futures, one of which will be lived and one of which will become a ghost that haunts the one that is lived.

"In order to get to exponential," Garrison continued, "you need to stop thinking about bakeries. You need to think about data. Every small business that uses your platform generates data. Customer preferences, purchase patterns, demographic information, location history. That data is more valuable than the connection. That data can be packaged and sold. That data can be used to predict consumer behavior at a scale that makes your fourteen percent monthly growth look like what it is: a rounding error."

"That's not what we built this for," Diana said.

"What you built it for is irrelevant. What matters is what you can build it into. The internet is not a democracy. It's a land grab. The people who got here first are planting flags on every hill. If you're not planting a flag, you're providing the soil for someone else's flag. I'm offering you a flag."

Point B-Prime: April 2001, a coffee shop on University Avenue, Diana sitting across from a reporter from the San Jose Mercury News who was writing a story about the dot-com collapse and had asked for an interview with a founder who had lived through it. The reporter's name was Melissa, she was twenty-four, and she asked questions with the earnest intensity of someone who had not yet learned that most stories had no neat ending.

"Everyone talks about the money," Melissa said. "The billions that were lost. But I want to know what it felt like. Not the numbers. The feeling."

Diana stirred her coffee. The coffee shop was one of those places that had multiplied in Palo Alto during the boom, with exposed brick and reclaimed wood and a name that was a single abstract noun: Verve, or Hearth, or something equally suggestive of authenticity that cost five dollars a cup to experience.

"It felt like being two different people," Diana said. "The person who started the company, and the person who ended it. And the strange thing is that I can't identify the exact moment when I stopped being the first person and started being the second. It wasn't a decision. It wasn't a single event. It was a thousand small adjustments, each one so minor that I didn't notice it at the time. Saying yes to a feature that collected more data than we needed. Saying yes to a partnership that monetized the data we collected. Saying yes to a term sheet that gave investors control over the board. Each yes made sense in isolation. Together they added up to a company I didn't recognize."

Vector Between: The technology Lumina built in 1999 was designed to make people feel known. The business Lumina became in 2000 was designed to make investors feel rich. The space between those two purposes was not a gap but a gradient, a continuous slope down which the company had slid one compromise at a time, each compromise so reasonable that objecting to it would have seemed irrational.

And Diana had been the one making the compromises. Not Marcus. Marcus had left the company in August of 2000, six months after the Series B closed, because he could not stomach the pivot from empathy to data extraction and because he was not, by temperament, capable of doing work he did not believe in. Diana was. Diana had discovered, to her horror and fascination, that she possessed a capacity for self-justification that was almost infinite.

"It's a means to an end," she had told Marcus during their last conversation before his departure. "We take the investment, we build the scale they want, and then once we're too big to ignore, we pivot back to the original vision."

"Have you ever heard of a company that pivoted back to its original vision after taking eighteen million dollars from Peter Garrison?"

"No."

"Neither have I. There's a reason for that."

Marcus had been right, of course. The company had not pivoted back. The company had pivoted further away, into a business model that Diana could not explain to her parents without using words like monetization and data arbitrage and conversion optimization, words that had not existed in her vocabulary when she had stood in front of the whiteboard on Emerson Street and drawn circles that represented human connection.

Point C: August 1999, the Series A pitch meeting. Diana and Marcus presenting to a room of twelve investors, all men, all white, all wearing variations of the same Patagonia vest over the same button-down shirt. The prototype was working. The first fifty small businesses had signed on, lured by the promise of a platform that would help them understand their customers in a way that felt personal rather than predatory. The pitch deck was twenty slides, and every slide contained at least one photograph of an actual small business owner, an actual human being, someone whose life would be improved by what Lumina was building.

"We're not competing with Amazon," Diana said. "We're not even in the same category as Amazon. Amazon wants to replace small businesses. We want to make them indispensable. The internet should strengthen local communities, not dissolve them. Every dollar that flows through our platform stays in the community where it was spent. Every connection we enable is a connection between two actual people who live in actual places. That's not just a business model. It's a moral position."

After the meeting, one of the investors, a man named Tom who had made his first fortune in enterprise software and his second in something that involved fiber optics and right-of-way agreements, pulled Diana aside.

"That was the best pitch I've heard all year," he said. "And it's never going to work."

"Why?"

"Because you're not building a company. You're building a protest. Protests are noble. Protests are necessary. Protests do not generate the kind of returns that justify the risk profile of a venture capital fund. If you want to change the world, run for office. If you want to build a company, build a company. And if you want to do both, you need to understand that the second thing will eat the first thing, every single time."

Point C-Prime: June 2001, Diana walking alone along the Stevens Creek Trail, the only place in Palo Alto where she could pretend, for an hour or two, that the world was not made entirely of ambition and money and the wreckage of ambition and money. The trail followed a creek that ran through the city like a vein, and on this particular afternoon it was empty, the California sun pressing down through the oaks like a physical weight.

She walked and thought about the photograph on her desk, the forty-nine people with their cheap champagne and their radiant grins. She thought about the bakery in Des Moines, which had been one of Lumina's first clients and which had closed in March, unable to survive the recession that had followed the bubble's collapse. She thought about Tom the investor, who had been right about everything and wrong about the thing that mattered most, which was that the world needed protests disguised as companies or it would become a place where nothing mattered except the returns.

But Tom's rightness was a fact, and Diana's idealism was a feeling, and in Silicon Valley in 2001, facts had won. Facts had annihilated feelings with the same efficiency that the market had annihilated valuations, and what remained was not a lesson but a landscape, a geography of failure that stretched from Sand Hill Road to the horizon.

Interpolation Complete: The Diana of 1999 had believed that technology could be a force for human connection, that the internet could make the world smaller in the way that made strangers into neighbors rather than consumers. The Diana of 2001 believed that technology was a mirror, that it reflected back whatever humanity aimed at it, and that humanity had aimed greed and extraction and the ruthless optimization of attention, and the mirror had reflected exactly that.

These two Dianas were not different people. They were the same person viewed from different angles, the vector between them not a transformation but a revelation. The idealism had not been replaced by cynicism. The idealism had been tempered, which was a different thing, a more durable thing, a thing that could survive contact with reality because it had already absorbed the worst that reality could deliver and was still standing.

She did not start another company. She did not write a memoir. She did not become a venture capitalist or a consultant or a cautionary tale on the conference circuit. What she did was simpler and harder: she lived with the knowledge that she had been both the person who wanted to change the world and the person who had let the world change her, and these two truths were not contradictory but complementary, two poles of a single life that existed, fully and completely and without resolution, in the space between them.

The river of capital had flowed on, indifferent, past Lumina and past Diana and past the forty-nine people in the photograph and the bakery in Des Moines and the office on Sand Hill Road with its photographs of presidents and its Lucite tombstones. It had flowed into the next boom and the next bust and the next generation of founders who would stand in front of whiteboards and draw circles that represented their dreams and would learn, or not learn, the lesson that Diana had learned: that the space between what you want to build and what the world wants you to build is not a distance to be crossed but a territory to be inhabited, and that living in that territory, in the gap between ideal and reality, was the only honest way to live.