the thing about the whole wattpad migration is
people who have never actually thought about the process of site moderation, don’t seem to realize that the kind of censorship they want is not only ethically wrong, but logistically impossible on the scale of a large website.
the websites that do have policies banning smut– or banning certain types of it– don’t consistently enforce such bans.
Because consistently enforcing them isn’t possible.
either they have “enforcement” that doesn’t work, and doesn’t really even try to (like here on Tumblr, where completely innocuous posts get censored for no good reason all the time while straight-up porn ads get through unimpeded)
or, they go long times without any enforcement and then have occasional random purges that freak everyone out (seems like this may be what’s just happened with Wattpad?)
actual, reliable enforcement just… wouldn’t work at all. Think about all the definitions you’d have to agree on. Think of the edge cases where mods would have to argue about whether something fit within a certain definition or not.
What counts as sex? What counts as an explicit description? What counts as consensual; how clearly do the characters have to agree to everything? What counts as incest, when all humans are somewhere in the range of 50th cousins or closer? What counts as underage, if the characters’ ages aren’t explicitly stated, or if they live in a world without the same measurements of time, or the same biology of growth? What counts as bigotry, or violence, or an abusive relationship, and what counts as the “author’s endorsement” of whatever’s being depicted?
Think about moderators having to actually read hundreds and hundreds of dark, disturbing fics a week– fics that they might personally find triggering– just to determine which ones actually broke a rule as they defined it.
(Because if they didn’t read them, and just banned the ones that had gotten reported by a certain number of people– then their “enforcement” is at the level that doesn’t even try to work, where the only thing you’d reliably get censored for is being unpopular enough with the nastiest group of bullies.)
if you think for a few minutes about what it would take to actually, reliably keep any particular type of fiction off of a fiction site– you’ll realize it can’t be done. Not beyond the scale of maybe a small website with a few dozen posts a week.
And this is why a site with an anti-censorship ideology like AO3 is the only kind of site that can even come close to reliably enforcing its own rules.
Hell, even with their very small list of offenses that a work can actually be removed for (plagiarism, harassment, commercial promotion, etc) they’ve already got their hands full, and they already run up against a fair number of cases where their decisions get criticized for being unfair.
Because enforcement of rules is always an imperfect endeavor. Rules require definitions, and definitions are imperfect.
So, if you’re gonna make a rule and tackle the task of trying to enforce it, for real– it had better be a rule that there’s a rock-solid important reason to enforce. And that you’ve actually got the time and manpower to enforce on the level it needs.
Anyone who wants websites to do more than that, has not actually given more than a couple minutes’ thought to how running a website works.

“but we’re not talking about the borderline cases, we’re talking about the ones everyone agrees are clearly just some person’s disgusting fantasy about real-life harmful things they actually want to do!”

There is no place where “everyone” draws that line.
Even if all moderators agreed on where to draw the line, they’d still have to read every fic to decide which ones crossed it. Meaning there would have to be more moderators working more hours than any free website can afford.
Even if those unambigiously bad fics could somehow be eliminated– then the window for what’s considered “unambiguously bad” would just shift, and something currently in the mid-range would be the new Evil Stuff.
We’ve seen that happen in every setting that manages to have high levels of control over what people are allowed to do.
There’s very little actual research on whether “liking fiction about doing depraved evil things” actually correlates with “wanting to do those things in real life.”
A lot of people like to claim (without sources to back it up) that “the people loudly against censorship are always the people who want to commit actual sex crimes”
But in my own experience, they’re mostly people with pretty tame preferences– because we like it when fiction about the extreme subjects is “allowed and well-tagged so we can avoid it,” rather than “forbidden and therefore posted without tags so we keep stumbing onto it unwillingly.”
And, even in the case of “people who read fiction about depraved evil acts because they want to do those things in real life”– there’s not much research on whether reading fiction increases or decreases the likelihood of them actually doing those things in real life.
As far as we know, it could do either, or both, or different things in different circumstances.
And, in the absence of actual knowledge of what works– “trying to eliminate the fictional outlet” is probably not the safest choice.

exploration of what I mean:

Now, personally, the sexy things I fantasize about are pretty tame, compared to a lot of the stuff that gets debated in "problematic content" discourse. But I think some of the same general tendencies hold true.
Namely: when I get into a sexual kink, one that I've never done IRL, and start exploring it by reading fiction about it-- there are, in my experience, four different things that can happen to that kink over time.
1. I may get increasingly interested in the kink, to the point that fiction no longer satisfies the desire and want to try it in real life.
2. I may get tired of the kink, and move on to entirely different interests.
3. My interest in the kink may remain the same, and so does the degree to which I'm satisfied with fictional depictions of it.
4. My interest in the kink may get increasingly focused, to the point that only very specific fictional scenarios will satisfy it. These scenarios never trend toward getting increasingly easy to achieve in real life. They often trend toward becoming literally impossible in real life (e.g., it only turns me on when I imagine specific kinds of fictional characters doing it).
If the kink was something truly harmful, only Option 1 would be a problem.
All the other ways that reading fiction about a kink can affect it-- all the possibilities from 2 through 4-- would actually make it become less of a problem in terms of real-world harm it could do.
So I guess the question is... whether real-world harm from option 1 is common enough to cancel out real-world harm reduction from options
2 through 4?
And as far as I can see, there doesn't seem to be any actual science exploring this question.