-So are we supposed to keep a file of our rejects so we don't forget what we already submitted?

I always do so i can keep track of what ive sent where and whats accepted.  It WAS useful for SS when they went through the random batch-reject nonsense not that long ago, you knew to just resubmit the next day.

-SS is using AI to decide if the image was previously submitted.
Wonder if we can fool it with a new pixel size, upscale, color rebalance and recrop?

Lets face it, the algorithm (hate the term AI - its an algorithm) has been capable of identifying identical and similar images BUT its never been applied.  People can upload stolen content identical to another users already online and its accepted despite the original one being shown clearly in the "similar images" section of that same page.

Most rejects except for IP get "accepted" for data licensing anyway so its online and stored and is then trivial to compare that to content coming in from that same user to check.  That said, the bug a few weeks ago where some rejected content from several years ago appearing and getting rejected suggest to me they're not entirely honest about deleting ANYTHING as such.