In this extract, Delaney goes to work in a department called TellTale. The manager, Alessandro, explains what the department does…
“People pick up a book,” he said, “and stop in the middle.
Why? With e-books we can study all of this in aggregate. We can take, for
example, two thousand readers of Jane Eyre and see who finished it. We actually
did that. Turns out 188 people did finish it. That’s not good, right? People
who read it all seem to like it. It’s at—” He tapped one of his screens a few
times and got the answer. “It’s a 83 per cent approval, which is high for a
dead author. So we dig deeper and see that of the 2000 people who started Jane
Eyre, most quitters put it down around page 177. So then we look at what
happens on page 177, and we see that people don’t seem to like this character
called Grace Poole. They find her scary and depressing. They want more of the
romance with Mr. Rochester. Now, if the author were alive, we could tell him—”
“Her,” Delaney corrected.
“Her?” Alessandro said reflexively, then went pale.
Oh no, Delaney thought. She’d seen this happen.
Alessandro stood up. He didn’t know what to do. He allowed
the vines of his hair to obscure his face completely, and seemed likely to
douse himself with gas and light a match.
“It’s OK,” Delaney said. “You meant Mr. Rochester.”
“I did?”
“I know you did. Like, if Mr. Rochester were alive, we could
tell him that his parts of the book were intriguing.”
“Okay. Right.” Alessandro began warming to the idea,
believing it his only way out of ruination. “That’s exactly what I meant. Where
were we?”
“You were saying you could fix Jane Eyre.”
“Well, we could,” he said, “but because the author is dead,
she”— he pressed the word so hard it erased centuries of chauvinism and
ignorance, “can’t learn from the data herself.”
“Sucks for her,” Delaney said, and saw Alessandro smile.
“But a living author, or a publisher, can avail themselves
of the numbers and act on them,” he said. “And this kind of data has been
invaluable to publishers and some of their authors already. Just those
data-points, sales versus book-starts versus completions, that’s huge.
Completions, of course, convert favourably to sales of the author’s next book.
So for publishers, figuring out where and why people are stopping is crucial.
Sometimes it’s obvious. Like unlikable characters. We can help fix that. We’ve
actually written a pretty simple code for turning an unlikable character into,
like, your favourite person.”
“Wow,” Delaney said.
“The main thing is that the main character should behave
they way you want them to, and do what you want them to do.”
“That’s just common sense,” said Delaney.
“Right? It kind of makes you worry about a lot of writers –
the fact that they didn’t know this. But we’re making inroads with colleges and
MFA programs, so now they have the information. We give it to them for free, as
a public service.”
Delaney made a grateful, admiring whistle-sound.
Alessandro moves on to discuss screen-writing and the algorithms The Every has developed to “help” screenwriters create more engaging movies (“For example, 82% of the best scripts have the Catalysing Moment on page 11…") These formulas don’t just apply to the creative side of movie making, but also to the critical side…
“Remember when assessments of the quality of a movie were an
unruly mess? Before critical aggregation, it was totally random. For a certain
movie, you had one critic saying one thing in Los Angeles, and another one in
Oslo saying something else, and there was no chance of order or consensus. But
when we applied percentages to each review, we could average them together, and it
became far more clear. A regular human in a fast-moving society doesn’t have
time to read twenty-five, or even three, reviews of a movie before they go see
it. But seeing it’s received an aggregated score of 74.61 percent – that’s
clarity.”
“And clarity is objectivity,” Delaney said.
He looked at her and grinned. “You belong here.”
“As you know,” he continued,” the aggregates worked so well
that they quickly moved from movies into painting, dance, sculpture, poetry. I
mean, you should have seen how low sonnets scored! That’s why you don’t see
those taught so much any more.”
“Sonnet? What’s a sonnet?” Delaney laughed.
Alessandro’s eyes were wild with mirth and inspiration.
“Then we brought numerical specificity to so-called fine art museums. In each
case there was some initial pushback, but the undeniable comfort of the numbers,
of simply knowing the quality of a work of art by its percentage, was soon
embraced by the overwhelming majority of people. Eighty-eight percent, to be
exact.”
“That’s the only reason we finally know who the best painter
was,” Delaney said. The Every had released their results a few years earlier,
the product of 32.1 million respondents. The greatest artist of all time was
Norman Rockwell, followed by Dale Chihuly, Frida Khalo, Pablo Picasso and
Patrick Nagel.
They shared a smile. “I liked when the Louvre started
showing the aggregates,” Delaney said.
“Right,” Alessandro said, “I mean, the Louvre came to us.
There were all these places doing bootleg aggregates and they wanted it done
right.”
“It was fascinating when The Last Supper was only a
sixty-six percent,” Delaney said.
“It had been so overrated for centuries! See, that kind of
thing was revelatory. We’re averaging together tens of thousands of ratings, as
opposed to taking the received wisdom of a few academics. The aggregates are
more democratic and egalitarian. Before the aggregates, it had all been so
hierarchical and subjective.”
“Subjectivity is just objectivity waiting for data,” Delaney
said.
“Right!” Alessandro said, and she saw him pause, again
working out whether any kind of compliment would be acceptable. He decided not
to risk it. “I mean, I think creatives are starting to acknowledge that we have
a valuable role to play. This high wall that was built between art and data had
to be torn down.”
I think I should read these books... even though the underlying themes are likely to depress. Hopefully there's no algorithm to feed my depression back to the author (yet).
ReplyDeleteDave Eggers appears to be a total technophobe. He writes his books on an old laptop that he refuses to connect to the net in case it tries to update the software he's been using for years. So I reckon he's fenced himself from the algorithms as much as he possibly can.
DeleteHell in a handcart springs to mind.
ReplyDelete