Hacker News Reader

Recent revelations show government agencies could manipulate online polls.

It would be trivially easy for them to poison this database.

> Rascals and scammers like these failed Kickstarter bros would be shut down after a single scam ... they would not be able to rent or scam another person after a single scam.

Well, you don't know they're a scammer until they've been to court and proved guilty or does no-one care about innocent until proven guilty anymore? And what's to stop them getting a new identity?

> Imagine the first time some innocent girl in high school is unjustly called a slut, and is defamed by her own school as part of some bullying clique.

That is a fucking awful example. You say "some innocent girl is unjustly called a slut" - this implies that if the girl was not innocent that it may be acceptable to call her a slut (as an insult). Then you say "[...] as part of some bullying clique" which implies that the problem is with the false accusation of sluttiness, and not with the fact that some bullying clique exists.

> I believe the network can ameliorate if not outright prevent such clique abuse by enabling users to rate reviewers based on the reviews they have given. If you abuse people with your reviews and ratings, it can damage your own ratings.

The clique just vote-rings and they cancel out bad votes. Or they buy votes off MTurk.

I don't have any problem with women or men having sex. I don't think the term slut is ever appropriate.

Bullying is always bad.

Hope I didn't piss you off too much. I just wrote that example quickly, without much thought.

A system like this would still not account for outliers, nor deal well with temporal changes. People can be completely well behaved for long periods of time and suddenly really hit rock bottom. If you're the person who they are interacting with at the time, the trust system fails. Sure they get a lowered rating now, but then you have to ask the system, was this person who had a bad interaction being honest and how much of a ding is this to a long record of good ratings.

There's no catch all, even a single, centralized system wouldn't be very successful at dealing with the outliers in human behavior on top of the questionable honesty of the people adding ratings to the system in the first place.

I honestly don't understand how anyone could even begin to think such system is a good idea. We already have credit ratings, they are far less invasive than what you've described, they cover only one particular aspect of trust and they still suck. It will only get worse from there.

The only rating system I've seen that mostly works is Ebay. The only reason it works is because it rates clear-cut transactions that are limited in time and scope, and where both parties have something to loose. And even there you have problems, like its notorious bias against sellers.

Decentralized system would be even easier to game.

That sounds horrific, and basically like a trial by popularity.

Count me out.

I'm not so sure. Privacy is a big deal for me, but I'd sign up to something like this. If:

If it uses a mechanism like CardSpace (http://en.wikipedia.org/wiki/Windows_CardSpace) had:

=> I own my own data (client-side encryption on my client)

=> I consent to others accessing my reputation. Note that access means others can see, not get a copy of, and not keep.

=> I can withhold my reputation completely, or delete it at will.

In such a world, as Airbnb for example, I might make it easier for someone willing to share their reputation to rent a room. A non-existing or withheld reputation might require other checks (like a credit check), or verifiable references, for instance.

If thought through this might have uses. Primarily I'd need to know that it abides by the 7 laws of identity (http://technet.microsoft.com/en-us/magazine/2006.07.7laws.as...). A lot of this technology exists. CardSpace is one example of a client app (I wouldn't use a web app to manage my reputation, but would be ok to use a service to let others leave feedback about me). Attribute provisioning in identity federation is another, and could be used as a mechanism to check someone's reputation, or to give feedback.

Something like this COULD work if done well. at the same time,it is also hard for one to protect oneself against identity thieft.But I guess this kind of public only social network could also be used by third party websites as an OAUTH provider. I post as little as possible on the web about myself in fear that others would use my identity for a scam.So if someone can fix that problem,it could be a killer product.

something like the Stasi files? http://www.spiegel.de/international/germany/stasi-files-revi...

Exactly wrong. Stasi files are held closely by the state.

This network would be public, and everyone would be able to see what the scum are up to.

If everyone had Stasi files, it wouldn't really be better.

Say again how many Yelp reviews are genuine and how many are planted by SEO outfits?

I doubt that 20 % of Stasi informers were double agents, if you led the Stasi on there were consequences.

It may be irrelevant to your analogy (it'd be more difficult to identify "personal bias" in reviews of people, after all), but the "20% of Yelp reviews are fake" meme is highly misleading. The study in question looked at what percentage of Yelp reviews were hidden, by Yelp, for looking "fake."[1] This provides us with little additional information on the percentage of visible Yelp reviews (which is what most people will assume you're talking about when you say "Yelp reviews") that are fake.

It's a bit like noting that 96% of Harvard applicants are rejected and concluding that "96% of Harvard students didn't deserve to get in," with the caveat that by "Harvard student" you mean "a student that applied to Harvard."

[1] http://people.hbs.edu/mluca/FakeItTillYouMakeIt.pdf

Until such a system is in place I have a simple hack that has worked well for me - being googlable.

When you use the same name on all social media, have a domain with it, a blog on it, and openly invite people to google you if they feel like it ... and you aren't a complete arsehat, great things happen. Things like finding a place to live in SF become relatively easy, getting jobs/work becomes simpler, hell, even Tinder works better.

We already have the trust networks in place. Why are so few people using them is a better question.

> Rascals and scammers like these failed Kickstarter bros would be shut down after a single scam ... they would not be able to rent or scam another person after a single scam.

the guys rans 2 scams on kickstarter. peer-to-peer trust rating networks wont help if a plateform does nothing to prevent scams at first place. that guys shouldnt have been allowed to run a second kickstarter even though he didnt deliver a product in the first one.

Wouldn't matter what Kickstarter allows or doesn't. People would look them up on our hypothetical trust network, and tell them to get lost after seeing that they are shameless.

I don't feel these guys are scammers at least when it comes to kickstarter. They have plenty of evidence to show that they have built something (even though it's incomplete).

I feel it's more accurate to label these guys as talented people who don't know how to manage a project.

As for AirBnB, the label is accurate. They're probably mining bitcoins as well since the electricity bill is so high.

http://signup.meowmeowbeenz.com/

There's a startup that's trying to do this: https://myrepp.com/. Not sure why they're not using this case as an opportunity to get some press.

I'm pretty sure they don't allow users to rate one another though as the OP is suggesting. Instead it looks like you can create an account for yourself, verify who you are, and run criminal/sex offender checks against yourself. Then you can share your Repp profile with others (when selling on Craiglist, for exampe) so that others know you aren't a psychopath (or at least not a convicted one). It wouldn't really have prevented the squatter's issue unless the squatter has a criminal record.