Give-em a timeout Time based mutes

Or in general - drop the event from your feed that interacts with the anoying pubkey/event

Introduce a slap. Slap people who follow shitty people to make them unfollow them. If they keep your wot shitty slap them again or unfollow. New nip idea. (I won't put it on GitHub, I never managed to convince the nip cabal of any nip idea I had)

I mostly just scroll past it, but that wordl5 spam irritate me lol

I tried a WoT filter on my relay and it was a shitshow because some people follow everyone. Had to turn it off.

That is why I don’t like web of trust. I might even be tempted to follow my enemies for short amount of time… I had seen someone who was talking about building the web of trust through interactions with notes rather than follow lists, by tracking how the users in your network and yourself react to various notes. I can’t find it anymore… it had a two wheels interface to give the feedback on notes, it looked interesting. Now I have to find it…

It's mostly only effective against low quality bots seems like, not post quality for sure

Web of trust is deranged. People block content they don’t want to see, but in no way think should be completely censored from view. It turns everyone into an accidental censor. Furthermore it provides perverse incentives. People with large accounts will intentionally block political enemies. Israelis will be paying popular accounts to block people if Web of Trust is widely used so that they end up censored from everyone’s timelines. It’s just absolutely ridiculous and perverse. Email solved these kinds of reputational problems with RBLs. People should be able to subscribe to appropriate RBLs and the RBLs should provide a reasonable delisting process for accidental inclusions.

No, but I'm only annoyed by these various online games like wordle that keep on going through but I'm too lazy to dig into it. 🫠

It would depend on the garbage-to-gold ratio of their output.

Had not thought about how easily it could be weaponized via large accounts, even worse.

The issue is there is no cost to an npub, so it doesn't scale to just block one when every reply can have a fresh nsec We've already seen this attack on nostr where every thread is flooded with gay porn from a fresh account

It might be a good idea to stop necessarily conflating follows with 'trust'. WoT based on follows is a good starting point; but, ultimately, more granularity will be needed before it can be an actually useful model for reputation. Do I unfollow the person? No - but, I follow plenty myself just for signal; not because I would say I necessarily 'trust' them.

unfollow if i have no rapport to them, but I'll invariably put them in my "nostr TV Rap" list. Which i frequently consult, on my own terms when the time is right.

Most likely. If it’s someone I enjoy I may give a heads up first.

I’d be happy if my follows would stop reposting and following Flash. If I wanted black pill propaganda I’d watch MSM

🤣

👍

This is also true. But it’s not a defense of WoT — because the same thing applies! Intentionally bad posters can make a new nsec and flood gay porn and nobody has blocked them yet. This only works as a defense if you tune WoT to make it so new users have no way to communicate on the network, and Nostr essentially becomes a closed registration social graph where nobody who doesn’t get vouched via a side channel can ever participate. Completely dead in the water. The porn flood problem is most efficiently solved via open relays blocking any post with attached media from open proxies/dronebl/etc and also any IP that has been the source of gay porn in the past. Boom. Zero impact to free speech on Nostr, complete nuking of the actual problem that is infuriating everyone. I’d like to mention that my current concerns about WoT are not some abstract thing. I am repeatedly getting people not receiving my replies and DMs until they turn off WoT in their clients. Even though *these people follow me*. I am not a bot. I am not spamming porn. I am a real person that a lot of people don’t like because of my politics. The current WoT model has you getting fucking completely nuked, even via the people who intentionally follow you, in multiple clients:

Blacklists aren't efficient on Nostr. Only whitelists work. The best usecase for Nostr is not a global square but the people you already know, your friends and family. If we build for that all the problems disappear.

Also, ignoring the scaling problems of blacklists in general, I think a small on-device LLM would be better than a blocklist because you can tell it "hide anyone who's mean to me" or whatever, full control over the rules on the user side.

Email has blacklists and moderates at greater scale than any protocol on earth, because just like making an nsec, sending email is free. There’s battle tested protocols that already have dealt with this shit for ages. Email uses a combination of blacklisting AND whitelisting, and this problem is intractable if Nostr doesn’t have both. There has to be a whitelist where you can reasonably appeal and a list of nsecs that WoT should never censor. I totally agree with the whitelisting part. I don’t believe in absolute blacklisting, but I believe that IP reputation should at least play a part as to if an open relay will accept a post with a media attachment for an nsec with no real history. People can then get whitelisted for media posting, perhaps even without their knowledge, by participating constructively in the network.

Relays don't moderate spam though. That's a generalization. Many of us have tried. But if the user enters just 1 relay that doesn't, then all your effort was wasted. So relay-side filtering isn't a reality in the present version of Nostr. The other option is client-side filtering. But then a spam attack, even if you can't see it, will effectively DoS the client because the client is still forced to download it all. You need a whitelist so you can collect only the information you want. Any filter without an "authors" property is doomed.

There's no such thing as email anymore really, if Gmail blocks you, you can't reach 90% of people

Absolutely

Slapping a boost or quote would be 🔥🔥

seems like client developers and users could have choices of relays with sanely maintained RBLs and not

Nostr can benefit from blacklists but it also needs real-time filtering (not whitelists) more than email. Rspamd is a great example with content scanning and local reputation scores. RBLs work on finite identifiers that have *some* cost. Nostr works on infinite numbers of pubkeys. This is the approach I am working on with nostr.land, where it combines multiple different types of information. Such as blacklists, historical reputation, but also content analysis, network, etc. The goal is to make it work for 99% of new users. The other 1% can solve a CAPTCHA and wait a bit.

WoT and PoW are futile solutions, as they are actively hostile against new users, and have a huge asymmetry between real users and spammers. AI spambots when designed right can easily manipulate WoT at little cost, while WoT excludes new users. PoW is just a joke. It’s designed entirely for huge server farms, and people are expecting anything outside of that.

The only “web of trust” is simply people following people. Find an associate and follow some of his associations, etc. Repeat and prune when needed. It’s how humans have networked for thousands of years. Once you impose rules and rewards you don’t have a web of trust. You have a community with titles, censorship and an algorithm you don’t control feeding you slop.

No, it's worth seeing outside your bubble.

yes. WoT is awful. I’ve had about a dozen people using multiple clients tell me about how they don’t get my replies or DMs until they turn off WoT: And yet a significant fraction of them have received gay porn spam. WoT is anarcho-tyranny, where you can be abused by malicious, politically motivated actors, with no actual ability to stop receiving repulsive annoyances. Spamassassin uses Bayesian statistics. I’d like to see more advanced kill files available on Nostr, and the ability to easily share them. If Nostr clients had not just mute words, but full regex instead of just words, and the ability to assign a Bayesian score to each line, I think I could come up with a good ruleset. Also WoT could be a great positive input on filters. It’s useful data, but not a great sign of whether someone should be blocked as spam or not by itself.

Even without wot filters this still happens on nostr because clients don't use outbox, or use different DM protocols, or have broken wot implementations Wisp users won't report this especially if they follow you

Nostr is janky af if you haven't noticed 🤭😂

Hey, but you gotta admit that it's less janky now than it was when people were posting notes into the future 😂

I think that for public posts as well, a remote scanmer could be used as well. This would be a tradeoff.

yes, Wisp is a better implementation than most. It becomes hard to even have these discussions with the lack of a formal spec, because arguing against “WoT” becomes something ill-defined that a ton of people are going to have problems with. Am I arguing against the very idea of WoT (no), the general Nostr consensus of what WoT is (yes, which is actually in conflict with traditional ideas about WoT utility), or Wisp WoT?

Nobody knows we all say WoT then do different things 😂 like Amethyst does WoT with mute lists, wisp does it with follows, others use relays, shit is chaotic, but that might be the best way, I don't have any good answers here

Basically I just wait for attacks then try to defend it

could be fixed with a separate filter that removes that account influence on the WoT without unfollowing (still show what the account posts but filter follows of said account)

It depends on the amount of garbage that originates from that follow. The occasional garbage could be tolerated.

yes. I follow people who follow non-garbage