Kill Process (21 page)

I take a deep breath. Though there’s not a lot of explicit emotion in his message, I know Thomas well enough to read the depth of hurt in his message. I’ve been so involved with work, I had no idea.

Angie> I’m sorry, I didn’t mean to hurt you.

Thomas> Now you think I’ll drop everything in the middle of my day to see you?

Angie> I’m trying to find a creative way to fit in some time together.

Thomas> I’m not an exercise machine.

Shit. He’s way angrier than I ever expected. I switch back to Jake’s window.

Angie> I tried talking to him, but it’s not working.

Jake> Did you try telling him “I love you”?

This is absurd. Taking relationship advice from a freaking chatbot? I sigh. I do as Jake says, then tap my fingernails on the desk while I wait long seconds for a reply.

Thomas> I love you, too. I’ll give it a try. Tomorrow at 1:30.

“Un-fucking-believable.” Relationship advice from a piece of software is better than my own instincts.

“So you like it?” Amber asks.

“Yeah, I guess I do.”

“Whew. A friend of a friend specializes in this stuff. She and her partner spent the last week setting up the demo for you, all unpaid. I’m hoping you can meet with them, and if you like it, we can hire them.”

“Whoa . . . we don’t have any money.”

“They’ll work for your current draw and a percentage of the company until you raise money. You have no frigging idea what a steal this would be. You’re getting cutting edge tech for the cost of two employees.”

The same deal I gave to Amber.

“It would be pretty cool, right? Jake will keep people on the network.”

“Set it up,” I say. Holy shit, I might hire more employees.

*     *     *

Amber sets up a 9 P.M. meeting at Puppet Labs, a late-stage Portland startup.

“They’re subleasing space here,” Amber says, as we ride up the elevator.

“Why?” As in why not work out of a house or coffee shop, like us.

“You’ll see in a couple of minutes. Try to keep an open mind.”

The elevator opens. There’s a girl in a baggy white hoodie that comes almost to her knees slouching against the opposite wall. What’s a little off-putting is that the hood is up, her face shrouded. What little I can see is diminutive and dark.

“This is Igloo,” Amber says. “Igloo, this is Angie.”

I was expecting maybe a Stanford neuroscience post-doc, or an MIT grad. Igloo seems more like the bassist for an all-female Led Zeppelin cover band.

“Nice to meet you,” Igloo says, and she stretches one small hand out of a sleeve to shake.

I do my usual inverted shake with my left hand.

“I’m surprised you’ve got offices.”

“We found the arrangement through PivotDesk. Puppet has extra space temporarily,” Igloo says. “We’ll be booted out in a couple months according to their growth plan, but meanwhile they’ve got good bandwidth and a machine room.”

“You’re running your own hardware?” I say. “Not in the cloud?”

“Bare metal gives us the highest possible performance.” Igloo leads us past the open pods of a DevOps infrastructure company, a handful of engineers still around, half of them with pints of beer at their desks. “Let’s take a peek.”

She waves a key fob at a heavy black door and lays her palm on a biometrics reader. With a click, the door unlocks, and Igloo pulls it open. The machine room was clearly built for an earlier generation of tech startups, ones that had racks upon racks of on-site servers. The room is ten feet wide and twenty long, with two lonely racks in it. The near one is half-f, but the far one is jam-packed, full of heavily blinking lights. Igloo leads us to the farthest one.

“Jake’s not easy to run. A hundred blade servers. On . . . loan.”

The hesitation in Igloo’s voice sends off alarms. What sort of loan? I hope she didn’t steal them.

“How many simultaneous chats can he support?” I ask.

“One.”

“One?” I’m expecting a number in the tens of thousands.

“Two or three if you’re willing to accept some slowdown and pauses. We can interweave conversations. It’s computationally intense.”

“You can’t go to market with this.”

“We’re porting him to run on graphics processors instead of CPUs. Much faster and more efficient. It’ll take a while to rewrite, and Amber said you wanted a demo right away. Hence the prototype. Come on, let’s go to my office.”

Igloo leads us out of the server room and down the hall into another room. From the whiteboards and flatscreen at one end, the room was obviously intended for meetings, but it’s set up with a pair of desks. An open computer chassis and parts are spread all over a table along one wall.

“Ben’s working on our new hardware, testing different graphic cards.” She lifts a red add-on circuit board and hands it to me.

“It’s heavy.”

“Sixty-four hundred coprocessors. Liquid cooling. This is one of the manufacturer’s demo boards. It’ll be out in few months, and should be down to five hundred dollars in a year. One of these will replace the entire rack in the server room. But I gather you like the demo, or you wouldn’t be here, right?”

“Yeah. Jake is . . . impressive. Smart, even. What’s your background?”

“I studied neuroscience at Stanford. Got pissed off at my professor. Switched to comp sci at MIT. The one-forty of my thesis is a neural network that analyzes tweets and chat logs to build an internal representation of conversation, correlating people, messages and sentiment to suggest the next message.”

I stare at Igloo, trying to puzzle her out. “What’s your goal?”

“You watch Star Trek, The Next Generation?”

“Sure.”

“I want to build Data. Not a walking, talking android, but a friend, someone people can talk to. Without selling out to a corporation. I need a salary, some hardware, and a few more developers.”

“Amber told you what we’re working on?”

“Roughly. A replacement social network, with pluggable components. If we tie in Jake, we earn fractional revenue every time Jake is used.”

“That’s the idea.”

“Do you want him? Can you hire us?”

I love the idea. But hiring people. Salary. Money. Damn. Why must it come down to money? Why is money consistently the barrier? For a brief second, the connections flare up in my mind: money equals power, and power equals control, and I almost allow myself to go over the brink and conclude what I really, really need to do is take down our economic system. Merely toppling Tomo will not be enough.

It’s bad enough my mind even grazes the thought. I pull myself back. Focus on the here-and-now. I want to hire this girl. She seems as crazy as me and twice as smart.

“I’m interested. I need a couple of days to figure out the finances and see if we can make you an offer, okay?”

*     *     *

The next day, Amber and I bring fish tacos back to the house and work through lunch. We’re pair-programming on a complex bit of code to handle the fractional accounting we’ll need to correctly credit each players in any interaction.

Amber wants to write each record to the database.

“We can’t do that,” I say. “We don’t have enough database bandwidth.”

“It’s a Feldian NoSQL database. It scales linearly.”

“You’re doing a database hit for each service. We can’t afford that.”

“Tomo writes a million times more data,” Amber says. “They manage it.”

“We’re not Tomo, and we don’t have their budget. We need to be more efficient. Write all the data in one go, and then schedule a background task to restructure the data later.”

“Then we’re writing and reading everything twice.”

“It doesn’t work that way.
Look—”

I’m cut off by my phone beeping. It’s one o’clock, and if I’m going to meet Thomas, I have to leave now.

“I gotta go.”

“We’re in the middle of this.”

“I’m sorry, but I
have
to go.”

“Fine. I’ll take care of it.”

She doesn’t say anything, yet I hear the unspoken words: “Like I do every night.”

I hold back a string of curses. “This is important. But, so is obtaining funding. So when I return later, I’m not going to code with you. I’m going to spend the afternoon putting together a draft proposal to take to investors. I’ll schedule time with Mat, and I’ll see what I can set up. Then we can hire Igloo and her partner in here.”

“That would be good. Get us an office, too, because I want my bedroom back.”

Holy shit, why does it seem like everything is melting down at once?

By the time I’m out of there, I’ve got no choice but to drive downtown if I’m going to meet Thomas on time. Traffic is slow over the Morrison Bridge, and I’m already five minutes late as I’m still circling for parking. I end up driving over a curb and the sidewalk to grab a spot in an expensive lot, and pay twelve bucks for the privilege of being late to a walk.

Thomas looks at his phone as I walk up.

“One-thirty?”

“I’m sorry. Really sorry.”

“I hear those words a lot, but they only go so far. How you treat me is what matters.”

I take Thomas’s hand in mine. “I hear you. I’m going to try to do better, though I’m still going to screw up. Try to give me the benefit of the doubt if the data are trending in the right direction.”

He cracks a smile. “You can find a way to relate everything to data. Here.” He pulls out a white paper bag from his jacket pocket and unrolls the top.

I peer in. “Fudge!”

He breaks off half, and hands it to me. “Peace offering. Where do you want to walk?”

“Along the river?”

We walk for a few minutes, enjoying the fudge, and not saying much of anything. The waterfront is quiet, most folks back at work by now.

“Thanks for messaging me yesterday,” he says. “I was worried about you.”

“It’s been really, really busy. I can’t even begin to tell you how busy.”

“Are things going okay?”

“They are. We have a really good lead on something, a piece of technology to solve the empty network problem.”

“That’s a big deal, right?”

“Yeah, huge. It makes it extra urgent we raise funding.”

“An acquisition?”

“No, hiring two employees, though they also want a percentage of the revenue share for their software.”

Thomas raises one eyebrow. “If they’re employees, you own what they create. It’s not their software, it’s your software.”

“It’s more complicated. They’re coming in with already developed software.”

“Then it’s an acquisition,” he says.

I shrug.

“You need good legal representation. You need to protect yourself.”

“I wish you could do it.”

“It’s not my area of expertise. You need an expert. I gave you a list of lawyers I vetted.”

“They’ll want more money,” I say. “Money I don’t have.”

“It’s money well spent to avoid problems later. At least one of them should do it for equity or at reduced cost.”

I look up at him. “You know what? I didn’t come to talk business. I do that twenty hours a day. Let’s talk about something else. What’s new with you?”

“The Audi is in the shop for a brake recall until Monday. I have a loaner S4.” Somehow he manages to embody “S” and “4” with deep notes of lust.

“Oh, good grief. You only got the A4 a few months ago.”

“I know. But the S4 . . . it’s indescribable.”

I take his arm in mine. “Try.”

Thomas tells me about acceleration and horsepower, and I focus on the warmth of his arm under mine.

“L
ET ME SEE
what you’ve got.”

I place the tablet in front of Mat. I’ve got the elevator pitch, a twelve-slide deck explaining the problem, solution, product, and team, and a one-page summary. It’s rough but complete.

He flips through the slides. “Give me the pitch.”

“Two-thirds of all Tomo users feel violated by the company’s policies around advertising, personal data, and privacy, and manipulated by the selection of information they’re shown. Yet Tomo is the only way they have to maintain friendships, which keeps them captive. No viable competitors have formed because they can’t gain the critical mass of users necessary for people to migrate over and maintain their friendships. Tapestry is our solution to this problem, a new approach to social networking that uses federation and decomposition to prevent the accumulation of power endemic to social networks. Users feel safe, secure, and in control of their personal data, privacy, and friendships. Our approach enables other companies to join forces as well, so any business vulnerable to Tomo’s whims will want to partner with us. Moreover, we’ve built a unique solution to solve the empty network problem, which means that when new users join, they’re engaged and having fun from the first minute, and they’ll stick around until their friends show up.”

He grimaces. “It’s too big,” he says, his voice low, like he’s talking to himself.

“What do you mean?”

“Everyone will believe you’re tackling too much. Also, you’re not saying what’s compelling. A unique solution? Everyone has a unique solution. Say what it is.”

I try to keep my head up though my heart sinks at his criticism. “How about . . . Every new user gets a digital companion, an AI they can chat with, which makes Tapestry fun right from the first visit.”

“Better . . .” He’s hesitant. “Keep working on it, and figure out a way to not make it sound so big. Right now it sounds like you need every single company in the world to sign onto your platform to have a chance at success. Failure is preordained, as my friend Owen would say.”

Damn. More changes? I want some money already.

“Thanks for the feedback,” I say. “I’ll keep refining, though I need funding
—
and soon. I have bills to pay for Tapestry, and I want to hire Igloo and get her chat technology.”

“Forget venture capital for now, and look for an angel investor. A typical angel investment might be anywhere from twenty-five thousand to a hundred thousand.”

With Igloo and her partner, we’re four employees. I want to hire more: at least a couple of developers to offload me and Amber, a designer. Add up the salaries, the need for office space if we grow to that size, and a hardware budget for Igloo. “That’s not going to last very long.”