my teammates’ sense of entitlement,
their natural ease. I began wearing flan-
nel. I incorporated B Vitamins into my
regimen and began listening to E.D.M.
while I worked. The sheer ecstasy of
the drop made everything around me
feel like part of a running-shoe ad or a
luxury-car commercial, though I couldn’t
imagine driving to E.D.M. Was this
what it felt like to hurtle through the
world in a state of pure confidence, I
wondered—was this what it was like to
be a man? I would lean against my stand-
ing desk and dance while pounding out
e-mails, bobbing in solidarity with the
rest of the team.
E
ach new employee, regardless of
department, was required to spend
a few days at the Solutions cluster, an-
swering support tickets––like work-
ing the mail room in Hollywood. The
C.E.O. believed that this experience
built empathy for our customers. It did
not necessarily build empathy for Sup-
port. The engineers and salespeople
tossed off replies to customer inquiries
and rolled their eyes at developers who
did not understand our product. The
engineers had been hired at two or three
times my salary, and their privileged po-
sition in the industry hierarchy should
have exempted them from such tedium.
It wasn’t exactly that they harbored con-
tempt for our users; they just didn’t need
to think about them.
In theory, the tool was straightfor-
ward. But when users—engineers and
data scientists, almost all of them men—
encountered problems, they would level
accusations and disparage the company
on social media. My job was to reassure
them that the software was not broken.
Looking at their source code or data, I
explained where things had gone hay-
wire. Some days, helping men untangle
problems that they had created, I felt
like a piece of software myself, a bot:
instead of being an artificial intelligence,
I was an intelligent artifice, an empa-
thetic text snippet or a warm voice, giv-
ing instructions, listening comfortingly.
Twice a week, I hosted live Webinars
for new customers. I asked my parents
to join, as if to prove that I was doing
something useful, and, one morning,
they did. My mother e-mailed after-
ward. “Keep that perky tone!” she wrote.
After two months, the Solutions
manager took me for a walk around the
neighborhood. We passed a strip club,
a popular spot for parties during devel-
oper conferences, which my co-workers
claimed had a superlative lunch buffet.
We circumvented people sleeping on
steaming grates. He looked at me with
kind eyes, as if he had given birth to me.
“We’re giving you an extra ten thou-
sand dollars,” he said. “Because we want
to keep you.”
T
he simplest way to solve users’ prob-
lems was by granting the Solutions
team access to all our customers’ data
sets. This level of employee access—
some of us called it God mode—was
normal for the industry, common for
small startups whose engineers were
overextended. It was assumed that we
would look at our customers’ data sets
only out of necessity, and only when our
doing so was requested by the custom-
ers themselves; that we would not, under
any circumstances, look up the profiles
of our lovers and family members and
co-workers in the data sets belonging
to dating apps and shopping services
and fitness trackers and travel sites. It
was assumed that if a publicly traded
company was using our software we
would resist buying or selling its stock.
Our tiny startup operated on good faith.
If good faith failed, there was a thor-
ough audit log of all employee behav-
ior. The founders tracked the customer
data sets we looked at and the specific
reports we ran.
Early in the summer of 2013, news
broke that a National Security Agency
contractor had leaked classified infor-
mation about the U.S. government’s
surveillance programs. The N.S.A. was
reading private citizens’ personal com-
munications, crawling through people’s
Internet activity by gathering cookies.
The government had penetrated and pil-
laged the servers of global technology
“There she is. The girl I’m going to marry.”
• •
THENEWYORKER,SEPTEMBER30, 2019 59
