So how can a great system solve less than great recordings

"Fidelity by the way means truth. High fidelity means highly truthful. Does not mean good. Truth ain’t always pleasant to hear now, is it?"

Dead on MC. And I think the current take away from this discussion is clear - no system can make a lousy recording sound good and any system that does will likely make truly good recordings sound bad.

But what appears to be missing from this discussion is mention of all that happens AFTER the recording is finished. Great recordings poorly mastered sound terrible! I just found a great copy of Neil Young's After The Gold Rush that is a true demo disc. I've had 3 other copies and none of them sounded remotely as good as this one. If you only heard a bad one, which if you consider the odds is fairly likely, then you'd think it wasn't a great recording, which it is.

The fact is, audiophiles such as ourselves need the right records to play first, then we can make the choices about our systems necessary to make those records sound the way they should. That is, IMO, the way forward in audio. And that's why analog is so special! 

If you're interested in more of I have a lot of content on this very subject here:

https://www.thebrokenrecord.net/category/mastering/

OP, good topic. To restate your points:
1. High fidelity systems currently don't make up for problems in the source material. Rather, HiFi systems can emphasize problems in production and mastering.
2. Remastering has not solved all the problems in the original production and mastering. 

I wonder if AI upscaling/upsampling in audio is the next step in remastering. I believe that we could begin to see an AI upscaler/upsampler improve poor recordings in ways that past remastering efforts hasn't. Also, I believe that an AI upscaler/upsampler could fit into an existing HiFi system. 

The strides in AI upscaling in the video space could show us the possibilities in the audio space. We've had video upscalers in TV's and receivers for a while, and older versions of upscaling have been impressive for past TV resolution standards. But with the release of 4k TVs (and soon 8k TVs), upscalers haven't kept up. Most receiver manufacturers seem to have 4k upscaling limited to taking 1080i up to 4k. Anything lower than 1080i is SOL. Many TVs today don't even have upscaling anymore. And the quality among the 4k upscalers in receivers and TV is all over the place. 

The best 4k upscaler available right now is from Nvidia, in their Shield TV streamer. It combines traditional upscaling tech with their own AI tech. Nvidia, which was a computer graphics card company, found itself on the right side of history when graphics cards became the best tools for AI research. And their Shield TV streamer arguably started as a streaming box to allow Nvidia graphics card owners to stream their games across their home networks, from the computer in the office to their TV. At the intersection of AI and consumer streamers, Nvidia is the clear leader in home video upscaling. Here's an example of what Nvidia upscaling can do. Make sure to view it on a high res screen:
https://www.nvidia.com/en-us/shield/sliders/ai-upscaling/

I know that the HiFi space has some companies that dabble in upsampling. My understanding is that the quality varies widely and that current upsampling offers another audio flavor rather than something so convincing that we leave the upsampling button toggled on. 

But some Nvidia level upsampling could arrive in the HiFi space, one day. The market for high fidelity audio lags behind the market for home cinema, and, as a result, the biggest companies have largely left the HiFi market underserved. Also, due to the software requirements for upsampling, the more boutique a company, the less likely that company is to maintain it's own software and make advancements in audio upsampling. (NAD and Roon strike me and the best pieces of software in HiFi right now.) Here are my takes on where impressive upsampling may one day come from:
1. The home theater space expands its interest into quality audio. (Nah... who am I kidding.)
2. FPGA-based DACs provide the inroad to opensource software, engaging a community of software hobbyists in audio. (Wouldn't it be great if PS Audio or Chord opened up their software to their communities? I should make an effort to pitch this idea to these companies. As an analog, the video game Skyrim was one of the biggest video games of the last decade. Part of the reason for Skyrim's success was because it's smaller developer, Bethesda, opened up its code to the gaming community for modding. That approach turned out to be wildly successful, and allowed Bethesda to compete with the bigger developers like EA and Ubisoft.)
3. The big streaming platforms (in both video and audio) perform the upsampling to their streams. I see this as the endgame. But just like Apple Music and Spotify waited until now to release a CD quality tier, these big players likely won't provide upsampling until they are forced to, or until they can easily check that marketing box. 

All I know is good quality mono recordings still sound like the performers are playing live in my room with my Ohm Walsh speaker setup. This setup has opened me up to so many older recordings that I would probably not pay much attention to otherwise. The best for bringing old mono recordings to life. Example: remastered Muddy Waters CD recordings originally from the 50s. He is there in the room. Not too shabby with stereo either. Coherency and carefully managed wide dispersion in the unique Ohm Walsh CLS design pays off!

https://photos.app.goo.gl/9mMf2dXS51F1oGH1A

I have always been very careful to navigate my evolving system towards greater fidelity but to stay slightly back from ultra detailed / revealing. One step to far and many recordings sound bad... there are at least two kinds of bad... lack of dynamics and noisy harsh. I am very turned off by noisy / harsh. I want to listen to the music not the system... I want an emotional connection with the music, not to be sure I clearly hear the 2nd violinist move his foot. 
Realizing this was critical on my turn to a far greater musically satisfying system. You can get caught up in perpetually evaluating sound and forget what the point was. So I am pretty sure @Mahler123 and I would react completely differently to equipment. We have two completely different design goals.

Just celebrated my 50th year of audiophilia.
With some luck, dedication and, gratefully, this Forum I have evolved my analog only system to a place where there is no such thing as an unlistenable record. All are enjoyable... some more than others.
The most significant evolution came with the addition of my AGD Audions. They were brought to my attention by several Forum member’s rave comments. Then, checking the reviewers, they constantly use the rare phrase “best regardless of price”. I was inspired to order a pair. Unbelievable! Not only the “right on” tone and SQ, I hear something never heard. I hear the musician’s phrasing. Music is made up of phrases tied together. To clearly hear the beginning, end and nuance-the humanness! I always liked Ella but only now do I fully understand her greatness hearing her phrasing. And this critical essence of music is evident in every record. For sure, my speakers, front end, cables and power all contribute. But the real change came with AGD.