I've also heard "it's better to finish something than to start something." At the same time, don't fall for the sunk cost fallacy. Finishing a dead idea only makes you feel good.
If a system has to operate the same way under rotation/reflection, that strongly constrains what the system can be doing. Think crystals or isotropic forces or, in the day job, what should happen if you negate a feature going into a neutral network then retrain. Should it matter from which perspective you look at a thing?
If a system is continuous then smooth changes to inputs should cause smooth changes to outputs. The system probably will be robust to small perturbations because they will result in small output changes.
https://en.m.wikipedia.org/wiki/Turbulence#Kolmogorov.27s_th... is something of an example.
What does this mean?
Time is a finite resource for everyone. Savor it. And, respect the time of others.
C'mon, we're all dying here.
Finishing is a skill (per Derek Yu).
All measuring devices have noise.
Symmetry and continuity impart incredibly strong constraints on systems.
A dollar isn't worth a dollar to everyone.
I am probably overreacting or missing something.
Most of the value can be had by delivering less than 100%.
Many problems cannot be solved, simply changed into other problems.
If my team succeeds then I succeed.