Bookworm and Dave Eggers

Today I was listening to the Bookworm podcast -- Bookworm is a radio show from KCRW that is syndicated on many public radio stations and available as a podcast that, naturally, focuses on books. The guest on today's episode was Dave Eggers. Eggers was there to talk about his new book The Every. Michael Silverblatt, the host, asked Eggers to read this material from the book flap of the hardcover edition. (Note: the hardcover was sold only in independent bookstores; other versions are available from amazon.) I laughed out loud as I was listening.

No one reads flap copy. No one ever has. And yet it persists, in every hardcover book and even some overly elaborate/usually French paperbacks. Countless hours are spent writing this flap copy, editing it, printing it, and then it is ignored by all. This is an unpardonable waste of resources, and proves that publishing, perhaps more than any other industry, is primed for disruption.

At the Every, we look to disrupt. We ask questions. We seek answers. We solve for solutions. And when we find the solutions we solved for, we implement them with grit.

When we took a long look at books, we found much room for improvement. Some of our initial questions were ones you have probably asked, too: Why is text read sequentially? Why do we need authors? Does anyone really like gerunds? Should there be books, when they take up space, kill trees, and require reading?

For now, let’s assume that books should exist. How do we use technology to make them better? Let us count the ways! First, characters. Not too long ago, books were full of characters who said the wrong things and did the wrong things. These characters were needlessly complex and often unlikable. We asked: Can’t there be a better way?

There can. There is. Introducing FictFix.

We don’t claim that our FictFix algorithms can remedy every error in every novel, but we can say, with empirical certainty, that it can fix 86 percent of errors in 92 percent of novels. Starting with characters. For centuries, readers were baffled by the choices certain characters made, and annoyed by certain things they said. Very often—really, too often—these characters behaved inappropriately. Especially in older novels, characters said and did things that we now know are incorrect. FictFix can and has addressed these errors, and readers have responded. Seventy-one percent of readers have found our FixedBux an improvement over their unprocessed predecessors.

But what about structure? Pacing? What about the length of sentences, of chapters? What about the removal of unpalatable ideas and the insertion of tasteful romance at predictable intervals? All of these things are solvable by solutions—we just need the data and the tools, and grit.

Questions that had, for centuries, puzzled academics—that were sometimes treated as unknowable—were easily answered by Every algorithms. How long should a book be? Now we know: no more than 368 pages. How many jokes should be in a book? Jokes should not be in books, but if there are jokes in books, there should be no more than twelve. How many sick dogs and characters named Rowena can be in any given novel? One and one.

Feeling relieved? Numbers do that. They give us certainty, and certainty makes us comfortable, and life is only livable if we are free from discomfort, or the possibility of discomfort.

But isn’t there one last question out there, difficult to answer—frustratingly subjective? Yes. Goodness. Quality. How do you measure it? How do we know which books are good and which are not? We’ve solved for that, too. Books should be rated on a website that aggregates customer reviews into a five-star average. To facilitate the greatest possible precision (and thus comfort) this average is rounded to the hundredth decimal point. This is how we know which books are good and which are not, and it’s how we know that Don Quixote is a 3.89.

But of course that rating can, and will, be improved.