From what I've seen over the years, getting rid of obvious spam makes some of the spammers switch to something that's much harder to detect.
When plagiarism got busted, they switched to a homograph attack. When that was made impossible (by theymos), they switched to word spinners. Now it's AI-spam. The main problem is: it becomes less and less obvious what's spam, and more time consuming to detect.
Yeah, that's true. But, I like to think that I can tell when I'm reading a redundant post [1], and I would completely trust
your judgment, and o_e_l_e_o's, Foxpup's, n0nce's, DarkStar_'s, DireWolfM14's, etc. in determining the same. (Not exhaustive and in no particular order, I'm just picking a few names off the top of my head.)
I really feel like people are sleeping on what a
good thing a properly implemented version of this idea would be. Anyway, my thoughts on this are spread out over my last few posts, and I've done a bang-up job of derailing NAT's thread, so I'm bowing out for now. (PM me for continued discussion, if you like.)
[1] I guess, my definition of "redundant" (for the purposes of this system) is when I can
defensibly say "no" to the following question: "Is this post likely to be missed (by anyone but the author) if it were deleted?". Now, before anyone says: "Yeah, but who are
you to determine that?", that's the
whole point of the system, no individual would have the power to affect things too much, and if mobs form that end up misusing the system en masse, then an "open logs" policy and the threat of negative feedback will put things right again pretty quickly.