Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

We’re receiving about 3,000 reports/hour

bsky.app

Post reply on HN
Print view View original

51–60 of 274 posts

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

#51

Post by bigbones »

photochemsyn wrote: post #43
bigbones wrote: post #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. Looke…

The whole point of online discussion IMO is not to join some little hive mind where everyone agrees with each other (eg many subreddits) but rather to have discussion between people with different information bases and different viewpoints. That's why it's valuable, you learn new things and are exposed to different points of views.

[flagged]

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

#52

Post by bdangubic »

luckylion wrote: post #38
paxys wrote: post #19

Earlier quoted context omitted.

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.

you really think Google/Facebook/… can’t do it reliably? :-)

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

#53

Post by OutOfHere »

pfisherman wrote: post #24
OutOfHere wrote: post #4

[flagged]

Why use an LLM as opposed to a more narrow purpose built model? LLMs are not beating smaller, purpose built models on tasks like POS tagging, NER, sentiment analysis, etc. And the inference costs scale quite poorly (unless you are self hosting llama or something).

That's where "rapidly" comes in. Also, LLMs allow very high customization via the choice of prompt. It's a lot quicker to adapt the prompt than to retrain a fine-tuned model. I think the outputs of the stabilized LLM could later be used to properly fine-tune a custom model for efficient use.

As for sentiment, even embeddings can do a good job at it.

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

#54

Post by Waterluvian »

paxys wrote: post #19
voat wrote: 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 have a feeling that this is going to create a weird thing of some magnitude where accounts end up on popular blacklists for poor reasons and have no recourse.

I’m concerned that in time it might develop into zealous communities of samethink where you have to mind any slightly dissenting opinion or you’ll get blacklisted.

I think what I’m thinking about is essentially that judges cannot be replaced by community opinion. (Not that Twitter moderation was less bad).

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

#55

Post by nicce »

DrillShopper wrote: post #36
calebh wrote: post #35

Earlier quoted context omitted.

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 inte…

> 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

It depends how the TPM utilization was applied in practice. The initial manufacturer key (Endorsement Key) is hardcoded and unextractable. All the long-lived keys are derived from it, and can be verified by using the public part of the EK. Usually EK (or cert created from it) is directly used for remote attestation.

More here, for example : https://learn.microsoft.com/en-us/windows-server/identity/ad...

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

#56

Post by almatabata »

bakugo wrote: post #49

The audience Bluesky is currently cultivating is the kind of audience that mashes the report button every time they see something they disagree with, so this isn't surprising. If the user base actually keeps growing at a steady rate, I don't see how they'll get the resources to deal with so many reports (especially since they don't seem to have a plan to monetize the site yet) without resorting to the usual low-effor…

> without resorting to the usual low-effort solutions, such as using some sort of algorithm that bans automatically based on keywords or number of reports.

Or you prioritize reports from accounts with a proven track record. If I consistently report posts that clearly violate the rules why shouldn't my report count more than an account that just got created?

If you consistently report nonsense then you should accumulate negative karma until at some point you can safely ignore whatever they report in the future.

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

#57

Post by ben_w »

ranger_danger wrote: post #7
hipadev23 wrote: post #6

Due to how easy it is to setup accounts and post on Bluesky, it’s likely many of the same operatives behind the propaganda and bot armies on Twitter are now pushing the same vitriolic content, and triggering these reports. If they can negatively impact Bluesky at a critical moment, it’ll reduce the flow of users who will quickly surmise “oh this is just like twitter”

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.

> I think if you can realistically solve that you'd be a millionaire already.

Please.

If I knew how to do that, or even how to reduce bots even with SMS verification etc., I'd be a multi-billionaire at least.

Making a twitter clone is relatively easy, making a community with a good vibe that's actually worth spending time using is the one single problem that makes none of the clones stand out to normal users.

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

#58

Post by ltoph »

bigbones wrote: post #51
photochemsyn wrote: post #43

Earlier quoted context omitted.

The whole point of online discussion IMO is not to join some little hive mind where everyone agrees with each other (eg many subreddits) but rather to have discussion between people with different information bases and different viewpoints. That's why it's valuable, you learn new things and are exposed to different points of views.

[flagged]

[flagged]

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

#59

Post by ben_w »

jrvarela56 wrote: post #9
whaaaaat wrote: 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’.

Counterpoint:

  while True:
    requests.post(api_url, json={"username": "@jrvarela56", "message": "hello"})
If this was allowed to run without moderation, targeting your account on some social network, it would effectively jam you from receiving any other messages on that network.

Moderation is noise reduction, undesirable content (junk mail, male junk, threats, even just uninteresting content) is noise, the stuff you want is signal, usability requires a good signal to noise ratio. Speech can be either signal or noise.

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

#60

Post by bdjsiqoocwk »

fullstackchris wrote: post #26
hipadev23 wrote: post #10

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!

So maybe think with your own head instead of just taking the average of everyone else's opinion.
Post reply on HN