I have a Vizio P75-C1. And I'd agree that the upscaling from 720p or less isn't all that great. But I think it does a nice job upscaling 1080p content--though I admit to not being able to compare it with Sony or Samsung. But I don't have any "complaints" with the 1080p upscaling--it's the 720p or less that has visible room for improvement. So in that regard, for non-native 1080p or 4k content (typically, this is content from my DirecTV box), I actually DO use an AVR to do the upscaling to 1080p (or 1080p/24 if it's a movie).
I know the article suggests against this approach, but for my chain, I actually think it does quite a nice job. And I rarely notice any artifacts or general "issues". I'm quite pleased with it. I have a 2013 Pioneer SC-1323-K--which is a pretty decent AVR--retailed at about $1300 (though I paid about half that when I bought it in January 2014). Incidentally, it also has 4k upscaling--but not all the way to 60 frames. At that time, 4k was in its infancy and was at 30 frames per second. I've watched movies using the receiver to upscale to 1080p/24 and to 4k/24. With the latter, the Vizio doesn't upscale further (at least that's my understanding). So this would be an instance where only one device did the upscaling (in this case, the Pioneer AVR is doing the whole enchilada)--which the article suggests is the preferred method. But at least in my chain, I respectfully disagree. To my eye, having the AVR do the 1080p/24 upscaling, and letting the Vizio do the rest gives a more pleasing picture experience. Similarly, if I take the AVR out of the chain and let the Vizio handle the entire upscaling process, it gives an inferior picture, imo. It's subtle--and I don't know if I could specifically identify what's different. It just seems better to my eye to let the Vizio do the 1080p to 4k part, and the Pioneer AVR do the 720p/1080i to 1080p/1080p24 part, even if it introduces two upscalers in the chain.
So I wouldn't dismiss using a high quality AVR as an upscaler. I've used a colorimeter and performed various calibration tests through a direct chain and through the AVR upconversion chain, and the results are basically identical. So as far as I can tell, the AVR isn't introducing any residual impact on the grayscale, black levels, brightness, or color readings. It's just doing partial upscaling. And at least for my situation, I'm very satisfied with the results.