What?
Been there since pre-8.2? Who's been laughing at us flailing around with locked-down multi-VI XNodes (or multi-multi-VI polymorphics) for years? This seems to answer 99% of my (and everyone elses) requests for a simple type-adapting VI. Looks like a very simple XNode without any type-checking (e.g. wire a Boolean to the Delay input and it generates the code but breaks the calling VI). But if the programmer can be trusted, it looks incredibly useful. I would presume it has the same caveats as using an XNode, but if it's now "released" (thanks jkodosky, who I imagine has some mandate to do so) it would be good to know anything to be aware of.
BTW, this is certainly not the same as the buggy, non-supported Generic VI that was dangled in front of us some time ago.