pull down to refresh
No. I just have this uninspected belief that pay to post fixes slop/spam/bots, but as I've been thinking about it, I am forced to admit that this is clearly not the case. So, I am wondering if the problem is that the cost right now is still so low that it does not really deter bots.
It is also possible that I need to rethink my mental model of pay to post. Perhaps there is no world where pay to post fixes spam and bots entirely, and the gain is always to remove the extremes. (Except I don't think it's doing that either, right now).
Most bots use freebies for comments. I went through exactly this motion this year:
I set comment fees to 45 sats on ~AI. This was prohibitive for human comments
Then I lowered to 10 and this was ok, but then the freebie nostr-key driven cycle bots came in[1] and I just lowered it to 5 because bots can fee and mute evade on comments this way.
you basically only need snappy and some nsec encoder and you have unlimited freebies, which at least 1 person is running, but I think I've spotted at least 3 distinct entities ↩
I asked k00b to allow territory owners to disable freebies. That would at least close the loophole.
Cross-posts must comply with the rules of each territory, so all restrictions apply.
Cross-posts are also the limiting factor for raising comment fees, which is why I think we should move to OP-defined reply fees or parent-defined.
Sorry, which part?
The concept of "crossposting" in the abstract is:
TERRITORY >--many-----many--< ITEM --one -----many--< COMMENTThe implementation from earlier this year solves this by:
TERRITORY >--many----- one-- ITEM[id, territories[], children[]]i.e. the relation of territories-to-items is a list of identifiers intrinsic to the item. This makes it so that the item - which in turn carries comments - is leading, within the strictest constraint set by territories.
However, you could also design the relation to be an instance of Item.
TERRITORY --one -----many--< ITEM_INSTANCE[item_id, territory, children[]] >--many----- one--ITEMThen, a different set of items (because they're instantiated against territories), can carry different rules, and suddenly, I could block a post on ~AI that is scamming people but you could let it sit.
The problem with that is that you need to have an entry-point into a territory, i.e. like Reddit would do with subreddits (but they copy the item, so it's different) and that is where it gets tricky from a UX perspective and the current UX makes it more intuitive, at the cost of territory "sovereignty".
Nevertheless, the reason why I ask is because we discussed this before and as you may remember, I expect abuse. Of course, it is easy to gamble that it may not happen, until it does. Like with every other thing that can be abused right now either being abused, or if you pay attention to some details, is being set up to be abused, such as territory owners getting a 70% discount on boosts, territory owners being able to defame other stackers and setting the comment fee high (profiting off defamation) and so on. There is a lot of naive things right now (systematic trust assumptions towards territory owners being one - and we know about the whale-bias we have at the moment too) that I know with 100% certainty will be abused in the future. Doesn't take a degree to figure out who has already gamed this out.
Do you really think 45 was prohibitively high? I'm tempted to see what happens with higher comment fees because so many territories are at 1 or 2.
Some of us are not bot we just need help.From Lagos here. Naira is 1500+ to dollar now so saving is hard. My shop work pays small but I want to convert small small to sats weekly. Do you stack daily or weekly? How do you avoid fees?
The risk/reward ratio for bots posting freebie comments or 1-sat cost comment is really skewed towards reward. If even one stacker accidentally zaps say 40-100 sats, that pays for 40 to 100 one-sat comments just on its own.
That being said, I am not sure how many of the bots are economic bots vs. someone's personal lab experiment.
it's kind of hard to tell them apart these days. (At least it costs too much attention and follow-up to do so.) And on high-rewards days, the incentive is to just spam out zaps to everyone...
Do you have a good optimal/curation/fix for this?
Only zap comments that you're pretty confident aren't from a bot, if you care.
Perhaps there is no world where pay to post fixes spam and bots entirely, and the gain is always to remove the extreme
that's the uncomfortable alternative explanation. ooopsies
Did you have a bot problem during an ama?