Umesh Vazirani's
quantum computing work at
Berkeley may soon be able to disprove it, though. I believe they're at a stage now where
(
P = NP || quantum computers can work exponentially faster than
DTMs ).
But I'm ignorant, so don't quote me on that.