Do note that duplicate detection in DT is weird and doesn’t work the way most would think, but works The One True Way that DT thinks is right. It is based on the rendered version of documents like markdown, rather than the source.
One of many of my annoyances with DT.
I agree with @ibuys ‘ recommendations above, develop your systems and mental models, then select tools that can help implement those systems.
As has been discussed before, Devon Technologies is one of the Henry Fords of “artisanal” software development. They produce software they think works the way you want, and generally are not open to feedback about what people actually need.
People have also experienced unexplained data loss (files being emptied of their contents). There were no indications from DT that it had corrupted the files, and in some cases the corrupt files were discovered after old backups had been overwritten, leaving no way to recover.
For me, DT has trust level 1 out of 10. I only keep things in it (or indexed by it) that I can recreate from other sources if needed. For instance, I store invoices for autographs I’ve bought to which I’ve added pictures of the autographs. I could recreate those from the auction’s website or from paper records. On the other hand, I would never let DT touch my source code, figures, writing, etc. for my research.