Google is developing a service that will tell you what you’re reading. The Prediction API uses your historical data and labels to perform text analysis. It can easily do simple stuff, like distinguishing whether a page is in English or French. But if you feed it enough data, it can approximate whether an op-ed is “conservative” or “liberal” based on vocabulary. It is still in development and available only to registered testers at this point; but here’s a demo whipped up by Forbes‘s Jon Bruner to let you give it a try. [%comments]
Let Google Decide for You
and 
. You can also watch it on 
The Met Office is developing something which will tell you what you can see if you look out of your window. That doesn’t work either.
From the description at the end of that link, it sounds like it just builds lots of histograms of pre-selected observables like word frequency and separation, then uses those to make the decision… I think I’d prefer just to see the histograms.
It’s a bit more sophisticated than that, Ian, but the fact that it’s a “black box” service, where your data goes in one end and answers come out the other, with nothing in between, bothers some people. This will probably be of more interest to hackers looking for quick ways to build out innovative software than to developers with big budgets who want to maximize accuracy and reliability.
Plagiarism is Epidemic in our schools. Google should develop a Plagiarism tool to help highlight frank stolen passages in a pile of electronically submitted term papers. The tool should be available on line to both students and teachers for instant comparisons and citations of plagiarism.
Daylight is the best disinfectant.
this one’s easy: 6th grade reading level = conservative, 9th grade = liberal- so can i get royalties or something?
@drill-baby-drill
I think something simular to that exist but the user has to manually enter the part they think is plagiarized. A friend in college got caught plagiarizing by a professor and the professor told him that he entered it in said program. Thinking about it now it could just be that the professor made up this tool knowing it was a better way to keep students from plagiarising papers than saying he recognized it.
Now it’s easier for the government to censor hate-speech as double-plus-ungood!
And that tool is useful because…?
How about a filter that identifies well reasoned pieces based on objective standards versus romantic drivel or demagoguery.