no, i think it's worse than that, and it's right in the study: "the nuclear correlation is just a function of which days Palomar was observing on... even that is not statistically significant. It's just noise." - https://www.metabunk.org/threads/transients-in-the-palomar-o...

another POV is the paper is sloppy in the parts that matter

Machine Learning goes both ways. A chatbot is not predisposed to ruin aliens enthusiast's days. It just does what it is told to do, like repro a paper, and it can tell you the problems in some limited, but globally important, objective way, and it did, and the paper has problems, and they're basic.