Chapter 10 : First Pitch, First Pain
[Philz Coffee, 24th Street. January 2014, 1:47 PM]
David Park was shorter than his LinkedIn photo suggested and younger than his title implied. Mid-twenties, Korean-American, in the uniform of a junior VC. Fitted chinos, a quarter-zip Patagonia pullover, and a leather messenger bag that cost more than Ethan's monthly rent.
He stood when Ethan approached the table, extending a hand with the practiced warmth of someone who shook hands for a living. "Ethan. Been a while, man."
"David. Thanks for making time."
They shook. David's grip was firm but brief. He'd been taught that too long signaled neediness and too short signaled disrespect. Ethan ordered a mint mojito iced coffee because the menu at Philz was an adventure and he'd given up pretending this body had established preferences.
"So," David said, settling back into his chair. "The pivot. AI, you said? That's... unexpected."
"Unexpected is the point."
"Walk me through it."
Ethan opened the laptop. Nine slides. Sarah's architecture: the consumer-friendly deck stripped of technical jargon and loaded with applications. He turned the screen toward David and began.
Slide one: What if a computer could write?
David's eyebrows lifted. Good sign.
Slide two: Market applications. Content generation for marketing agencies. Document summarization for legal firms. Automated customer service. Each one with a market size number that Sarah had pulled from industry reports.
David nodded. Scribbled something in a small Moleskine he'd produced from the messenger bag.
Slide three: The technology. Deliberately vague. "A novel machine learning architecture that processes language the way humans do, reading entire documents simultaneously, understanding context, generating coherent responses." No mention of attention heads. No mention of Transformers. Just the output.
"Interesting," David said. "Really interesting. But here's the thing. I'm an associate. I can bring deals to the partners, but I can't write checks. If you want Basecamp's money, you need to pitch Alan Rao. He's our managing partner. He does the final call on anything over fifty K."
"Can you set that up?"
David tapped his pen against the Moleskine. "I can get you fifteen minutes. But Ethan, be straight with me. Your company's been dormant for months. You've got no team, no traction, and you're pitching into a category that most of Sand Hill Road considers academic. What changed?"
A reasonable question. An honest question. The kind of question that deserved a truthful answer, which Ethan couldn't provide.
"I found something," he said. "A technical approach that nobody else is working on. It's not incremental, it's a generational leap. I can't explain the full architecture yet, but I can tell you that the math works and the early results are promising."
David studied him. The Moleskine stayed open. The pen stayed still.
"Okay," he said. "I'll get you in front of Alan. Thursday, 3 PM, our office on Sand Hill. Bring your deck and whatever demo you've got."
"I'll be there."
---
[Basecamp Ventures, Sand Hill Road. January 2014, Thursday, 3:02 PM]
Alan Rao's office was on the second floor of a building that looked like every other building on Sand Hill Road. Glass and beige stucco. Built to project wealth without looking showy. The lobby had a Keurig machine and a receptionist with a smile optimized for hospitality.
David met Ethan at the door. "Alan's running two minutes behind. He's got a four o'clock, so you've got maybe twenty minutes instead of fifteen. Don't waste the first five on pleasantries."
Solid advice. Ethan gripped his laptop bag and followed David upstairs.
Alan Rao was a different species from David Park. Mid-fifties, silver at the temples, wearing a sport coat that cost more than Ethan's checking account. He stood behind a mahogany desk cluttered with business plans and a framed photo of himself shaking hands with someone who might have been Marc Andreessen. His handshake was not practiced: it was automatic. He shook hands like he'd evaluated ten thousand pitches and stopped counting.
"David tells me you're doing AI," Alan said, sitting down without waiting for Ethan to sit first. "Machine learning, specifically."
"That's right."
"All right. Show me."
Ethan opened the laptop. Nine slides. The same deck that had made David lean forward. He started with slide one — What if a computer could write? — and delivered the pitch the way Sarah had coached him. Output first. Applications. Market size. Technology overview. Team, just him and a CTO he was onboarding, which was technically true as of three days ago.
Alan listened. His face revealed nothing: the practiced neutrality of a man who could reject founders without guilt and write seven-figure checks without excitement. He let Ethan get through all nine slides without interrupting, which was either a very good sign or a very bad one.
Then the questions started.
"The technology," Alan said. "You say it generates text. Like a language model?"
"More advanced than anything currently in production. Our architecture processes entire documents simultaneously..."
"So it's like Siri?"
The words landed like a brick on glass. Ethan's jaw tightened.
"No. Siri retrieves information. Our system creates content. New text. Original paragraphs. It's generative, not responsive."
"So it's like autocomplete? Like the iPhone keyboard?"
"It's..." Ethan paused. The gap between what he was describing and what Alan's frame of reference could accommodate was an ocean. The man wasn't stupid. He was contextless. In 2014, the most sophisticated language technology a non-technical person had encountered was autocorrect and voice assistants. There was no bridge between that experience and what a Transformer could do.
"Think of it as a writer," Ethan tried. "Not a search engine. It reads documents and produces new documents in the same style. Marketing copy. Legal briefs. Customer service responses. Content that currently requires human labor."
Alan's eyebrows pulled together. The expression of a man trying to fit a square concept into round experience. "And the moat? If this works, what stops Google from building it?"
"The architecture is proprietary. Three years ahead of anything in academic literature."
"How do you know it's three years ahead?"
Because he'd read the paper that invented it in a timeline where it had already been published. Because he carried the blueprint in his skull. Because ChronoCloud's servers ran on hardware that wouldn't be manufactured until the next decade. None of which he could say.
"I have good technical intuition," Ethan said. The words tasted flat.
Alan glanced at David. David gave a micro-nod: the kind of signal that meant "I vouch for this, but not with my career."
"Interesting technology," Alan said. "But the market isn't clear. You're describing something that replaces human writers, which is a hard sell to companies that employ human writers. No traction, no revenue, no team beyond you and a CTO. I'd want to see product-market fit before we commit."
"What does product-market fit look like for you?"
"Paying customers. Three to five. Businesses using your product and willing to say so publicly. Come back when you have that, and we'll have a different conversation."
The rejection was polite, professional, and absolute. Alan stood. The meeting was over. Twelve minutes.
David walked Ethan to the parking lot. "Don't take it personally. Alan passes on ninety percent of what he sees. The fact that he asked questions is a good sign."
"He compared it to Siri."
"Everyone's going to compare it to Siri. It's the only reference point people have. You need a demo that makes the comparison impossible."
Ethan shook David's hand. Got in the car: the dead man's car, a 2009 Honda Civic. It had sat in a parking garage since October, retrieved yesterday for the drive to Sand Hill Road. The engine turned over on the second try. The gas gauge showed a quarter tank.
He drove back toward the city. The 280 freeway was clear, the peninsula rolling past in its usual California-perfect greens and browns. His mind replayed the meeting in a loop. So it's like Siri? The question that would haunt every pitch until he could put a working product in front of someone and let the output speak for itself.
A bar materialized on his left as he exited toward the Mission. Dark interior. Neon sign. The kind of place that served whiskey without asking questions.
Ethan parked. Walked in. Sat at the bar. Ordered a Jameson, neat.
The bartender: a thick-armed woman with a tattoo sleeve and the efficient disinterest of Bay Area service workers, set the glass down without small talk. Ethan drank. The whiskey burned going down. A good burn. The kind that replaced the cold knot in his chest with something warmer and less productive.
Two PM on a Thursday. Drinking alone in a bar on Valencia Street. The pitch that Sarah had helped him build, nine slides, stripped to essentials, built to carry across the comprehension gap. It died in twelve minutes against a man who thought AI meant a phone assistant.
The bartender refilled his glass without being asked. "Rough day?"
"Had a meeting. Didn't go well."
"What do you do?"
"I'm building AI that writes."
The bartender considered this. "Like Siri?"
Ethan stared at the glass. Drained it.
"Exactly like Siri."