Over simplification but partly it has to do with how LLMs split language into tokens and some of those tokens are multi-letter. To us when we look for R’s we split like S - T - R - A - W - B - E - R - R - Y where each character is a token, but LLMs split it something more like STR - AW - BERRY which makes predicting the correct answer difficult without a lot of training on the specific problem. If you asked it to count how many times STR shows up in “strawberrystrawberrystrawberry” it would have a better chance.
Am I the only one who doesn’t care that much? People should be putting as much of the text/image directly into the body of their lemmy posts anyway to save clicks, a link to reddit when attribution feels necessary seems more appropriate in most cases because if I click the link I probably want to go to the actual source of info not an archive.
I’ll passively watch reddit kill itself with bad decisions I just don’t see the point in going out of my way to avoid linking reddit.