It Starts With Science


I am Brandon Weiss. I build startups and web apps. I also Twitter.

Unwarranted Skepticisim

Text

I get recommendations all the time.  Everyone does.  Your friends tell you to see this movie, or read this book, or gawk at some cult classic that you’ve somehow missed all the time.  But there is no one with universally good taste.  Everyone likes something that someone else dislikes.  After a while you get to know what you like that your friends also like, and when you get recommendations you (without thinking about it) mentally cross-reference and calculate the odds that you too will like it.  A ton of stuff goes into it, even more subtle things like your friends’ enthusiasm relative to previous enthusiasm.  There are hundreds of variables and it all happens in a split second.

But every once in a while, despite the odds being very, very low, a great show will be recommended and all the signs will point to it not being very good.  Recently that was Pushing Daisies.  A bunch of people told me it was unbelievaby good, and it wasn’t until Christy told me to watch it that I downloaded the first episode and got hooked.  Just to be clear about how wrong I was, I would consider Pushing Daisies to be one of the best shows I’ve ever seen.  Ever.

But even more recently, indeed as of only a few days ago, was Cowboy Bebop.  This is on basically every top ten list of the best anime shows, and virtually impossible to not hear recommended if you’re talking to someone into anime.  But somehow I lumped it in with all the other terrible shows like Dragon Ball Z (for the record, Z sucks, but Dragon Ball was incredible).  But I was in the mood for some anime the other night, and something reminded me of it, so I downloaded it.  And it’s fucking incredible.  Just as everyone said it would be.  It’s right up there with FLCL or Ghost in the Shell.

Periodically I’ll have moments of clarity where I’ll pull back and see how our brains are just like processors or some other techno-analog, but I’ve never felt more like a computer than in realizing the algorithm I use (albeit subconciously) to determine the probability that something is awesome based on recommendations gives false positives.  Hmm, or would it be true negatives?



January 14, 2008, 1:04pm