I dd some quick checking though my notes and played with it. There is an undocumented override to the maximum number allowed.

The issue is that from a computational point of view drag & drop is far more efficient. You select the files in iTunes and start the drag. iTunes does all the work in machine code. The Yate function posts a single AppleScript call which then enumerates the selected tracks and has to extract each track's address and add it to a list. When the list is complete, it is returned to Yate. This is done in AppleScript. Yate is blocked during this process and if the request takes too long it will time out. I've just tried it with the max set to 5000. It works and I cancelled it once control was returned to Yate. Yate did spin for about a minute before this happened. I can't remember what the system timeout is for AppleScript events but I think it's around 3 minutes. So, in theory if the max was north of 15000 it more than likely would time out.

As I said it's not great to open tens of thousands of tracks in Yate by any means. Depending on your system memory it might be fine or it could cause a crash.

If you want to try it reach out to support@2manyrobots.com and I'll tell you how to do the override.

I'm a big proponent of working on a single album at a time. My personal library has about 45k tracks in it. I've saved the PIDs in my files (originally through the Batch Extract Persistent IDs process and now through Auto-Import of the PIDs). I tend to drag in an album from the Finder and link it; do my work; save the files which updates iTunes. Whenever I have to do something over my entire collection I use the Batch Processor. I do not do anything manually in iTunes (or now Music) ... everything is driven from Yate.