Not sure about Spotify and Tidal as I don't use any streaming services myself, but if you compare a properly encoded flac file with a properly encoded 320k mp3 (using one of the modern lame presets), encoded from the same source, played on the same hardware, same player, etc I'd be massively suprised if you could notice a difference immediately under double blind test conditions. Under these conditions I can't notice any difference until I get to very low bitrates like 128kbps and even then, it's because I know the tracks very well and know what bits of the music will sound different with lots of compression.
Put simply, all else being equal, listening to 192kbps mp3 and above (or better still, one of the VBR presets) is going to be virtually indistinguishable to flac for 99% of people under normal listening conditions with high quality gear. It's been shown time and time again, with audio hardware and testing conditions superior to anything anyone is using here.
Despite this the vast majority of my music is encoded in flac, but that is only because storage is cheap and having a lossless audio files provides better flexibility if I need to encode into different formats in the future. When I see people using massive flac files, or worst still, 192kbps/24bit lossless files with their portable gear, I always laugh. I also find it annoying when people call mp3 compression garbage (or any other lossy encoder for that matter). Purely from a technical standpoint it's pretty damn impressive, especially when you learn about the psychoacoustics part of it. It's not mp3 that is garbage, but more likely your flawed testing conditions or poorly encoded music. Using youtube (for example) as a basis for calling lossy compression garbage is stupid, as youtube does way worse things to the audio than using a lossy encoder (AAC in this case). If you're going to call something garbage, you should at least ensure everything else in the chain is the same.