PBRs are pretty much slanted towards game production, which is where they come from. Physically based rendering is only new to games. Every other aspect of 3D has been approaching materials partially or entirely from a physical standpoint for more than a decade. For instance, those of us who have been following Bagginsbill since the beginning have been talking about conservation of energy since before Poser supported GC. Really, the only area besides gaming where people were still faking the funk was film, and while Renderman makes faking the funk efficient and useful, renderers like Vray have been pushing film towards more accurate raytracing for a long time. As far as I can tell, arch vis, product rendering, prototyping, etc. has been going the full realism, unbiased route since I started using Poser in 2000 or so.
Most PBR materials aren't nearly as accurate as, say, custom materials in Luxrender or Octane. A cross renderer standard is _invaluable_. Much like HTML/JS makes it great for us to develop web apps for various platforms, PBRs are great for making materials that can translate to other renderers. But like HTML/JS, the vagaries of different platforms matter _hugely_.
Iray has a nasty edge artifact when you apply blurring to reflections, and only supports scattering as a volumetric effect that works like water, gems, or fog. A lot of people (including DAZ, AFAIK) incorrectly use translucency instead for SSS materials. Translucency is supposed to be for single poly surfaces imitating things like curtains or lampshades, and used incorrectly it has lots of weaknesses in various lighting situations. It's possible to use the scattering for skin, eyes, and teeth, but it's a lot harder than with other systems. Sort of how blurred reflections in Iray are best handled with light blurring in the settings and lots of fine bumps in normal maps. Also, you need to set your lights to test your materials, and Iray's lighting gives _entirely_ different results depending on the film and exposure settings. To even start testing materials, you have to at least have a clue about fStops and white levels. Maybe that's common knowledge, but I _really_ had to research to build a solid test environment in Iray. Because the default settings were _very_ extreme, and I had to change a lot to get them to a sort of normal, point and shoot camera with an average ISO level. It's awesome power to have, and I _so_ appreciate it, but it adds to the variables you have to consider when making and testing materials.
Cycles, on the other hand, handles reflection blurring and regular BSSRDF just fine, in addition to having volumetric absorption and scattering (things you can't exactly play with using the PhysicalSurface). But its caustics are somewhat limited and inefficient, and can be slow and hard to control. Cycles doesn't support true diffraction on its own (think the rainbow sparkle of diamonds, which Iray handles _beautifully_ and Cycles can only have with a fake), and even regular glass shading can take _ages_ to render with full caustics in certain environments. There are several fake caustic shaders for Cycles out there, the best being comprehensive, proprietary, and reasonably expensive. I really, really wish that the Cycles team gave caustics a higher priority, but the some in the community are fighting to keep caustics unimportant and not very good. It also doesn't have film settings, or other more physical considerations, because that's just not what it's made for. Even though Iray is made by nVidia, and therefore made mostly with games in mind, it's even more suited for stills than Cycles is. Cycles was largely designed for film/video, so it's weaker on realism while having all _kinds_ of features for stylization.
For instance, before Cycles supported SSS, you could use the raytracing nodes and fake Beer's law. What that gave you was a kind of internal depth map. Even with SSS and volumetrics, you can't as easily take depth and use it to, say, control a whole rainbow of colors (see the ColorRamp node in Poser and Blender) based on depth. You can get the same effect using plain volumetrics, but it's not as easy to control.
Both Iray and Cycles are great renderers. It's easier to get them to perform their best if you know what weaknesses to compensate for and what strengths to play to.
So sure, you can certainly focus the PhysicalSurface and ignore all the particulars of different systems. IMHO, PBR's in general are kind of bad for skin and natural items. I've seen lots of attempts at PBR skin, in various renderers, and I've not seen one that looks more like skin and less like painted wood or some other substance to me. The PhysicalSurface doesn't have any features that support making velvet, satin, terrycloth, or any other complex cloth shader. For instance, there's no anisotropic factor in the PhysicalSurface node, Firefly doesn't have an anisotropic reflection node, and plugging Superfly's anisotropic reflection node into another specular node is just completely inaccurate. Since PBR materials are from gaming, they focus on all the materials that are central to games, and therefore the kind of materials that are cool to your stereotypical 15 year old boy. So plastic, metal, glass, grunge, and such are great and covered in detail. Nylons, velvet with gold embroidery, silks, even lifelike and detailed skin, not so much.
To me, PBRs are weak in exactly the areas that tend to be important in our community. We have a gender balance and focus on rendering female figures in fancy gowns that's unusual for all aspects of the 3D community I know of.
While gaming is a part of the Blender community, more of the community does stills and movies. I've literally never looked at material tutorials for games rather than stills or video, and I've been using Blender since the 2.4x days. I'm not interested in the gaming industry, so I've avoided those tutorials and never felt the loss. I'm sorry, but I can't think of a particular piece of advice for finding Cycles material tutorials. I just haven't had much of an issue, so I don't know what I've done right, so to speak.
That said, PBRs are the rage now, and I'm not finding anyone with what _I_ would consider great advice on velvet or skin (good yes, great, no). I have to admit, though, I've been studying layered skin shaders since P6 days (the main paper I read was for MentalRay/Maya, IIRC, and it went over the physics of each layer and aspect of skin in detail), so my standard on skin shaders is pretty high and probably pretty esoteric. I'm not big on, say, using cyan or green to counteract burned in artifacts (I have better solutions), and I'm much more into the combining and balancing dermal scattering (in the skin, yellowish, pale, and waxy) and subdermal scattering (under the skin, red, and very dense) than pretty much everyone I've come across except that original skin shading paper I read. And the velvet I developed for Firefly involves quite a lot of playing around (including procedural textures) to fake the effect of a volumetric surface. The best alternatives I've seen use actual volumetric shaders. So you might find the information out there on skin and velvet more valuable than I do.
