If there is a huge discrepancy in dates, your browser could be loading for significant amount of time (and show a huge number of revisions) before getting to the other user.

This isn't necessary. Instead of fetching list=contribs&ucuser=User1|User2, make your first request list=contribs&ucuser=User1 and your second request list=contribs&ucuser=User2. Take note of the earliest timestamp of the 500th most recent edit for both queries, compare them, then continue the query of the user with the latest timestamp and update the timestamp to the 1000th edit. Compare timestamps again and repeat. This easily generates sensible partial results.

Less important now that we're doing T187406: If no Start Date is specified, use the more recent registration date (or the more recent first edit date)