Today Rachel Greenfeld and Terence Tao announced via the arXiv that they can disprove the periodic tiling conjecture (arXiv:2209.08451). I do not want to discuss here in detail the contents of the conjecture or their approach to disprove it (if you are interested in this, you can read about it on Tao’s blog).
What I want to do instead is to utter my astonishment about this practice: Why putting an incomplete paper on the arXiv claiming to have disproved an at least 50 years old conjecture, if the complete version of the paper is only a few weeks away (as Tao himself says on his blog)? If it is about priority: The chances that somebody else finds an argument for a 50 years old problem at the same time as you are negligible I think (and since so many people follow Tao’s blog, it would have sufficed just to write for him the blog post).
The arXiv has already the problem that it contains many wrong preprints, where the authors even know that they are wrong but still do not withdraw them. If everybody now additionally starts announcing his or her results by putting half finished papers on it, where I am sure that half of them will never see a final version, the arXiv will be clogged with erroneous stuff.
(I am not saying here that the argument of Greenfeld and Tao will turn out wrong. I am just expressing my personal opinion that it will be quite bad for the arXiv if it becomes standard practice to upload announcements.)