My Kindle Mocks My To-Read Shelf: Machine Learning, Bestsellers, and the Future of Publishing

Standard

This short (>8min) Vox video on machine learning is terrific. It’s a fascinating look at the way work is being automated.

It also reminds me of last month’s post about the academics who created an algorithm to analyze books to see if they’ll become bestsellers or not. Brief summary: they subjected thousands of books to several kinds of analysis in order to identify traits that the bestsellers had that the non-bestsellers did not. They found nearly 2800 distinct differences.

The algorithm couldn’t create a bestseller, and in their book the academics were clear the technology was a long way off, if it was possible at all.

The video above explains why that is, and why the software’s ability to teach itself is so interesting. Recommended.

Since that last post, the academics who developed the algorithm and wrote the book have opened a consulting service. Of course, right? It’s the natural next step. As an author, I guess I’m supposed to find this threatening/a sham/the end of literature, but I don’t. It’s just information. The only real question is whether it’s good information.

I won’t be worried until the day editors stop reading manuscripts my agent sends them unless they’re accompanied by an Archer-Jockers Score(tm). And I don’t see that happening in my lifetime.

But no, seriously, that’s an interesting video up there.

As I write this, The Twisted Path has eighteen reviews on Amazon. My short fiction collection, which includes a 20P novelette, has eleven. It would be extremely useful if those numbers could be boosted to twenty-five. Fifty would be even better. Amazon has algorithms of its own, and works with a number of reviews that pass a certain threshold get more prominent placing in search results.

It’s all pretty opaque, but what it boils down to is more reviews=more visibility. If you read and enjoyed either of those works, please consider dropping a review for it.