I have used a long-term mount of CLU without issue. In my case it was the CPU, not the TIM that degraded from abuse. CLU will exhibit some of the behavior noted in the thread linked by Yuriman, but in my case, it has never resulted in massive "clumps" that caused problems.
It may be that using too large a quantity or allowing air bubbles to form during initial application of the TIM will lead to long-term issues. I use lower-than-normal amounts during application. You can spread tiny droplets of the stuff really thin.
Well, it just begs the question as to whether use of CLU on bare die is as good an idea as we thought.
What I'd read about liquid metal or even the Indigo pads: it will indeed make an amalgam with the copper. With an IHS either with nickel-plate or copper, it does the same thing. It might require some patience separating a sink from an IHS -- I'm not sure.
But it wouldn't form an amalgam with the silly-cone die.
So maybe uneven application is a risk, or when the CLU "sets" it can leave "inadequate" spots.
And I just stumbled in on this thread! I acquired a 3570K and was thinking to make preparations with a razor-blade. I think I'll shelve that idea for a bit . . . The IB wasn't the focus of my hardware-coveting, so I can just use it with lesser ambitions.