Unique sequence vs PCR duplicates

I apologize in advance if this is an obvious and dumb question.

What is the PRECISE difference between a PCR duplicate
versus
A unique sequence?

I tried looking for it all over, but was unable to obtain a clear explanation. Now, my thought process is something like this.

A PCR duplicate is something which is right off from the NGS instrument. i.e., you look for PCR duplicates from your Raw Fastq file.
Whereas a Unique sequence is …?

If you have a sequence as

Read 1 - ATCGCCCTA
Read 2 - ATCGCCCT

Would these be TWO unique sequences? If so, what is the algorithmic THRESHOLD in Mothur Unique.seqs to call two sequences as Unique versus Non-Unique?
Thanks for the favor of a reply.

For two reads to be identical, they must have the same length and sequence. So the two reads you presented would be different by unique.seqs.