I think that what he is saying is that the stylus tracks an arc across the record, which means it is at some point slowly moving forward (retarding, in terms of time), then at the top of the arc, it starts to retreat (speeding up). Both the slowing of time and the speeding up covers the entire side of the record and covers such a small number of degrees of arc (hence small fraction of one cycle of the record) that it has nothing to do with what can be perceived in terms of pitch change or timing.
TT speed
When I use a protractor to align the stylus I do the alignment at the inside, and then rotate the platter maybe 20 degree when I move the arm to the outside of the LP, or protractor.
On a linear tracking “arm” it would not need to rotate at all.
At 33-1/3, then 15 minutes would be about 500 rotations. And that 20 degrees would be a delay of 18th of a rotation.
So a 1 kHz tone would be about 0.11 Hz below 1000.
It is not much, but seems kind of interesting... maybe?