In my late 20s I had just started my first real job as a programmer at a small company of about 10 people, I was kinda shy and hadn’t talked to people much. One day a woman was talking about her trip to New York and how there was this famous hotdog stand near the statue of liberty and how the seagulls just terrorised her and people buying hotdogs, swooping in to steal the sausages right out of your hands.
People were laughing and chatting about it and I joined in by wondering out loud if you could put a mannequin on a bench there and put hotdog meat in its eyesockets and teach seagulls to think that human eyeballs were made of hotdog.
The room went completely silent.
I’m now married to the secretary from that job. This was the first thing she remembers ever hearing me say.
The difference between auto-generating hundreds to thousands of words, versus having to manually read and type all of those words out yourself is very obviously a lot more than “nothing”.
And that’s not even taking into account the fact that we don’t know what their method of watermarking is. From how they’re talking it sounds like it will be algorithmically embedded in the text itself somehow, through specific choices of words or something similar. If that’s the case, typing it out word for word will retain the watermark.