Some of the Hubble results were also raising questions. At the same time, I read one of the papers on the galaxy stuff, and what struck me was they were identifying galaxy shapes by counting the pixels each galaxy had, so there are definitely some question marks over how they do some of this.

You would expect more background pixel fuzz when centering an image kernel over an artefact.

In Hubble, that fuzz was marked. With Webb, far less so.

I think these are real true positives