In fact, for almost all real machines, the total work increases when you use the machine! That's because almost all machines introduce some frictional forces which waste some of the input energy. For a simple spanner, that could include deforming the nut being tightened as well as air friction while rotating the spanner (although this is absent in vacuum and negligible for virtually all applications in atmosphere). The spanner itself might also undergo small plastic deformation as well (at the point where it contacts the nut, or even in the handle, under large stresses).
Now, ignoring the pedantry above, let's think about a counterfactual universe in which simple machines could change the work performed. There are two possibilities, and without loss of generality, we'll just consider the first one: work done decreases with leverage. So, let's say that a 20 cm spanner requires less work than a 10 cm spanner under the same conditions (tightening/loosening a nut).
What this implies is that we can build machines that use less fuel simply by increasing their leverage! So, for instance, a bicyclist would be able to travel much further simply by riding in a higher gear. In fact, high gears would always be favored. And the same is true for automobiles! If the work reduction is proportional to the leverage, then we could probably double or triple the mileage of almost any powered vehicle simply by increasing the leverage used to move the vehicle! The transmission of a car might end up being many times larger than the engine, simply because adding leverage is more cost-effective than adding power.
But what about going the opposite direction? It means that shorter lever arms do more work. It means that shorter people do proportionally more work than taller people just to move about every day! Scaling down to insect sizes, it means that instead of carrying hundreds of times their weight, ants and spiders would get exhausted just walking around! Of course, they might compensate for this by having extremely long legs, possibly cantilevered at the joints, just so they could walk! But it would be incomprehensible to imagine grasshoppers jumping dozens of times their body height under these conditions! They might burn up a whole day's worth of calories doing that!
Go even smaller still and imagine the flagellum of a bacterium. This is one of the smallest machines we know of, and so, it would have the largest proportion of work expenditure due to its lever arms being literally molecular scale! It might not even have sufficient metabolism to operate it. Of course, cilia and flagella are rampant across the microbial world, and so we know that tiny levers/motors/machines work just as efficiently as large ones (or perhaps more efficiently, given the internal frictions mentioned above).
And if it turns out that longer levers require more work, then just reverse the argument above w.r.t. scale. ;)