This is a problem with the turbos. What happens is that the cruise is designed to maintain the constant speed very well - but the turbo introduces a power curve that is nonlinear over time for a given throttle opening due to turbo lag. At speeds ~55 and below, the turbo does not produce a lot of power, so the 'hunting' is generally not noticeable. However at the 65 range, as the cruise opens the throttle to correct speed, it will continue to open the throttle until the speed comes up - but now the turbo starts winding up and boost increases, developing much more power and the cruise overshoots. What is needed is actually a system that analyzes second order effects - i.e., one that measures and acts on acceleration signals, not just speed signals.

Unfortunately there is no design adjustment for time constant for the system. What I recommend is that if you have a programmable boost controller, you use a low boost setting when using the cruise. Another option would be to insert an orifice in the vacuum line to the cruise actuator. This would effectively change the response time constant. If you experiment with different orifice sizes, you should be able to get the system to respond smoothly in a given speed range, however, you will likely get a slightly wider variation in speed - i.e., the speed will droop more going up hills and overshoot more on downhill roads.