What a nice idea. I expect you would have the issue that the bearings for the gyro always transfer some torque to it (how do navigational gyro's solve this?). But if you could solve that, then yes it would certainly work as a clock.

An adjustment to the unlocking timing would promote a release just as the pendulum pallet enters the arc of the path of the escape tooth instead of waiting as long as it does in the video. Furthermore, the arc of the pendulum is so small in this video that the impulse would end up being quite long but at least better balanced about the pendulum zero-crossing. Interestingly, I have seen other videos of clocks built with the chronometer escapement much lower on the pendulum, resulting in the pendulum pallet being completely clear of the escape tooth arc for a much larger portion of the pendulum arc. What really hurts this design as shown is the contact between the fixed pin and the weighted "arm" that is attached to the pendulum. Note, however, that Ken had to "add" that feature to show the function of that mechanism. If you look at the sketches in the background, you can see that Berthoud doesn't show how he keeps that "catch" from dropping too far. I would suspect there is a stop fixed on the pendulum rod.