Here's an interview with the two guys who started the project. Gabriel Ortiz and Paul Starr seem like OK fellows, and it's good that they're busy at work in these times, but, they're probably doomed to failure (at worst) or (at best) will come up with a better definition of stupidity than typos and SMS abbreviations.

[Addendum: It just dawned me, how badly the filter might do with mathematics, computational, or lexicographical texts (i.e., with any texts that use abbreviations). I just spent yesterday reading up on the programming languages APL, J, and Haskell.]

Ceci n'est pas un seing.