From a writer's point of view, knowing that an electronic Neural Net / Crystal Ball existed that could analyze my script and tell a studio what the movie would make is somehow deflating. A release that would get $50 million in box office receipts is referred to as "a dog."
You gotta read it.
Of the many excellent points the article makes -- creating more than enough fodder for several dinnertime conversations -- two things stand out:
1.) Script reigns. Stars, directors, budgets, F/X... You name it. It doesn't matter. Everything rests on the story. Hollywood pays lip service to this concept, but I don't think for a moment that anyone in the industry actually believes it. Still -- facts don't lie. At some fundamental level, audiences agree. It's all about the story.

It would appear that though we can create machines to operate with ruthless impartiality, there is still a need for the human element. The machines can analyze scripts until the cows come home. That doesn't mean they like the words they're processing.
I'd rather watch a compelling small film than the current blockbuster ANY DAY. That's why I, for one, consider it a huge blessing that, so far, most studios are skeptical of Epagogix' uncanny accuracy and are not using the technology to analyze everything they option.
I happen to be a dog lover. I'm just glad that the occasional "dog film" can still make it through the studios' crystal ball mine fields and find its way to the screen...