I will agree with your thought experiment, but how many cycles will it take to fail the aluminum bar? I agree that it will fail "some time", but the number of cycles it would take, make your thought experiment just that: a theoretical experiment. My apologies if I used the term incorrectly, but the "line" for aluminum is not the same as the "line" for steel; we agree there.
I get what you are saying, there is no minimum level that aluminum can stay under and live forever because it depends solely on cycles, but as long as you keep the stress per cycle under the allowable stress according to your S-N curve for the useful lifetime of whatever part/thing is made of aluminum you would never see aluminum fail. If this is used in a car, depending on the analysis, the aluminum driveshaft may not be what sets the lifetime. If the part is used on a firearm, the barrel may become unusable before the aluminum part fails, or the stock will fall apart, etc.. Whilst both scenarios are of items that can be rebuilt, it may or may not be an item that you would even rebuild if another major system failed. So using aluminum in targeted use is completely fine, so long as it is designed appropriately.
Edit: just realized that this in the art of bolt/lever/pump. My apologies for taking this off track.