I spend most of my days working at a studio that primarily relies on Maya, Renderman, Houdini, and Clarisse iFX (among others).
That said, I can see Blender becoming a more integral part of our pipeline over time. Everything you mention is true, but some of it could also have been said of things like Softimage and Shake. I remember when we switched from Softimage to Maya back at ILM. It was a painful transition, but it was accomplished in less than a few years.
I don’t see anyone (other than Blender enthusiasts) seriously suggesting that Blender will supplant Maya any time soon. But that doesn’t mean that it won’t start to find a foothold now that 2.8 has managed to kill a large part of the “non industry standard stink” that the previous versions had.
Here are some reasons why I think Blender will actually start to make inroads faster than your article would suggest (inroads means some adoption, not becoming the primary application used by studios):
1) Samiam suggests that open source is a liability. I would disagree. Open source is actually a benefit. Most major VFX studios actually rely quite heavily on open source software: Linux. Having the ability to customize the specific functions of your 3D pipeline is part and parcel of a larger studio. Often it involves writing glue code between commercial packages, but it can also benefit from modifying the actual functionality of the tools themselves. Specifically referring to the alembic exporter in Blender. The deficiencies that it has in this regard is theoretically possible to fix at a studio level (Not super likely for the reasons pointed out in the article – other packages already in use do it better). But all it takes is one studio making that change and upstreaming the changes for all studios to benefit. This is a snowball kind of effect. As smaller changes get made, adoption goes up. As adoption goes up, more interest in making changes (and upstreaming them) happens. This drives even more adoption. Note: I am not saying this WILL happen. Just that it can (and there is a precedent for it in the rest of the open source world as open source projects matured they became as good as or better than commercial systems in SOME cases.)
2) Finding professional level artists who are familiar with Blender. This may also start to change over time. As smaller studios are much more price sensitive (and are less reliant on older, established pipelines) they may begin to adopt Blender for some of their workflow (now that it is so much easier to start using without massive re-training). Once you have smaller studios showing some demand for this package, you will get a few more artists who are comfortable with Blender. These artists will then start moving to larger studios and occasionally bring with them the skills and/or desire to use it at their new jobs. Again, to be clear, I am not suggesting that this WILL happen, nor that it will be widespread. But it is a way that new tools make their way organically from one market (small studios to larger ones). This is only feasible because 2.8 has a significantly lower barrier to entry (ease of use) than before. The quality of the tool (for specific, limited pipelines) does matter. It is a similar reason why ZBrush, even with its non-intuitive interface, took the market by storm (over a fair number of years actually). It is a high quality tool. Blender is starting to become a high quality tool in some areas. This will help with its adoption on the fringes. And things on the fringes, assuming the quality of the tool encourages it, will start to move to the mainstream.
Ultimately I think the question of whether Blender will supplant Maya is really only interesting to hardcore Blender enthusiasts. Professional artists will choose the tools that:
a) are available to them
b) help them get their work done faster, better, and cheaper
Blender has long solved A. It runs on every platform and has zero cost. But it suffered quite a bit when it came to B (though, in some cases, more as a perception issue than a real one). With 2.8 they are starting to address B. As they start to provide small solutions to real problems that are better or different than other packages, people will start to use it for that. Even if it is just for a few minutes a day for a specific, well defined purpose. But this will drive familiarity and acceptance. As long as the Blender developers (who are AMAZING BTW) continue now to focus on B (even if some of their changes infuriate older Blender fanatics) they will find the adoption rates increasing. Will it supplant Maya? Who cares. But I think it will find a home in the studio environment.