But I guess you could cheat by having multiple engines along the way that all accelerate in lockstep.
But what if the thing I push on is a quantum particle? Does this same thing happen at the smallest scales? If one end of a quark is pushed on does the other end move instantaneously or is there a small(!) delay?
Probably the answer is just "that's not how quarks work" but I've always been curious.
For subatomic particles, the most intuitive way to think about things is to adopt the "fields are real" mindset. Here fields are the underlying reality, and particles are just a pattern of waves excited in the fields. Disturbances in all fundamental fields we've discovered propagate at the speed of light, and we have pretty solid reasons for believing no future discovery will contradict that, as it would break causality in a fundamental way.
Quarks don't have an "other end." To the best of our knowledge, particles are points.
This is a classical picture, but the quantum picture is similar: evolution is generated by a local Hamiltonian constructed out of field operators attached to every point of space.
So, both classically and quantumly, relativity demands the existence of fields filling space to propagate causal influences at finite speed.
The speed of light in a vacuum happens to be the best representation of the maximum speed of causality
Which also makes more sense why you can't do things like travel faster than light (your effect would precede the cause), and why two protons going past each other in opposite directions don't violate this law