Finally had a chance to verify. Not just my imagination.
Off by almost exactly 1 degree, and narrow at the top face on all 4 sides, so it was not machine or workpiece alignment.
2-degree beam cone, focused at the bottom of the cut. Just enough to make a few fractions of a mm difference at a quarter-inch thickness.
Lasers are generally near-parallel generating a beam depending on the source type, but when used for cutting they are collimated to a sharp point to increase effective power in watts per area. Results in a conic beam, focused at the cutting plane.
Knew it in theory. First time I've actually seen it "in the wild".
When I corrected the inside cut size, I fixed the cutout wall angle to be normal to the workface, as best I can with handtools.
Measures as within a quarter-degree, which is good enough as the router bit gets its angle from the face, not the cut edge.
