T> The requirement to type "will cite" has always bothered me a bit. 

And I agree: It is not ideal.

You can read the background in the FAQ:
https://git.savannah.gnu.org/cgit/parallel.git/tree/doc/citation-notice-faq.txt

T> I must use literally thousands of software packages in my (biosciences) research, and only only "GNU parallel" does this nag. For example, I use parallel to launch multiple copies of a file conversion software. It is not customary in my field to cite the file conversion software even though this is essential to the research at hand, and parallel is a (very welcome! much appreciated!) convenience with a time-saving contribution.

As mentioned in the FAQ:

"""
If you feel the benefit from using GNU Parallel is too small to
warrant a citation, then prove that by simply using another tool.
"""

There are even links to many alternatives in 'man parallel_alternatives'.

You may think it is a small thing, but it is really a matter of
long term survival. From the FAQ:

"""
> How important is the notice for the survival of GNU Parallel?

Citations is what indirectly funds maintaining GNU Parallel. Before
the citation notice was implemented hardly anyone cited GNU Parallel,
and that would not have been sustainable in the long term. Funding
development aligns well with "We will give back to the free software
community" and "To accelerate innovation and underpin operations".

Therefore it is more important to keep the notice than to be included
in different distributions. Specifically, it will be preferable to be
moved from Debian main to Debian non-free over having the notice
removed (and staying in main).

In other words: It is preferable having fewer users, who all know they
should cite, over having many users, who do not know they should cite.

This is because a long-term survival with funding is more important
than short-term gains in popularity that can be achieved by being
distributed as part of a distribution.

If the goal had been to get more users, then the license would have
been public domain.
"""

T> Many journals restrict citations to the <50 most pertinent

I note that you do not mention a single journal by name. I have yet
to come across a journal, that restricts citations to 50: 

Nature does not: https://www.nature.com/articles/s41598-020-65360-y
PLOS does not: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0220953
CELL does not: https://www.sciencedirect.com/science/article/pii/S009286741930220X
The Journal of Cell Biology does not: https://journals.biologists.com/jcs/pages/manuscript-prep#ref

But to give you the benefit of the doubt: let us assume you can
find a journal that has this restriction, then you could simply use
an alternative, if you do not want to cite GNU Parallel or pay.

If you neither want to help financing development by citing or by
paying, then you should contemplate whether GNU Parallel is really
the right tool for you. If you do not want to help, I will prefer
you use a different tool.

We see a similar situation with companies wanting to use GPL'ed
software, but not wanting to abide by the GPL: They like the
software, but dislike that they have to give the users freedom.
But you cannot have it both ways.

Maybe it is you Nadia Eghbal addresses in
https://www.slideshare.net/NadiaEghbal/consider-the-maintainer:

"Is it alright to compromise, or even deliberately ignore, the
happiness of maintainers so we that can enjoy free and open source
software?"