You can actually do it both ways, but which is easiest for whom is different.

Also keep the size low. If you’re having a hard time at 20cm from a 4k 30” monitor, it won’t come easy. Zoom out.

I cross them first, then slowly relax them and as the two pupils start to move slowly back out I tell myself that the middle image is further away and "real".

It helps me to see the depth and then properly focus to cross them very slightly to start, then as I see the image my eyes adjust to pull it in focus properly.

Yep, I didn't need to fully cross them. Which is good, because that is painful.

There are two methods, either you cross them either you do like you’re describing.