This is obvious though

currently, you might test a drug on mice, then on primates, and finally on humans (as an example). It would be faster to skip the early bits and go straight to human testing.

...but that is very, very, very wrong. Science of course doesn't care about right and wrong, nor does it care if you "believe" in it, which is the beautiful thing about science

so a scientifically sound experiment is a scientifically sound experiment regardless of ethical considerations. (Which does not mean we should be doing it of course!)

Now, taking a step back, maybe you're right that, in the long run, throwing ethics out the window would actually slow things down, as it would (rightfully) cause backlash. But that's getting into a whole "sociology of science" discussion.