Oh my god she's right. From the hashtag mentioned. on damus and nostr.band. i don't know what Amethyst does, or if i have a setting on, but it has all the account's images blocked for me. Don't want to accidentally send someone to it without those filters off.

Reply to this note

Please Login to reply.

Discussion

this is even beyond hiding it, or a malicious relay. Its infected a popular hashtag.

They do that to attack the network and upset users. Fed posting.

Sounds about right.

The problem is the media is being hosted on primal. They don’t seem to be scanning for this stuff like nostr build does

I would contact the primal team with a link to the user and ask them to start scanning for this stuff or else it’s gonna keep happening.

Would also be good to see if they have information on their IP so they can send this information to the proper authorities.

They don't respond to me.

Does viewing of hashtags and global use specific relays or is it also using outbox model?

Hashtags and global are currently risky spots where attacker can post csam.

I think focusing on nostr is wrong. Best to put pressure on the media hosts who are actually hosting illegal content.

I think that's fair, but I also think that any solution that reduces the problem in any part of the stack is worth doing. Defensive clients, ML models, filtering relays, good Media servers, easy ways to report csam, Web of trust solutions, etc.

go back to twitter

Why? I have been here since 2021.

Are you seeing what this account is posting on this thread anyone? The one that says hi.

Unfortunately I can see it. Thankfully it’s not involving actual children but the content is still wrong.

Reporting and blocking their profile.

Sadly that’s the best I can do.

I thought scanning for that was industry standard practice.

If it is primal, then yes they need to preemptively screen that. But this a different point - I don't think there is any argument for the post itself to stay on the relays, even if it is 'just' text.

I, and other users want to interact as minimally as possible with that radioactive content. A relay report was sent to damus and nostr.band, not Primal, because those are the relays my client found the note. I'm not going to find the text link.

It is primal. Taking down the media is the most effective method. Notes and Links are easily copied and replicated. People could auto-mirror and blast the note to every public visible relay. nostr is censorship resistant this way. if it was easy to get every relay operator to take down the link it wouldn’t be. Taking down the media is the only thing that makes sense.

In the outbox model you don’t even get to choose while relays you pull from, so the person could just set their relays to ones that don’t delete things.

This is also why I am skeptical of censorship resistant media protocols. Sounds like a great way to empower pedos.

What I would like to see is a protocol i can implement to report things to media hosts

Is there a relay query that echos its configs so you can choose relays based on settings?

We’re moving to a more decentralized model where users don’t choose relays since that doesn’t scale

How does the outbox model even choose the relays then?

Is there redundancy in case a relay goes down?

Will clients even be in sync? I imagine global is worse off not showing everything

the way i would prefer it is use nip05 relays to get their relays, if they have no nip05 try your best based on popular bootstrap relays. Once you have their relays you know where a particular user reads from and writes to. So when sending them a note or replying to them you select their read relays. When reading their notes you pull from the relays they are writing too. Pretty straightforward.

Gotcha. I bounce around and use all sorts of clients right now to feel out how various features are being implemented. I’m not client committed.

What if I just want to push to my own relay but want to use a good client I found?

Would client selected relays start segmenting the landscape of users?

Client influenced segmentations of communities isn’t ideal if I want unlimited reach and exposure to varied content. Is that the eventual result if my client chooses relays for me without my input?

Or any I missing something?

The point is that you should be able to set any relay you want and clients should be smart enough to pull from the right relays. Outbox reduces segmentation. It would be transparent.

Users are segmented on clients that strictly follow a users relay list (like damus ios atm)

Thanks for the clarification. 🤙

When I first showed up I thought I’d be able to run my own relay and broadcast to that for my content to be found there, archived there, and just self hosted. But most clients won’t even respect my ws://[address].onion relay address.

Fixed.