Live data from Hacker News

We’re receiving about 3,000 reports/hour

bsky.app

31–40 of 274 posts

Re: We’re receiving about 3,000 reports/hour

#31
post #5

Earlier quoted context omitted.

Moderation is orthogonal to free speech. They are separate concerns.

No it’s not. More moderation, more false positives, less free speech. Just having ‘moderation’ means the speech is not ‘free’.

So if I posted your home address, social security number, bank account and routing numbers, your work address, pictures of you, your spouse, your kids, the schools they go to, your license plate numbers, pictures of your car and its real time location that moderators can't take that down if they believe in free speech?

Interesting world we live in then.

Re: We’re receiving about 3,000 reports/hour

#33
This is the big challenge of any platform for user-generated content and it's incredibly difficult to scale, do well and be economical. A bit like CAP, it's almost like "pick 2". You will have to deal with:

- CSAM

- Lower-degree offensive material eg Youtube had an issue a few years ago where (likely) predators were commenting timestamps on inocuous videos featuring children or on Tiktok videos with children get saved way more often. I would honestly advise any parent to never publicly post videos or photos of your children to any platform, ever.

- Compliance with regulation in different countries (eg NetzDG in Germany);

- Compliance with legal orders to take down content

- Compliance with legal orders to preserve content

- Porn, real or AI

- Weaponization of reporting systems to silence opinions. Anyone who uses Tiktok is familiar with this. Tiktok clearly will simply take down comments and videos when they receive a certain number of reports without it ever being reviewed by a human, giving you the option to appeal

- Brigading

- Cyberbullying and harassment

This is one reason why "true" federation doesn't really work. Either the content on Bluesky (or any other platform) has to go through a central review process, in which case it's not really federated, or these systems need to be duplicated across more than one node.

Re: We’re receiving about 3,000 reports/hour

#35

Earlier quoted context omitted.

How would they make it harder / reduce bots without sacrificing privacy (such as SMS/ID verification/etc.)? I think if you can realistically solve that you'd be a millionaire already.

That problem is unsolvable

What about using TPM modules? I've been researching these modules lately, primarily for use in online video games. From my understanding, you can use TPMs to effectively ban players (TPM ban) based on their hardware. This would mean every time an account is banned, the bad actor would have to switch to a different TPM. Since a TPM costs real money, this places a limit on the scalability of a bad actor.

Re: We’re receiving about 3,000 reports/hour

#36
post #35

Earlier quoted context omitted.

That problem is unsolvable

What about using TPM modules? I've been researching these modules lately, primarily for use in online video games. From my understanding, you can use TPMs to effectively ban players (TPM ban) based on their hardware. This would mean every time an account is banned, the bad actor would have to switch to a different TPM. Since a TPM costs real money, this places a limit on the scalability of a bad actor.

Cool, if you can require them for every possible interaction on a platform but even that violates privacy if you have one universal value that ties it all together (the identifier of the specific TPM).

It's just the phone number/email issue but tied to hardware. If you think these things won't leak and allow bad actors to tie your accounts across services then I have some lovely real estate in Florida you may be interested in.

It also appears that resetting a fTPM works around this since it fully resets the TPM. Even if it didn't then people buying used CPUs could find that they're banned from games that they've never even played or installed on their system before

Re: We’re receiving about 3,000 reports/hour

#38
post #19
post #13

I'm interested to see how Bluesky ends up handling bad actors in the long-term. Will they have the resources to keep up? Or will it become polluted like the other large platforms. Also, if a part of their business model will be based off selling algorithmic feeds, won't that mean more spam is actually good for their bottom line because they'll sell more algorithmic feeds that counter the spam?

The AT Protocol already accounts for this. There will eventually be community-built content labelers and classifers that you can subscribe to to rank and moderate your own feed however you want.

I understand the moderators working for the big social networks have a terrible job and often see the worst the internet has to offer.

Who is going to do that job as a volunteer? Or is that expected to be solved by technology? Hard to imagine them achieving what Google, Facebook etc could not reliably.

Re: We’re receiving about 3,000 reports/hour

#39

Earlier quoted context omitted.

I don’t think you realistically can. I’d instead approach it from limiting the reach of new accounts until proven as good actors. Or switch it back to invite only, as there’s a massive userbase now, and if you invite a problematic account it becomes a problem for your account too. Operate on a vouch system.

> good actors Aha... dont be naïve... what is the definition of "good" in 2024? Take the US population for example... 50% will say your intentions are "good", the other half will not!

This still better than existing system.

Re: We’re receiving about 3,000 reports/hour

#40
I think the central nature of moderation needs fixed, rather than moderation itself. Real world moderation doesn't work by having a central censor, it involves like-minded people identifying into a group and having their access to conversation enabled by that identification. When the conversation no longer suits the group, the person is no longer welcome. I think a technical model of this could be made to work.

Looked semi-seriously at doing a Twitter clone around the time Bluesky was first announced, and to solve this I'd considered something like GitHub achievement badges (e.g. organization membership), except instead of a static number, these could be created by anyone, and trust relationships could exist between them. For example, a programming language community might have existing organs who might wish to maintain a membership badge - the community's existing CoC would necessarily confer application of this badge to a user, thus extending the existing expectation for conduct out from the community to that platform.

Since within the tech community these expectations are relatively aligned, trust relationships between different badges would be quite straightforward to imagine (e.g. Python and Rust community standards are very similar). Outside tech, similar things might be seen in certain areas of politics, religion or local cultural areas. Issues and dramatics regarding cross-community alignment would naturally be confined only to the neighbouring badges of a potential trust relationship, not the platform as a whole.

I like the idea of badge membership and badge trust being the means by which visibility on the platform could be achieved. There need not be any big centralized standards for participation, each user effectively would be allowed to pick their own poison and starting point for building out their own visibility into the universe of content. Where issues occur (abusive user carrying a highly visible badge, or maintainer of such a badge turning sour or suddenly giving up on its reputation or similar), a centralized function could still exist to step in and potentially take over at least in the interim, but the need for this (at least in theory) should be greatly diminished.

A web of trust over a potentially large number of user-governed groupings has some fun technical problems to solve, especially around making it efficient enough for interactive use. And from a usability perspective, application onboarding for a brand new account

Running on little sleep but thought it was worth trying to sketch this idea out on a relevant thread.

Post reply on HN