Then the answer is just the speed of light in both cases, and that's really the crux of the problem of special relativity and that the speed of light is constant in all references frames. Then you work backwards to see that the observer that is moving must have time dilation effects.
I also liked the geometrical approach to Minkowski space-time a lot better than just chewing through formulas -- like Taylor and Wheeler's Space-Time Physics book. Drawing light cones and various grids and curves of constant space or time measurements relative to different observers made it all a lot more concrete to me.
The other thing that confuses people is presenting things from an 'independent' perspective in the diagram.
At what speed does the point of contact move, relative to the reference frame of either bar, or to a static external observer?
In that case, the "static observer" would see the point of contact moving at very nearly sqrt(2) times the speed of light (along the x-y diagonal). But that's okay, because no physical object (or bit of information) is being carried along at that intersection point. (If you put some little spindle at the intersection to be carried along, you'd suddenly have Problems.)
On the other hand, from the reference frame of one moving bar, the intersection point appears to be moving almost infinitely fast! If the "static observer" measured speeds u and v for the two rods (as fractions of light speed, so u and v are both close to 1), then an observer moving with one rod will see the intersection point move at speed v/sqrt(1-u^2). If we write u = 1 - ε for some very small ε, then the intersection point's speed in this frame is roughly v/sqrt(2ε). (The intuitive reason for this result is interesting: an observer on one moving rod will consider the other rod to be almost parallel, at a very small angle (rather than perpendicular), and still rushing toward her at nearly the speed of light.)
What's your field?
Tom Moore's Six Ideas That Shaped Physics textbook (Unit R) does a very nice job of this. But here's a really bargain-basement handout that I made many years ago to explain the idea before I started using Moore's text for this. (I'd probably do it a little differently now.) http://www.slimy.com/~steuard/teaching/classes/spacetime.pdf
My discussion of velocity addition there (at the bottom of page three out of four in that PDF) is much too terse for this context, I'm afraid. But if you draw my example there carefully and measure in the tilted coordinates as described earlier, you'll start to see why the leftmost observer still measures the rightmost observer moving less than the speed of light.