Space ship speeds are unlikely to keep ever increasing. In the limit you can’t do much better than turning part of the ships mass into energy optimally, eg via antimatter annihilation or Hawking radiation, unless you already have infrastructure in place to transfer energy to the ship that is not part of the ship’s mass, eg lots of lasers.
Mass drivers don't need to be a linear ramp, portions can be circular
It would work better for smaller, unmanned craft, especially when you consider g force limitations
NPP is only theoretical, and still has major problems such as finding a material that can withstand a nuclear detonation at point blank range. Mass drivers have been proven to work, albeit at a smaller scale
IIRC, Dyson proposed using a thin layer of oil on the surface of the pusher plate that would get vaporized with each shot, but would prevent the plate from ablating away. This effect was discovered by accident during nuclear testing when oil contamination on metal surfaces in close proximity to the explosion would protect them.
Of course, depending on how much oil you consume for each shot, you will degrade your effective specific impulse - I'm not sure by how much though.
The other issue which you can't really get around is thermal, that plate is going to get hot so you'll have to give it time to radiate heat away between shots. This may be less of a concern for an interstellar Orion since the travel times are so long anyway, low average thrust may not matter too much.
Mass drivers on a asteroids or the Moon could change the game