1) No, it's risk of silent corruption. No apparent failure, but a bit here or there different from what the stock machine would provide. With floating point, this could be nearly impossible to detect, without running the same data through multiple computers. I'm pretty sure IDC actually found this going on, with one of his older rigs.
2) With a Pentium or i3, it's not easy at all. With others, you still might need to raise voltage, and/or apply a beefier cooler. Then watch temps. Then, perform stress testing. If you get disk or memory errors, or BSODs, try to figure out if you went too high in Hz settings, need to adjust RAM timings, need to raise voltage somewhere, etc. etc..
In the last gaming build I made, the guy isn't so tech savvy, and wondered about doing it. I explained it as best I could, and he decided he'd rather have a GPU speed bump, instead (non-K CPU and non-Z mobo savings went straight to the video card). I think you're very out of touch with users who aren't the kind to browse tech forums and such 🙂.
Sure, it's far easier than back when we had to adjust things by DIP switches, and the CPUs had thermal protection, and no MCA, even. But, it's not without its effort, costs, and risks. If it's not something you've been doing, it can present a large learning curve, and time sink. If it's something you have done before, or want to do for the extra performance, then either way, it will be less of a hassle, and could possibly even be one of the reasons for the upgrade (to have a new rig to tinker with!).