[bbeonx]

Same. I guess my takeaway is that I'm happy people are looking into this because in ten (or twenty (or a hundred)) years this research will hopefully have paid off. I, however, want nothing to do with it.

Also, they don't do natural language processing but allow you to write method names like "the square-root of _x^2 + _x + _" where the underscores are arguments.

Thus their "efficient" compiler that they parallelize the parser to find an unambiguous parse.

I don't know. This seems like it would be super fun to write and to play with, and that there are probably some really cool new things being discovered that, when matured and properly integrated, may be workable into a usable programming language.