CD vs FLAC stored on flash drive vs streaming

That's generally what I've heard. My assumption is that the digital file via a CD player or via a flash drive has a shorter signal path to traverse and thus less chance to pick up analog noise along the way compared to the longer signal path that the streaming signal takes. Noise might include common-mode noise such as RF interference or  and differential-mode noise as typically caused by ground currents.

I have come to accept, "sounds good enough to be enjoyable" as my end point.  I cannot reliably determine a difference between my ripped CDs that are stored in my server's SSD and the same tunes played through Tidal (non MQA).  I get more of a noticeable change from a cable swap.  Once I reached the current level of digital playback equipment that I have, the convenience of having no physical media outweighed any small sonic differences.  I do plan to try Qobuz soon and look forward to hearing if there are significant sonic differences between that and Tidal.

At least with my system, I've found that the original source material largely rules.

Most of CDPs operate in real time.  For short scratches along the track (shorter than 4mm) they supply data from error correction.  For 4-8mm they interpolate lost data and above that they lose data (gaps).  Ripping is not real time process.  Computer can read each sector infinite number of times to obtain proper checksum.  I placed the limit of 200 attempts.  Theoretically ripped CDs should sound same or better than original, but for most CDs in decent condition it would be very difficult to hear the difference.  Connection to DAC or preamp is different story.  

All true. Copied CDs usually DO sound better than the mass produced originals especially on gold. I rip all CDs for convenience and with my playback setup I cannot ascertain any difference, except maybe the BluNode.