You either die an MVP or live long enough to build content moderation • Hacker News Board

Re: You either die an MVP or live long enough to build content moderation

#331

Post by prawn » Wed, Sep 29, 2021, 4:17 AM UTC

derekzhouzhen wrote: post #328

prawn wrote: post #286

Earlier quoted context omitted.

You will get dominated by bot-generated spam at the very least. The bots don’t care that their content will be seen by one person, they just hammer every form they find, and the recipients will be overwhelmed having to delete this, and they will criticise your platform. As well as bots, you will get people manually hitting it. They’ll often use software customised to spam you. There is zero chance this model survives…

I am flattered. Let's say my platform somehow get popular, and bots is hitting it for no financial reason at all, they still get the one reply per node limit. Also I have a global per user replies limit per day. If you have an orchestrated army of bots hitting me, then I would get DDoS'ed, just like everyone else. However, that has nothing to do with content moderation at all, right?

You don't need to be popular or flattered. If you are linked from anywhere and in search engines, they will find you. And they will reply to every node they can. And they'll create new user accounts to submit more.

I've gradually built my setup to filter based on words, URL fragments, block IPs, honeypots, shadowbanning, etc and they still get through. Take away all content moderation and I would be overrun to the point that my site was useless to my users. Those sort of methods are the easy bits too; dealing with greyarea trolls is the worst.

Good luck. You'll find out.

Re: You either die an MVP or live long enough to build content moderation

#332

Post by crakhamster01 » Wed, Sep 29, 2021, 4:22 AM UTC

dylanjha wrote: post #2

Hey there :). Author here. It was a fun experiment to play around with some different strategies for adding content moderation to https://stream.new Hive ended up being the one I landed on after trying Google Vision first ( https://cloud.google.com/vision ). The other one I was looking at is Clarity.ai but I didn't get a chance to try that one yet.

Great write up! Curious if you/Mux have ever considered offering content moderation alongside content hosting? Seems like most platforms do one or the other, but I imagine you could charge quite a premium if you offered both in tandem.

Re: You either die an MVP or live long enough to build content moderation

#333

Post by avs733 » Wed, Sep 29, 2021, 4:34 AM UTC

rtkwe wrote: post #71

It's really funny every time a "we don't censor" platform pops up catering too the American right they speed run going from moderation == censorship to we're moderating our platform in record time. Turns out moderation is really important to make a platform for a community.

your statement is worth unpacking a bit...because it suggests some difficult questions:

what does it suggest about how that group defines 'censorship'?

do they really start without censorship? or just with censoring ideas that don't comport with the american right?

what does it tell us about any argument or claim they make?

do you want these people as part of your community?

Re: You either die an MVP or live long enough to build content moderation

#334

Post by avs733 » Wed, Sep 29, 2021, 4:36 AM UTC

fragmede wrote: post #250

Sohcahtoa82 wrote: post #189

Earlier quoted context omitted.

If you create a platform with absolutely zero censorship, you become a repository for child porn. I participated in Freenet many years ago because I liked its ideas (And thought it would have been a nice way to pirate games without my ISP being able to know), but it got a reputation for being used for CP, and I promptly deleted it, because I want no part in that. If you merely censor illegal content, you will become…

> See Parler. Gab may be a more useful point of reference, but what's hilarious about right-wing platform is that their censorship is fine, it's the other people's censorship that's the problem. (Similar to how immigrants having fake papers is wrong but having a fake vaccination card is sticking it to the man). Go post some pro-vaccine or pro-mask mandate things, and see how long you last before being deplatformed.

Agnotology and bad faith arguments[0]

[0] https://medium.com/s/story/a-field-guide-to-bad-faith-argume...

Re: You either die an MVP or live long enough to build content moderation

#335

Post by nbardy » Wed, Sep 29, 2021, 4:36 AM UTC

Alir3z4 wrote: post #17

Around 1 year ago we got hit badly on our [blogging platform][0] by people/groups submitting fake customer support description of other big companies, either being Microsoft, Facebook, Comcast etc. Rolled out a machine learning model and trained it on the database. 99% of them vanished. Next day, the machine didn't work and success rate was around 5%. Found out, they have learned the trick and now using symbols from…

That sounds like a killer saas startup

Re: You either die an MVP or live long enough to build content moderation

#336

Post by marcthe12 » Wed, Sep 29, 2021, 4:38 AM UTC

bogwog wrote: post #199

duxup wrote: post #68

So true. In HN articles where we discuss social media moderation there's often this idea that "they shouldn't be doing this at all". But I think for most companies and users ... they won't like what a completely moderation free site looks like. So here we are with this painful problem. I kinda wish there was an imaginary "real person with an honest identity" type system that did exist where we could interact without…

> I kinda wish there was an imaginary "real person with an honest identity" type system that did exist where we could interact without the land of bots and dishonest users and so forth. But that obviously brings its own issues. That sounds like it could be done in a way that isnt terrible. As a user, you sign up with an identity provider by submitting personal documents and/or an interview to prove that you’re a real…

Well this sound very similar to PKI and domain registar so the issues with them are pretty same.

But like PKI, if we can have id providers delegated by high authorities, the roots could be governments which could delete to smaller entites and so on. Not to much diff from pasports or liscence

Re: You either die an MVP or live long enough to build content moderation

#337

Post by zerocrates » Wed, Sep 29, 2021, 4:40 AM UTC

jart wrote: post #173

airstrike wrote: post #97

Earlier quoted context omitted.

You should post this to Show HN. Also you have a typo on your README ("characgters")

The one built-in to Python will get you most of the way there: >>> import unicodedata >>> unicodedata.normalize('NFKD', '𝓗℮𝐥1೦𝗵𝗲𝗹𝗹𝗼') 'H℮l1೦hello' Obviously it isn't going to remap leetspeak characters like 1 -> l but it covers a lot of cases.

Obviously you're saying it doesn't cover everything, but a big thing it's not going to catch beyond leetspeak-type situations is the kinds of thing you (used to) see in internationalized domain spoofing: legitimate non-Latin-script letters that just look the same or nearly the same.

NFKC/NFKD will handle "this is another form of the Latin letter A" type stuff but not "Cyrillic A looks like Latin A."

Re: You either die an MVP or live long enough to build content moderation

#338

Post by zerocrates » Wed, Sep 29, 2021, 4:44 AM UTC

chefandy wrote: post #101

infogulch wrote: post #77

Earlier quoted context omitted.

If you can identify text written with mixed glyphs just ban it outright. Normal users don't use text like this, the pure binary presence of such "homomorphic" text at all is probably a better signal for spam than whatever your neural net when running it after normalization.

> Normal users don't use text like this I think that depends on the users. People copying and pasting bits of text that was in English or another common language— think documentation, code, news articles, tweets, etc.— with a different character set could be problematic. Also, 𝒮ℴ𝓂ℯ 𝒜𝓅𝓅𝓈 marketed as "𝔽𝕠𝕟𝕥𝕤 𝕗𝕠𝕣 𝕤𝕠𝕔𝕒𝕝 𝕞𝕖𝕕𝕚𝕒" would be ℭ𝔞𝔲𝔤𝔥𝔱 𝔲𝔭 𝔦𝔫 𝔱𝔥𝔦𝔰. (math symbols) A user base with…

I've noticed much more usage of alternative Unicode ranges for numbers/letters in email subjects lately to make marketing messages stand out, too (in addition to emoji of course), though I wouldn't necessarily mind banning that...

Re: You either die an MVP or live long enough to build content moderation

#339

Post by derekzhouzhen » Wed, Sep 29, 2021, 4:50 AM UTC

ketzo wrote: post #262

derekzhouzhen wrote: post #232

Earlier quoted context omitted.

Flame wars, abuse etc. only happen if there are other people watching. Why would anyone try to abuse someone of which they have next to zero info for no personal gain, and no one can witness your triumph anyway?

Is this a rhetorical question? Because people do that literally all the time. This is an extremely real problem.

There are genuinely mean people, and people financially or psychologically incentivized to be mean. Most, if not all, mean people I encountered in my life belong to the second group. Maybe you live a much more interesting life than mine. I dunno.

Re: You either die an MVP or live long enough to build content moderation

#340

Post by pitay » Wed, Sep 29, 2021, 4:50 AM UTC

ljm wrote: post #217

mrkurt wrote: post #74

Earlier quoted context omitted.

We shadow ban abusive users on Fly.io and it works great. Everything seems to work right up until they try to connect to a deployed app. It took me a while to realize that ramping up the frustration level is, itself, a helpful deterrent.

I remember an old mailing list discussion on sourcehut, because sourcehut provides a build service that you can use for automation. The decision from sircmpwn was, at the end, to charge money for the service. Charging money and KnowYourCustomer will kill most exploits dead. In this sense, this is turning the frustration level to 11. You can use the service to a certain extent, without frustration, but if you want to…

Sounds like a micropayment system where there is much more significant first buy in amount to be on the micropayment system than the actual (tiny) cost of a micropayment to use a service would make it worth far less worth it to spam online services.

The risk of having their payment identifier/address banned from services before significant use makes it very risky for them to use such a thing even if tiny micropayments a worth it to the spammer.

It certainly could have other problems with people getting banned from such a system for things other than: spam and other use detrimental to the service provider. There is also the issue of how the initial buy in fee is distributed.

But a high buy in for a system that many online service providers use would very strongly discourage use detrimental to the services providers (I think; this is only for discussion, as this is posted by someone with little knowledge on this. Micropayment systems have been talked about a lot, but I don't remember high buy-in mentioned).

Edit: Forgot to mention that the idea of this is that service providers can offer their services at lower cost because the risk to them from a account/address with a high buy in is lower than from an account/address with no buy in.