It doesn’t stop people posting AI slop, it stops people from posting AI slop more than once. If you ban somebody for spamming today, they just create a new account and keep on spamming. If you can determine they are the same person you banned before using verifiable credentials, it makes the ban actually effective.

Layer on captchas. It won't completely stop slop but it's an incentive against slop flooding. And I mean, nothing is stopping a human from just going into ChatGPT by hand and asking for output and copy/pasting that into an HN post box.