Diego Basch

At Flaptor we just released a WordPress plugin for our automatic tagger. It’s called TagMahal (click on the link for download instructions). It’s fun to play with, I just used to pick the tags for this post.

I read a post on Techcrunch about a startup called YouNoodle. Their idea is to try to predict the success of a startup using artificial intelligence. I doubt they can come up with something better than what an experienced technologist could do with a simple checklist, but I guess time will tell. There is a link to an interview with the founders on the NYTimes. What I found interesting was the response to the obvious question, have you applied the algorithm to your company?

“So far, we haven’t run ourselves through it,” Mr. Goodson said, adding that the results could prove baffling. “If it says we’ll fail, and it’s right, that’s something of a paradox.”

It is not a paradox at all. There are many non-paradoxical explanations for such a result, for example:

• The algorithm doesn’t work at all, and it guessed luckily. For example, here’s my algorithm: For a given startup, I predict that it will fail with a 90% chance. It will be right most of the time but it’s still worthless. If I started a company based on this algorithm, it would most certainly fail.
• The algorithm works, but the startup itself will fail. Wouldn’t that be ironic? You come up with an algorithm that can predict whether a startup will succeed, but you don’t have what it takes to make it into a successful business and the algorithm knows it.
• The algorithm works well enough but it’s not perfect. YouNoodle suceeds in spite of the fact that the algorithm guessed wrongly for them.
• The algorithm doesn’t work at all and the company still succeeds! I can remember several examples of this scenario.

As I posted on the Flaptor blog recently, we released a tool to automatically suggest tags for blog posts based on a learning algorithm. You can check it out here:

Flaptor autotagger