If you have a good DAC and a decent system, I think you should be able to hear the difference. For me, as an analogue person, I find the sound less fatiguing, less 'digital', with some high-res material. Of course, there are numerous factors that go in to this, in particular the mastering, but I do feel it is worth it and the MQA folding technology, in particular, is really useful here.
Notice how he did 13 comparisons in the first test, but only 7 in the last one? Notice how close the timestamps are? He simply selected randomly through a large number of tests, until he got a 100% success rate, and then posted the log, claiming his amazing hearing as the key. In fact he probably did dozens and dozens of runs, exploiting pure random chance to get a result that looked nice.
Secondly, as several people (including Arny Krueger, who came up with the jingling keys sample) showed how you could ace the test easily, by noticing defects in the software used, a tell-tale sign of switching between the two files, in other words simple cheating.
Even if the difference is actually audible, jingling keys is an utterly pathological test case, which doesn't transfer to actual music.
As I've said before in other threads and discussions, what you're hearing (improved quality) is not due to hi-res audio formats, but rather due to better mastering, because there is more care taken with releases targeted at audiophiles.
All the MQA folding does is fold ultrasonic content into lower bits, leading to a higher noise level when played on a non-MQA DAC. The fully MQA-decoded audio is equivalent to roughly 18-bit 96kHz audio (with lossy ultrasonics). Ironically, FLAC files in non-MQA 18-bit 96kHz audio take up less space than the MQA files, so it really is an utterly pointless and counterproductive format.
The downloadable files you linked to are interesting, but 1) you don't know if any different processing was performed, nor which sample rate converter settings they used, 2) your playback hardware may be non-optimal at the higher sample rates, leading to audible distortion, which can be perceived as "analog", and 3) the files are clearly labeled, which introduces bias. You need to properly blind test, for the comparison to have any value.
I have been using MQA for a while though I've let my Tidal subscription expire. There hasn't been a title encoded in MQA that sounded worse than the red book CD version of the same title. SACDs and other formats on the other hand have been virtually indistinguishable from the MQA version. In rare cases, the SACD version was better. This has been a totally unscientific assessment of about 100 titles, however. Nevertheless, I am convinced the SQ from Spotify would be vastly improved if MQA formats were streamed.