At first I wondered why you'd assume the latter – certainly something like a tiny rotary encoder is a simpler lower-tech solution than a MEMS inclinometer. But these days I'm not actually so sure.

They're using a pretty nifty differential Hall magnetic encoder, like this one:

https://www.novosns.com/en/hall-angle-sensor-4010

There's a fun Hackaday video using the same type of part in a rotary knob + LCD control.

https://www.youtube.com/watch?v=zIrAe23f8sg

Without checking the catalog, I would assume that a MEMS inclinometer is much cheaper in 2025 than an absolute position encoder.

Edit: catalog confirms.