Quick reply...
Not boring, not boring at all. And I wouldn't like a device which "does anything whatever I think" for two reasons. First, if technology exists to read a human mind, then there are far too many malicious ways to use it, including overwriting human mind; frightening opportunity for invasion of privacy. Second, if the device is capable of doing anything, then it's as scary as house elves in Harry-Potter-land: it is underestimated by the virtue of being a computer, it has unknown capabilities since its only limit is its owner's imagination, and it's capable of great destruction should owner wish it - or when the device turns out to have its own will, separate from the will of a human.
Just look at the influence Daneel Olivaw had on Foundation universe. He was benevolent, by his own choice or due to Asimov's laws, we don't know. But if the device doesn't obey Asimov's laws, or finds a way to disregard them, or is mistaken in some of the definitions, then the device may be, either deliberately or accidentally, malicious.