Why SWEs Can't Have Nice Things

Why SWEs Can't Have Nice Things

Not that long ago, and not that far away (from San Francisco, as it were) at a
software company, one of those billion-dollar concerns you've heard of, a Vice
President attended a dwarf-tossing-slash-coke orgy on his rival's yacht. He 
couldn't enjoy himself, for that intrusive realization: "This guy's boat's 
twice as big as mine! No wonder I haven't been promoted." So, when he returned
to work, he knew what he needed to do: "free up salary" (fire people) to justify 
a bonus for himself. 

He calls one of the firm's economists into his office, and describes a "special
project" he'd like the economist to do. It doesn't take long for the economist 
to realize that this "project" is the replacement of his entire department, and 
because his master's thesis was on the bargaining power of cartels, he brings
the other economists together and explains what the VP's trying to do. The next
day, they call a meeting with him and say they've formed a union. They agree to
do the project, but only in exchange for massive raises and a tenure system to
protect their jobs.

The VP gets furious, but he knows he's been beaten. "Fry my frenulum," he says.
"That didn't work."

Perhaps statisticians would be a better target, he decides. He calls his chief
data scientist into his office and describes a "special project" he'd like her
to complete. She realizes what he's really asking, but tentatively agrees to do
it. The next week, she presents the VP with a study that gave the project a 57 
percent chance of the project succeeding. "Well, 57 ain't bad," says the VP. 
"Go on." She continues to explain that her 95-percent confidence interval around
this 57 is quite wide (17 to 97) and further admits she's only about 40%
confident in her model. "These things are never actually normally distributed."
She begins to throw words like "heteroscedastic", "posterior probability," and
"curse of dimensionality", and by the time she's getting into her philosophical
spiel about Bayesian versus frequentist approaches, the VP gives up, realizing
he's out of his league-- the statisticians are far too smart for him to be able 
to evaluate their work. She keeps her job.

As soon as she leaves, he balls his fists. "Son of a taint!" For psychosomatic
reasons, he's walking funny the rest of the day. Goddammit, society owes him a
bigger yacht. He went to MBA school at Harvard, you prole fucks! His anger gets
the better of him and, in an unrelated meeting later that day, he goes off on a 
tirade reminiscent of Bruno Ganz's performance in Downfall, and his boss sends
him home.

Twice defeated, not ready to give up, really wanting to buy himself a new yacht,
the VP decides to go after the software engineers. They're just glorified 
mechanics after all, anomalously paid a middle-class salary. What do they even
get up to all day? He grabs a random engineer and pulls him into his office and,
yet again, explains the "special project" to automate and replace all of the
company's software engineers.

The engineer, like his predecessors, realizes quickly that he's being asked to
dig his own grave, so he goes to a whiteboard and begins making diagrams the VP
can't understand about the company's systems and the various programming 
languages used to build them. The VP slumps in his seat, realizing again that
he's out of his depth. The programmer, at the end of the presentation, concludes
by saying, "Really, I should be making your salary. I'm irreplaceable, bitch."

The VP put his hands in his pockets and folds in on himself, giving a 
nod of concession. The yacht of his fantasies, the yacht society owes him, 
recedes from possibility... a tear forms in his eye.

"However," the engineer continues. "If your goal is to replace all the other 
bozos on my team, this is what you need to do..."

Thanks to you, now whenever I watch GATTACA, whenever the readout says "INVALID", in my mind's eye it instead says "BOZO BIT SET".