Well, the speed regressions are mild and expected, and will be addressed "after 3.0 is released", which means "starting now", I guess. The post you've linked to is a bit misleading. Sure, the new ruby interpreter is expected to be faster, the post claiming "3 times" as fast. But Python is already 5 or 6 times faster than Ruby since Ruby 1.8 is a purely interpreted, whereas Python has always be a VM. The post implies that Python 2.6 and Ruby 1.8 both have comparable speed, which is not the case at all.
Actually, what the poster is missing is that 3.0 is not a whiz-bang, shiny, gotta-upgrade-now sort of release and was never supposed to be. Use it for green field projects, and if desired, continue on with 2.6 and 2.7, and possibly 2.8 and 2.9 for your existing code.