Yes, I've done the modification.
The bushes are an interference fit in the pulley bore, pressed in. There's a gap between the bushes, into which you can get a drift to knock them out, not difficult.
BUT…. a) use a brass drift, not a steel one, and b) for knocking out the first one file the end of the drift at an angle that provides a square seating against the end of the bush that's being driven out.
The problem I discovered is that the drift must inevitably be at an angle to the end of the bush being knocked out first as a result of it being pushed out of line by the bush that's not being knocked out ( if you get my drift – deliberate pun!! ), so a flat ended drift then bears on the inside circumference of the bush, and even a brass drift will cause some small deformations around the rim – sufficient to stop it going back onto the spindle and necessitating some fiddling about with a scraper. Chamfering the end of the drift should at least reduce the damage, and hopefully eliminate it.
For re-fitting to the new pulley, I bored the pulley an interference fit for the bushes, and pressed them in. Started them in the lathe (pulley in the chuck, bushes pushed by the tailstock via a shouldered mandrel) to get them aligned, then finished them off in the bench vice.
Oh, at this stage, the pulley steps had been left 0.005" over diameter and the grooves had not been cut. The o.d.'s and the grooves were finished and cut with the pulley mounted on another mandrel to ensure that they were concentric with the bush bores, and, of course, the spindle.