Is friction always constant? Can the norm of the friction force change with time? Notice I said the norm which is in Newtons, can it change with time? Sorry if the question seems a bit confusing but I've done my best to explain. 
 A: I'll assume that you're talking about friction between two solid surfaces sliding over each other. Other comments and answers have pointed out the friction is velocity dependant in gases and liquids, though I would call this viscous drag rather than friction.
When you first learn about friction you're usually taught that the frictional force is given by:
$$ F = \mu L $$
where $L$ is the load (normal force) and $\mu$ is a constant called the coefficient of friction. Your question then amounts to asking whether $\mu$ is really a constant or whether it depends on sliding velocity. The easy answer is to point you to a Google image search for friction coefficient velocity. This finds lots of experimentally measured graphs of $\mu$ against sliding velocity and you'll see $\mu$ is indeed dependant on velocity and isn't a constant.
Friction is an emergent property. It happens because when you touch the two surfaces together high points (asperities) on the two surfaces come into contact and bond due to the same sort of forces that hold solids together. To slide the surfaces you have to break these bonds, and that takes energy. The energy dissipation per unit distance of sliding gives the force.
The approximate equation comes about because the area of contact of the asperities is approximately proportional to the load. As you press harder you deform the asperities so they flatten out and form a larger contact patch. As a rough guide you'd expect the pressure at the contact point, i.e. the normal force divided by the real area of contact, to be roughly equal to the yield stress of the solid.
The dependance of $\mu$ on velocity is complicated and can go up or down depending on the system being studied. The rate at which asperities touch and bond depends on all sorts of things so it's hard to say anything definitive about it. All we can say is that there's no reason to expect the rate to be independant of velocity so we shouldn't be surprised to find that it isn't, and also that it varies from system to system..
A: (Air) friction on a pendulum definitely changes with time - because even in a simple model, it is a function of velocity.  
A simple expression for the friction would have the force proportional to
$$F = av + bv^2 $$ 
which is obviously a function of time (since $v$ is a function of time).
See (C2) of The pendulum - rich physics from a simple system for more details than you ever wanted.
