No, it really hasn't. It's a house of cards built inside a house of cards and it is fraught with peril.
I originally wrote the "Please don't virtualize" sticky in response to people trying to do stupid tricks with RDM, nonredundant datastore drives (which is IIRC what this post was originally about), and PCIe passthru on old janky pre-Sandy-Bridge platforms.
I quickly followed it up with the "guide to not losing your data" sticky, because it's really not a "this can never be made to work" but really much more of a "you need to think this through and be so careful in your choices and testing and it isn't guaranteed to work".
My posts all predated the iX blog post by like two years so it was kind of a rotten posting on their part as it added a bunch of confusion to the issue. Their developer team was basically like "oh, but, like, we do this all the time and it's fine." The trick being that there are things you can do in a development environment where you do not truly care about the safety and longevity prospects of your data. Yes you can just cram some VMDK's on RAID0 ESXi datastores and then make them into a RAIDZ, but you won't be getting health monitoring through FreeNAS and you can lose your pool because of it.
So, basically, yes, you can virtualize FreeNAS, but trying to do so with arbitrary hardware can end in disaster. If you pick a nice X9 or newer Supermicro board (almost all of which do PCIe passthru fine) and do PCIe passthru of the Intel SATA or an LSI HBA, and you build your system like you would a bare metal FreeNAS, you have VERY good chances of being able to make it work, but you may still have a few bumps and bruises along the way. The "guide to not losing your data" is literally the golden path to success I've used here for many years, and is a better guide than the iX blog post which seems to try to take tangents to cover random possibilities that might not be such good ideas.