Why does uniq seem to keep some consecutive identical lines?

TL;DR: The lines have different whitespace (possibly spaces) at the end.

This happens when you have lines that look the same but are are actually different due to characters that don't display in your terminal, usually at the end. Often these are are trailing spaces (as fkraiem suggested) or inconsistent line terminators.

You might expect that starting the pipeline, as you do, with cut, would prevent this. It doesn't, though. cut uses a tab as its default delimiter. (Readers who wish to verify this behavior--and its relevance to having unexpected duplicate lines after uniq--can try cut -f 1 <<<$'foo\nfoo ' | uniq, which prints two lines.)

The solution in your case is probably to use something other than cut -f 1 to select the fields. In particular, if the fields are separated by spaces instead of tabs--whether by a single space or multiple spaces, and even if the number of spaces is different in different records--then you can use cut -d' ' -f 1 instead, specifying space as the delimiter character. Or you might not want to use cut at all, but instead use awk '{ print $1 }', which prints the first field, taking any sequence of consecutive spaces and tabs as the delimiter.

You could alternatively strip the trailing whitespace, though this makes your command even more complicated. One way to do that would be by piping your text through sed -E 's/[[:space:]]+$//' before it goes to uniq.

As a side note, if whatever command you ultimately use still ends up piping the output of sort directly to uniq, you might consider just using sort -u for that instead.